Information_Retrieval_Management_1962 Information Retrieval Management 1962
Information_Retrieval_Management_1962 Information_Retrieval_Management_1962
User Manual: Information_Retrieval_Management_1962
Open the PDF directly: View PDF 
.
Page Count: 153
| Download | |
| Open PDF In Browser | View PDF |
INFORMATION
RETRIEVAL
MANAGEMENT
EDITED BY
LOWELL H. HATTERY and EDWARD M. McCORMICK
© 1962 American Data Prace~sing, Inc.
All rights reserved.
Library of Congress Catalog Number: 62-18060
first Edition
Printed in U.S.A.
About the Library Series . ..
The DATA PROCESSING LIBRARY SERIES was inaugurated in 1962 to
answer a widespread need for reference information in depth about
selected subject areas in the field of business data processing. To provide
this infprmation, American Data Processing, Inc. sought out leading
authorities in these subject areas and arranged for their assistance not only
in the collection of material, but also to· help ensure that the contents of
the Series are held to exacting standards of quality and timeliness. With
the library Series, the manager or researcher in business data processing
is assured a continuing supply of authoritative and definitive reference
books covering all important aspects of the subjects they explore.
About the Publisher ...
AMERICAN DATA PROCESSING, INC. (formerly Gille Associates, Inc.)
began research on its first data processing information service in 1950.
The resulting publication, The Punched Card Annual,· became the first
regularly Pl,blished national reference on data processing systems.
Since that time, American Data Pracessing has continued to lead the
way in making available the newest data processing information in· the
most usable forms.
These now include:
DATA PROCESSING -
a monthly magazine.
DATA PROCESSING EQUIPMENT ENCYCLOPEDIA - detailed standard
descriptions and specifications for all office data processing eqUipment.
COMPUTER APPLICATIONS SERVICE in hard-bound form.
case histories issued semi-annually
DATA PROCESSING LIBRARY SERIES
DATA PROCESSING YEARBOOK -
a reference guide.
PUBLISHED BY AMERICAN DATA PROCESSING, INC.
2200 BOOK TOWER
DETROIT 26, MICHIGAN
Publisher
Editor .
Assistant Editor
. FRANK H. GILLE
ALAN D. MEACHAM
. VAN B. THOMPSON
Advisory Committee
To The American University for the Fourth Institute
on Information Storage and Retrieval
SAMUEL
N. ALEXANDER
Chief, Data Processing Syste'tns Divisions, National Bureau
of Standards
LEA M. BOHNERT
Chief, Information Retrieval Section, LibraTY Branch, Federal
Aviation Agency
GEORGE P. BUSH
Emeritus Professor of Government and Public Administration,
The American University
W. CLAPP
President, Council on Library Resources, Inc.
VERNER
LOWELL H. HATTERY
Professor of Government and Public Administration, and Director,
Center for Technology and Administration, The" American University
F. HEUMANN
Director, Office of Documentation, National Research Council
KARL
W. HOWERTON
Vice President, Information f01' Industry, Inc.
PAUL
D. LEAKE
Retiring President, American Association for the Advancement
of Science
CHAUNCEY
EDWARD M. MCCORMICK
Office of Science Information Service, National Science Foundation
CATHERYN SECKLER-HUDSON
Dean, School of Government and Public Administration, The American
University
Contents
PREFACE
6
THE CONTRIBUTORS
8
1. The Systems Concept in Documentation
lowell H. Hattery
10
2. What Must Give in the Documentation Crisis?
Chauncey D. Leake
3.
........................ .
15
A System of Information Systems
Julius N. Cahn .............. .
21
4. Status of Technical Information Centers
Paul W. Howerton
31
5.
Functions of a Technical Information Center
John Sherrod ..
. ....... .
6. Technical Information Services in an Industrial Organization
W. C. Asbury and J. E. Moise
. . . . . . . . . . . . . . . . . . . . . . ..
7.
Designing an Information Center to Meet a Real-System Requirement
DeWitt O. Myatt .........
...................
34
37
45
8. Management's Evaluation of Information Services
James Hillier
9.
................................ .
Financing a Technical Information Center
Bernard K. Dennis
54
61
10. Documentation Needs of Scientists
Helen l. Brownson ..
76
11. Communication Between Computer and User in Inf~rmation Searching
William B. Kehl ........................................
83
12. Automated Intelligence Systems
H. P. luhn
..................................... .
92
13. Guidelines to Mechanizing Information Systems
C. Dake Gull .....
. ................................ . 101
14. Compatibility of Information and Data Systems within a Company
Hattie T. Anderson
.......................
111
15. Economic Justification - Factors Establishing System Costs
Simon M. Newman ..................................... 117
16.
International Activities ·in Documentation
Karl F. Heumann ...................................... . 120
17. Computers, Communications and Science - Extending Man's Intellect
Saul Gorn
. . . . . . . . . . . . . . . . . . . . . . . . . .. 123
18. The Management Process and Science Information Systems
Edward M. McCormick
......................
131
BIBLIOGRAPHY
136
INDEX
148
5
Preface
THE PAPERS IN THIS VOLUME are adapted from
the proceedings of the Fourth Institute on
Information Storage and Retrieval presented
in February, 1962, by the Center for Technology and Administration, School of Government and Public Administration, The
American University.
One purpose of the Institute and these
papers is to present a broader systems approach to the handling of technical inforrriation. A complete systems approach must
include many elements sometimes neglected,
including the user and, particularly, the
manager .. The manager is responsible for the
input to science information centers and for
utilization of the output from these centers.
Management is therefore so intimately associated with the scientific information problem
that its role in the system must be studied
and rationalized.
With what level of management are we
concerned? It certainly includes the managers
of technical information activities, but it must
go beyond that level. Upper levels of management are often "unsold" on the importance
of the scientific information function. Although executives are providing increased
support to these activities, from their point
of view the science information function must
be judged and weighed in competition with
many other demands on the resources they
control. There is general reluctance to give
all the support that enthusiastic directors of
scientific information centers would like to
have. This function is not alone in making a
case for increased support; generally, there
isn't enough to meet all demands. Management must decide what can be support.ed, by
how much, and what must be cut back.
Implicit in this consideration of the management aspects of science information
centers is the assumption that technical infor-
6
mation is valuable, in that it is rather directly
related to a commercial product or is valuable
to our defense efforts. Further, it is assumed
that a substantial portion of the cost of the
item is the cost of the scientific effort involved
in its research and development. If this cost
itself is not significant, then we presume that
the value of the time is significantly dependent
on the contribution made by scientists. In
addition, we presume that the effectiveness
of scientists can be substantially enhanced by
more effective communication of scientific
information. Even in considering the cost of
scientific information we have much yet to
learn; in considering value we know even less.
Yet, managers are and must be quite cost and
value conscious.
Another aspect of the cost versus value of
scientific information is contained in the
objection expressed by managers that their
activities are considered to be expendable.
Whenever there is a cutback, the library is
the first to get a cut. This attitude is reflected
in the fact that the costs of such services are
generally taken up in "overhead" accounts.
Sometimes an even more suggestive term is
used synonymously with overhead, the "burden". This could be interpreted as a negative
attitude toward the value of technical information. When management is in a position
to consider technical information to be a
major resO'urce, rather than a burden - a
valuable resource it is willing and able to
allocate in a positive sense towards the goals
of the company - then we will have
made a significant advance in the scientific
information area.
In every organization, in addition to' information for the scientist, there is obviously
a substantial class of information which the
manager needs for his own use. Although
this managerial information is outside the
scope of this volume, there are areas of overlap. Often there are common approaches to
these two classes of information. Sometimes
the same organization, the same personnel,
and the same equipment are involved. Thus,
although we cannot and do not always want
to separate them, the emphasis is intended
to be on management aspects of science information activities as needed by the scientist,
not by management.
The chapters in this book tend to fall into
five subject categories. Each represents an
area of management interest.
The five sections are:
The General Problem. Leake, Hattery,
Heumann, and Gorn consider technical information in broad, general terms. They examine
various aspects of the historical, social and
cultural backgrounds of the problem, and the
general factors· significant to date and of
probable significance in the future.
The Management Problem. Cahn, Howerton and Hillier set the stage for the
management situation with respect to science
information centers. They consider various
aspects of the management problem and some
of their interrelations. In these chapters, the
presentation does not consider specific areas
in detail.
McCormick, in the final chapter, summarizes the material on information retrieval
management.
Communications Aspects. These four chapters, by Luhn, Brownson, Anderson and Kehl,
are directed toward an important aspect of
management, i.e., communications. They consider the facets of communications which are
peculiar to science information centers. Luhn
considers the general business intelligence
system. Brownson indicates the nature of the
scientist. Anderson studies the compatibility
of these new information services with other
information activities, and some problems
involved in integration of these services. Kehl
considers the rather specific problem of cus-
tomer communication with a digital computer
information retrieval system.
Problem Definition. This section of three
chapters provides specific app'roaches to certain areas of interest to management. Myatt
reports on a survey technique used to determine design information for an information
center. Newman presents criteria for the
economic justification of a science information center. Gull gives guidelines for the
decision to mechanize a technical information
activity.
Systems Experience. This set of three chapters, by Sherrod, Dennis, Asbury and Moise,
reflects the operating experience of one government and two industrial environments. In
each case, they consider in some detail the
situation in which each exists, the types of
services provided to the customers, and problems associated with the operation of the
centers. Dennis particularly emphasizes the
problems of financing an industrial technical
information center.
As is evident from· reading the papers, we
are in the developmental stage of anew dimension in science information. No seminar and
no collection of papers at this s.tage can define
and solve all the problems of information.
storage and retrieval and associated managerial problems.
Nevertheless, we are progressing in the
ability to meet the challenge. It is hoped that
this volume will contribute to a better understanding of the science information process
and provide guidelines for managers and
information specialists. Our objective is to
present points. of view and reports of experience. The state of the art does not support a
single philosophy, approach, or set of conclusions. Therefore, divergent points of view
and opinions are presented. We believe this is
appropriate to the stimulation of thought,
experimentation and progress.
-The Editors
7
The Contributors
HATTIE T. ANDERSON is the Librarian at
W. R. Grace and Company, Clarksville, Maryland.
Her professional background is in chemistry. Mrs.
Anderson has made a study of the relationship
between the technical library and the other technical information activities within the company.
W. S. ASBURY is Vice President of Esso Research and Engineering Company for the Legal
Patent and Information Division. He joined Esso
Research Laboratories in Baton Rouge in 1927
and served in various posts in the United States
and ,abroad. In 1947, he was appointed Vice
President and a member of the Board of Directors
of Esso Research and Engineering Compahy.
>
HELEN L. BROWNSON is Program Director
for Documentation Research, Office of Science
Information Service, National Science Foundation.
Mrs. Brownson has been in scientific information
activities at the National Science Foundation
since its inception. Prior to that, she was Secretary of the Special Committee on Technical
Information, Research and Development Board,
Department of Defense. She also served with the
Committee on Medical Research of the Office of
Scientific Research an~ Development.
JULIUS N. CAHN is Project Director of the
Subcommittee on Reorganization and Internal
Organizations of the Senate Committee on Governmental Operations. He has had many professional Senatorial assignments.
8
C. DAKE GULL is a Consulting Analyst, Information Processing Unit, General Electric Company, Bethesda, Maryland. Mr. Gull is a librarian
with extensive experience in mechanized information systems. His experience includes assignments
at the Library of Congress and with Documentation, Inc. He is a past president of the American
Documentation Institute (ADI) and is chairman
of the U. S. National Committee for the International Federation for Documentation (FID).
LOWELL H. HATTERY, Ph.D., is Director of the
Center for Technology and Administration, The
American University. The Center's activities include institutes on information storage and retrieval, research administration, and electronic
data processing. Professor Hattery also directs
academic programs in these areas in the School
of Government and Public Administration, The
American University.
KARL F. HEUMANN, Ph.D., is Director, Office
of Documentation, National Academy of Sciences.
An organic ohemist, Dr. Heumann's scientific
information experience includes Minnesota Mining and Manufacturing Company. He was Director of the Chemical Biological Coordination
Center, National Research Council, and was Director of Research for Chemical Abstracts. Dr.
Heumann is a past president of the American
Documentation Institute and is a vice president of
the International Federation for Documentation.
BERNARD K. DENNIS is Manager, Technical
Information Center, Flight Propulsion Division,
General Electric Company, Evendale, Ohio. IUs
professional training was in education. He has
been concerned since 1954 with technical communications, and has been Manager of the Technical Information Cep.ter since 1957.
JAMES HILLIER, Ph.D., is Vice President, RCA
Laboratories, Princeton, New Jersey. Dr. Hillier
is a physicist who was a pioneer in developing
the electron microscope. Since 1953, he has been
concerned with the management of research
activities. Dr. Hillier is also on the faculty of
Princeton' University.
SAUL GORN, Ph.D., is Associate Professor of
Electrical Engineering, Moore School of Electrical
Engineering, and Director, Office Computer Research and Development, University of Pennsylvania. His background is in mathematics. He is
an authority in the digital computing field. Dr.
Gorn is engaged in formal language design and
learning models.
PAUL W. HOWERTON is Vice President and
Director of the Communication Sciences Division
of Information for Industry, Inc., Washington,
D.C. Mr. Howerton is a chemist and a linguist.
His experience in organizing and managing large
information centers includes service as Deputy
Assistant Director for Central Reference of the
Cent;ral Intelligence Agency.
WILLIAM B. KEHL is Director of the Computation and Data Processing Center and Associate
Professor of Mathematics at the University of
Pittsburgh. Mr. Kehl has developed an information retrieval system using a digital computer.
This has been used for legal information retrieval
at the Health Law Center of the University of
Pittsburgh, and has also been applied to other
information problems in a university environment.
CHAUNCEY D. LEAKE, Ph.D., is Professor and
Assistant Dean of the College of Medicine, Ohio
State University. His professional background is
in pharmacology, but his interests in science are
manifold. He is past President of the American
Association for the Advancement of Science, and
was responsible for establishing the section on
Information and Communications Science in
the AAAS.
H. P. LUHN is a documentation consultant recently retired from the International Business
Machines Corporation. During his career at IBM,
Mr. Luhn had been engaged in design and development of special purpose machines for information storage and retrieval, and has been identified
with several significant innovations including
Keyword-in-Context (KWIC), Auto-Abstracting,
and Selective Dissemination of Information
(SDI) .
EDW ARD M. McCORMICK is with the Office of
Science Information Service, National Science
Foundation. He is engaged in activities concerned with the use of digital computers in
documentation systems. His experience includes
management of a digital computer and data proc-
essing activity for the Navy. Mr. McCormick's
background is in mathematics, and he has held
positions as an electronic engineer. He is the
author of Digital Computer Primer.
J. E. MOISE is Acting Manager, Technical Information Division, Esso Research and Engineering
Company, Linden, New Jersey. From 1933 to
1958 he was on the staff of Esso Research Laboratories in Baton Rouge. He was appointed
Assistant Director of the Laboratories in 1951
and served in that position until his transfer to
the Department of Refinery Liaison of the Esso
Research and Engineering Company in 1958.
DEWITT O. MYATT is President, Science Communications, Inc., Washington, D.C. Mr. Myatt
is a chemical engineer with experience in technical
information with Atlantic Research Corporation,
The American Chemical Society as Managing
Editor of Industrial and Engineering Chemistry,
and the Tennessee Valley Authority.
SIMON M. NEWMAN is a documentation consultant. He has served the U. S. Patent Office for
over 30 years. Trained as a chemist and as a
lawyer, Mr. Newman had considerable experience
as a patent examiner. He worked in the Research
Department of the Patent Office in developing
systems for storing and retrieving information
for use in patent examination.
JOHN SHERROD is Chief of the Science and
Technology Division of the Library of Congress.
Prior to joining the Library of Congress in 1952,
Mr. Sherrod's professional activities were in
meteorology at Pennsylvania State University.
9
By Lowell H. Hattery
The American University
1. The Systems Concept in Documentation 1
IN THE FIELD OF SCIENCE INFORMATION we are
faced with confusion in terminology. Librarianship, documentation, science information,
technical information, science communication
are all used to identify areas of concern which
may be identical, partially duplicative or exclusive depending on the user of the term
and the situation.
Not only does the obfuscative practice of
inexact and unstandardized terminology make
communication difficult - it also tends to
insulate against detailed probes illto the
nature of operational science communkation.
Yet, a direct look at situational objectives and
processes is needed sorely at the present juncture. The look must encompass the total of a
given situation. Implications to be drawn
from description of a science information center if the center is the only agent serving a
group of scientists are quite different from
those to be drawn if a library alsq exists to
serve independently or in concert.
Libraries and information centers are not
the only agents to serve the science information function. Nor are the needs and processes
the same from one science discipline to another, from one science problem to another,
from fundamental research to development
and testing, from one organization to another,
from one point in time to another, or from
one group of scientists to another.
Although there are many va,riables in
science communication, and many conditions
which may contribute to confusion, it is possible to de'velop an orderly approach to science
communication in any given situation. The
approach is based on the concept of science
communications, as a system in which information is produced by Researcher! and
transmitted to Researcher 2' (See Exhibit 1.)
The next step is the production of information from Researcher2 , which may be
10
transmitted to Researcher! among others, thus
closing the loop (Exhibit 2).
produces
communication
Researcher2
EXHIBIT 1
communication,
produces
communication
EXHIBIT 2
Complexity is added to this scheme when
all researchers are substituted for Researcher!
and Researcher2 • Not only are numbers added,
but communication must pass barriers of distance, discipline, proprietary interests, p'atent
rights, language and political boundaries.
Each of these' and other barriers is a matter of concern in the study of science
communication.
The barrier of language, for example, suggests such problems as translation, automatic
translation, research in syntax, international
exchange of literature, and the potential of
universal languages such as Interlingua.
Illustrations of the characteristics of the lan-
guage problem are shown in Exhibit 3. Yet this
diagram only be'gins to suggest the facets Qf
communication across the language barrier.
Who shall dO' the translatiO'n? HO'W will it be
financed? What media shO'uld be used for the
Qriginal publicatiO'n and for pUblication at
subsequent stages? If an abstract only is
translated, what arrangements are to' be made
for full text available to Researche,r2 ? What
is the loss in accuracy and subtlety in translatiO'n? What is the comparative lQSS in machine v. individual translatiO'n? Are competent translators available? If sO', at what CO'st
to mO're direct cQntributiO'n to' science? We
might spin out many more questiO'ns O'r
problems related to' language in science
communication.
MEDIA
In transmitting infQrmatiQn frQm scientist
to scientist, variQus media may be used as
illustrated in Exhibit 4. It is, QbviQUS that the
list Qf seven media is not cQmplete. The
medium Qf professiO'nal jQurnals itself can be
subclassified intO' several significantly different grQUps. As one does sO', many related
questiQns arise - questiO'ns of editQrial policy,
financing, distribution, publication time lag,
cQndensation, abstracts, reviews, etc.
Relatively little written infQrmatiQn gQes
direct frQm scientist to' scientist even through
the media just referred to'. For example, a
scientist cannQt read the many journals which
may be related to' his research interest - not
O'nly language barriers but the sheer volume
of jO'urnals in his own field and in related disciplines make it impossible. He must depend
heavily on intermediary repQsitories, where
infO'rmation can be stored and later retrieved
thrQugh classificatiO'n and index systems. The
nature and rQle Qf the repository varies accQrding to needs of the perSQn served, nature
of the informatiQn handled, media, available
means and Qther factors. (See Exhibit 5.)
Not only dQes the matter of repositories or
stQres of informatiO'n raise a large number O'f
questions about organizatiQn, media, financing, persO'nnel; acquisitiQns" physical facilities,
and disseminatiQn; it alsO' carries with it the
necessity for effective systems of classificatiQn and indexing. TraditiQnal library classification must be modified or supplemented to
meet special needs. Since the primary need is
information and only secQndarily a document,
indexing systems take on great significance.
As methods and equipment for searching indexes are developed, imprO'vement Qf indexing
systems becomes more important. Effective
infQrmatiO'n storage and retrieval depends
very largely on indexing systems and the
qu'alityof indexing.
THE SCIENTIST
Thus far the discussion has been directed
to what happens between researcher and
researcher - events frQm origin to' destinatiO'n. Yet of transcending importance is the
scientist himself - the user and producer Qf
informatiQn.
Researcher2
(German - German language
only)
Researcher!
(American - English
language only)
Information
I
Complete
Translation
into German
~
Manual
Auto-mechanical
Translation of abstract into German
Interlingua
EXHIBIT 3
11
As a prQducer, the scientist's thO.ughts.,
QbservatiQns, and findings must be placed into'
transmissible fQrm. It is interesting to' speculate. hO.W many useful ideas are nQt recQrded
and transmitted and hQW many potentially
useful experimental results never progress
beyond the mind O.f the scientist-observer or
beyond his undecipherable nO.tes.
Turning to' the user in the system, it is a
truism to' say that infQrmation is useless
unless used. "Availability" to the researcher
is nQt achieved unless it is brO.ught into the
scientific activity Qf the researcher in his
effQrts to' prQduce new infQrmatiQn. When,
how, and in what O.rder O.f importance is
infO.rmatiO.n utilized by different kinds of researchers O.n different kinds of problems in
different kinds of research situatiQns? Some
assumptiO.ns and some understanding O.f this
question are fundamental to the purpO.seful develO.pment of science cO.mmunicatiQns
systems.
ideas and infO.rmatiO.n from Qne research
scientist to' another. Science cO.mmunicatiO.n
is a single, internatiO.nal, multi-lingual, multidisciplinary, multi-process system.
Planning Qne element in the system, therefO.re, is likely to' affect anO.ther. A change in
translatiO.n programs may affect the rO.le Qf
libraries. A different mechanism fO.r searching indexes may require a modificatiQn of the
index. Changes in media affect prO.blems of
financing. The system cQncept in science CQmmunicatiO.n is particularly significant because
Qf the understandable practice O.f specialists
who. attempt to' sQlve the prQblems Qf an
element withQut reference to' the cQmplete
system.
LOCALIZED SYSTEMS
We have extended our cO.ncept O.f system to'
all science cO.mmunicatiQn. We may also. fO.CUS
it to' a labO.ratQry, Qr to' a cO.rporatiO.n, O.r to' a
Federal bureau. ThO.ugh the number O.f individuals invQlved may be reduced to' a few
thousand, a few hundred, Qr even a few dQzen,
much of the detail of a total system cQncept
remains relevant. The set O.f elements, prO.cesses and relatiQnships is sufficiently cQmplex
to' demand explicit enumeratiO.n and descriptiQn and sets O.f running rules, understandings,
and limits.
As fO.r any administrative study, objectives
THE CONCEPT OF THE SY8"l'EM
All O.f the diagrams shQwn might be placed
into' a single cQmplex scheme. The effect is to
shO.W the interrelatiO.nship O.f media, libraries,
stO.rage, retrieval, translatiQn, indexing, abstracting, electrO.nic sO.rting, financing, persO.nnel, and Qther cO.mponent elements to' the
simple O.rdering Qbjective O.f transmitting
Media of Communication
Researcher 1
Book
Monograph
Professional Journal
Notes
Information
Researcher 2
Photographs
Punched Cards
Oral
Etc.
EXHIBIT 4
12
and needs set the frame of reference for defining an optimum local science information
system. Once these are established, organization, methods, and techniques for achievement
are defined. The present resources are assessed. Present limitations and barriers are
identified. Desirable changes and additional
resources are specified. These steps are well
known and accepted in principle. They are not
so well known in practice.
This approach immediately establishes the
relative roles of libraries and science information centers, formal and informal programs,
media, machines, and people. The contribution
of a central library is put into perspective
along with the contribution of a special or
branch library.
An interesting approach to defining the
roles for libraries and science information
(or documentation) centers has been developed for Russian mechanical engineering
industries. 2 Within a "technical information
bureau" are: "the technological information
group; the technological library; the technical
office; and the system of technical information
representatives and correspondents in the
plant's departments."3
Processing the input to the technical information bureau is classified into: "library
processing, bibliographic processing or information processing".4 A detailed table of processing functions· is presented for books,
periodicals, and other documents. The processing actions for each major activity are as
follows :5
Libra'rY Processing
1. Inclusion in acquisition plans
2.
3.
4.
5.
Cataloging
Classification and filing of cards
Processing and shelving of publications
Notification of publications received
Bibliographic Processing
1. Bibliographic description
2. Filing of cards in reference and bibliographic files
3. Reproduction of cards
4. Notification of individuals
Information Processing
1. Distribution of literature and documentation at request of departments
2. Information notification
3. Preparation of reference cards and
filing in reference information file
4. Preparation of abstract
5. Making of technical translation
6. Microfilming and photoreproduction
7. Reproduction of material
8. Ordering of material needed
9. Assembling selections on particular
topics
10. Inclusion in work plan, amendment of
plan.
This illustration suggests an approach to
a definition of functions and assignment of
responsibility which is an essential first step
to the development of an orderly, complete
libra.ry-information program.
COSTS OF SCIENCE INFORMATION
For advance planning, budgetary controls
and evaluations, cost estimates of the science
Intermediary Repositories
Researcherl
General library
Specia I library
Science or technical
information center
Information
Laboratory files
Researcher 2
Personal files
Tape for
electronic processing
EXHIBIT 5
13
information system are needed. Cost analysis
should include resources directed to informal
as well as formal processes. Cost evaluation
should take into account the penalties attached to lack of information as well as the
values of the information program provided.
As an increasing number of persons point
out, trends in expenditures for science information services run far behind other support
for research and development activities. It is
probable that one reason for this disparity is
that the need for experimental equipment
such as a new laboratory instrument is more
readily demonstrable than the need for information. This situation makes better cost
analysis of the science information function
more important.
.
RESPONSIBILITY FOR
SCIENCE INFORMATION
There are several factors identified or implied in the preceding discussion which relate
to the location of responsibility for science
information, planning, and operations. Scientist producers and users continue to be key
elements in the system. But the complexity of
the system, the extension of information
processes beyond those of the traditional
library, an extraordinary' concern with a
multitude of external relationships, the availability of new machines and techniques for
information storage and retrieval, and an
ingreasing tendency to develop and define
non-science information systems within organizations, all tend to push upward prime
responsibility for the science information
system.
Precisely where this responsibility should
lie depends. on many special conditions. The
slow but inevitable rise in status of the administrative information function has led to
the designation of a vice president or the
equivalent for internal information systems.
Although this arrangement does not ordinarily
include the science information system, it does
include information processing machines and
personnel essential to some aspects of a
modern, effective science information system.
It includes the systems analysts who can contribute to definition, plans, and operations for
the science information system.
Substantive planning and operational control, however, are properly assigned in part
to professional personnel associated with the
14
scientific and technical program, reporting
typically to a vice president for research and
development.
It is a short step to recognition that the
science information problem requires the
joint contribution of professional knowledge,
understanding and skills of the scientist, the
systems analyst, the librarian, and the information or data processing specialist. This
combination forms the basis for defining a
new professional field of science information
(or documentation). Several universities are
developing curricula designed to train science
information specialists. to meet modern needs
of documentation. The American Institute of
Biological Sciences has recently announced a
program, supported by the National Institutes
of Health, to train science information specialists in biology in a cooperative arrangement with The American University. Selected
trainees undertake internship projects under
the direction of the Biological Sciences Communications Project of the AIBS and enroll
for Master of Arts or Doctor of Philosophy
degrees in Public Administration at The
American University with special emphasis
on documentation and related fields. Several
schools of library science are extending their
curricula to include special training in science
information and data processing.
SUMMARY
The science information function can be
conceived and designed as a system. The
nature of the current science information
problem calls for high level organizational
attention. The planning and operation of a
science operation system and many of its
elements calls for a combination of knowledge,
understanding and skills which is bringing a
new profession into focus - the science information specialist or documentalist.
REFERENCES
Exhibits 1 - 5 and some of the descriptive material
are taken from Lowell H. Hattery, "Information
and Communication in Biological Science A
Report Prepared for the Biological Sciences Communication Project," American Institute of Biological Sciences, December, 1961.
2 Aram S. Melik-Shakhnasarov, Technical Information
in the U.S.S.R. Translated from Russian by Boris
Gorokhoff. Massachusetts Institute of Technology
Libraries, Cambridge, Mass., 1961. 122 pp.
3 Ibid., p. 36.
4 Ibid., p. 25.
S Ibid., pp. 26-35.
1
By Chauncey D. Leake
The Ohio State University
2. What Must Give in the Documentation Crisis?
SOMETHING HAS TO GIVE in our current crisis
in documentation. Like an overblown balloon,
science is ready to burst its conventional
bounds. If something does not give, we are
going to be drowned in the flood of our
scientific knowledge.
A moratorium on the publication of scientific findings has been proposed. This suggestion comes snideIy from uncomprehending
humanists who claim to want time for people
generally to be able to absorb what science is
all about. This is not a satisfactory answer to
the problem. We must continue the prompt
publication of the results of scientific studies
in order to accumulate that verifiable information about ourselves and our environment
which is essential for wise policy decisions,
whether private or public. Clearly the conventional methods of documentation are not adequate to the task of handling the current flood
of scientific information. What gives?
The bare statistics, whether for scientific
periodicals, for abstracts, or for indexes, are
discouraging in their mountainous size.
Thanks to various mechanizing devices, we
are beginning to see a little hope of being able
to handle the necessary indexing of what is
being reported. However, in the publication
of original scientific information, and in its
abstracting and reviewing, we are floundering
and desperately wondering what we may be
able to do about it.
PURPOSES IN DOCUMENTATION
In order to keep ourselves oriented in our
documentation problem, let us consider what
it is we are trying to do. To begin, I think we
can all agree that verifiable knowledge about
ourselves and our environment is essential
for us so that we may make satisfactory public policy decisions on ways and means of
maintaining or increasing our individual and
social satisfactions. John Locke (1632-1704)
made this clear in his consideration of the
psychology of democracy. His point was extended in our country by Benjamin Franklin's
insistence on free public libraries. It was
extended further by Thomas Jefferson's insistence on free public schools. In both
instances, the purpose was to make it possible
for people in every community to have access
to the same body of "sound information," as
Locke called it, and thus be prepared to make
wise public policy decisions. On this basis,
scientific information, the best of "sound
information," must continually be available
for our people.
The main function of science documentation, therefore, would seem to be to arrange
ways and means by which scientific information can best be collected, preserved, analyzed,
indexed, abstracted, reviewed, translated, and
interpreted, for two purposes: (1) that scientists generally might know what is going on,
and (2) that people generally might understand something of what science is about, and
what it may mean to them in comfort, health,
convenience, and satisfaction.
The collection and preservation of scientific information is the basic responsibility
of libraries. Libraries also have the taxonomic
job of classifying it, analyzing it, and indexing it, so that it may readily be found when
needed. Is this enough?
The collecting, classifying, indexing, and
retrieving of scientific information might
seem to be a simple clerical task. In order to
be useful, however, whether to scientists or
to people generally, the ever growing and
changing scientific information must be analyzed so that there may be an effective taxonomic approach for satisfactory indexing.
It is unfair to think of librarians in scientific libraries as being merely clerks. Skilled
15
librarians in good science libraries are really
part of the scientific team. They participate in
the group research endeavor by anticipating
what already published scientific material may
be helpful to the scientists who are at the
bench. The reference function of a good
science library is an extremely important one
in modern scientific effort.
. It is always interesting to watch capable
scientists work in a library. More and more it
is becoming necessary to have the current
scientific material right out on shelves where
it can readily be found. This is the prime
function of the current periodical room, which
is increasing in size every year. The rapid
scanning of periodicals by individual scientists, even if they are well acquainted with
their respective fields, is an increasingly time
consuming and difficult task. Such a device
as Current Contents gives an opportunity for
quick weekly scanning of major scientific
journals as they are published, so that scientists can title-scan the current reports as they
appear. They may then quickly note what it is
that they may wish to examine.
Increasingly, libraries must make readily
available the large and growing indexes to
current scientific information, so that scientists may be able to find quickly what specific
references they may wish. The library card
catalog requires continual revision and analysis, with cross reference, in order to aid in
quick retrieval.
No matter how carefully a library may
maintain the current periodical room and
index references, there may still be frustration for the scientists who do not know
exactly what it is they wish to find. This is
why it is becoming increasingly important
that current monographs and reviews, particularly those for the last ten years or so, be
kept on open shelves where browsing scientists may see the title that will give them the
idea of where they may find what they seek.
The essential purposes in science documentation are to handle the flood of scientific
publications in such a way that specific factual
data may promptly be found, as well as analyses of their significance. Adequate science
documentation also should include syntheses
of scientific trends, with indications of relations to other factual data, or to significant
applications.
16
INDEXING AND ABSTRACTING
Indexing always seems to be a compromise.
Most scientific publications contain much
more significant information than can ever
be indicated by a title. Titles are often deceptive. If an effort is made to be precise in a
title, it may become too long. Then it becomes
difficult to index satisfactorily, especially if
adequate cross reference is to be given .
Nevertheless, very helpful mechanical devices are being developed for useful indexing,
and these are being utilized by Chemical
Abstracts, Biological Abstracts and Index
M edicus. These are meeting many of the objections to the older hand labor indexes. At
least they can appear contemporaneously with
the material that is being indexed. Are these
mechanical devices to be adequate as scientific
reporting doubles in another decade?
Chemical indexes have a definite taxonomic
advantage over indexes to biological and medical reports. Chemistry has an exact system of
nomenclature, precisely organized so as to be
able to handle any conceivable chemical compound. It is taxonomically stable, even with
the introduction of new chemical concepts.
On the other hand, biological and medical
terminology is subject to great change. This
occurs with improved etiological understanding of disease, and with new insights regarding the relationships of various kinds of
living material. This causes many taxonomic
difficulties. With inevitable changes in terminology, the systems of indexing for biology
and medicine can have little stability. Much
confusion results on attempted retrieval of
past information, which may be indexed under
different terms than those which are used
now. Much potentially valuable scientific
information may thus get lost.
There is much speculation about the possibility of computers being used for medical
diagnosis. This will be practically impossible
until there is a stable taxonomy for disease,
and this may be a long way in the future. The
best that could be done now, even with our
most advanced concepts and mechanization,
would be to give the probabilities that a particular set of symptoms observed in a patient
might be diagnostically indicative of a specific
disease concept. Information on the key symptom for an accurate diagnosis, however, might
be lacking completely.
It is particularly important in effective
retrieval of scientific information to study
indexing in depth. This has been well attempted by Dr. Isaac Welt, first under the
auspices of the National Research Council,
and now working through the Institute of
Medical Communication. Indexing in depth
is applicable only to relatively circumscribed
areas of scientific interest, such as cancer, or
drugs acting on the cardiovascular system.
Indexing in depth involves the careful
reading of original articles and noting whatever scientific information or data of possible
significance is included, and making cross
references so that this information can be
retrieved. For example, in compiling an index
in depth on the effects of drugs on the cardiovascular system, one may list all drugs in the
literature examined against the biological
effects produced by these drugs. One can also
list various biological effects which are possible, and then place against such a list the
various drugs which produce these actions.
There is thus provided a cross-index of great
power and usefulness. The bulk of this sort
of information is not to be derived from the
titles of the articles, so ordinary indexing by
title would not give the information that is
available.
Many of these problems of indexing were
recognized a couple of decades ago in what is
now the National Library of Medicine. Under
the leadership of the late Colonel Wellington
Jones, General Joseph McNinch, and now Dr.
Frank B. Rogers, the National Library of
Medicine has exhaustively studied the indexing problem and set up a highly efficient
mechanized system for compiling the Index
Medicas. Nevertheless, this is under strain.
A computer center system is being devised. It
is called MEDLARS (not the dictionary's
"crabapple" but an acronymn for "Medical
Literature Analysis and Retrieval System").
MEDLARS is scheduled to begin operation
in another year or so. With it, Index Medicas,
which now indexes 130,000 articles a year
from 1,800 periodicals, should be able within
a few years to handle 250,000 articles annually from 6,000 journals. Subject headings
which now run about 5,000 can be increased
to 12,000. The articles will have to be coded
by human analysts, but this coded information can be placed in computer storage, so
that on command the computer can print out
a compendium of whatever set of entries are
desired in a format ready for publication.
This can compile bibliographies by subject.
Dr. Rogers is quite specific about what
MEDLARS can do: "It is not a system for
picking out and reproducing the articles
themselves . . . Something else is called for
which no computer can supply. This is intelligence ' - human intelligence. MEDLARS
will respond to questions which are asked as
precisely as possible in terms of the system.
Frankly, many people don't know how to ask
questions at all. A person who walks in and
asks for a bibliography on diabetes will have
to decide exactly what he wants to know about
diabetes. He may even have to define what he
means by diabetes. If he knows what he
wants, we can help him."
Here is the scientific situation in a neat
capsule: Ask the right kind of questions, and
scientific effort may be able to provide the
answers. It is the appropriate kind of questions that we need to bring up for consideration in respect to indexing, abstracting, and
other aspects of documentation. When we ask
these questions in an appropriate manner, we
may be able to get fitting answers that may
help solve our dilemmas.
Abstracting is an art. It is particularly
well developed in chemistry, medicine and
psychology. It is dependent entirely upon
human judgment, usually the judgment of
experts in the particular field that is being
abstracted.
The flood of articles worthy of being abstracted is, however, becoming so vast that
the system shows signs of cracking. There are
many practical and theoretical reasons for
looking carefully into the whole proposition
of abstracting scientific publications. The
abstracting process may break down in practice because of sheer bulk; theoretically, it
may be questioned whether or not even a
"good" abstract is much better than the title
of the article in a "good" index. Frequently
the original articles must be consulted by the
scientists who need specific information, unless the indexing has been in depth.
Abstracts may be unsatisfactory because
they are too brief and do not give all the essential information, or because they are as
detailed as the original article itself. Quite
often one will find an abstract in Chemical
17
Abstracts which will run to a column and a
half of fine print, when the original article
its.elf may only be a couple of pages..
Abs.tracts thems.elves. mus.t be indexed if
they are to be us.eful, and this. is. a problem
equally as. difficult as indexing original articles. There is. another trouble: an index to
abs.tracts. is. by no means. a comprehens.ive
index of the field. Abstracts. are s.elective. For
s.ome reas.on or other, many original contributions. may not be abstracted anywhere.
Worthines.s for abs.tracting is. us.ually an
individual judgment: who may s.ay when
s.ome other individual may find s.omething
quite worthwhile in the article that has. not
been abs.tracted?
N evertheles.s., abs.tract journals. do have one
great function: they do bring together at one
time and place the important contemporaneous
contributions. that are being made in any
particular s.cientific field. They thus. can s.ave
a great deal of time in s.canning. Certainly
the great s.ucces.s. of Chemical Abstracts, Biological Abstracts, Excerpta Medica, and Psychological Abstracts testifies. eloquently to
their us.efulnes.s. and need. Thos.e who have
been respons.ible for developing thes.e great
abstract s.ervices deserve the abiding thanks
of all scientis.ts..
An exceptionally well organized and s.tandardized abs.tracting s.ys.tem, with effective
cros.s.-indexing, has. been developed by the
Federation of American Societies. for Experimental Biology, under the direction of Milton
O. Lee. Nearly 3,000 abs.tracts. for the great
annual meeting are assembled, clas.s.ified, indexed, and dis.tributed within 10 weeks. time.
The res.ultingpublication runs well over 500
pages of double column offset, but it serves
as an indis.pensable adjunct to the meeting
program (of 212 pages itself!), which is classified by titles and cross references to the
abstracts.
One might think from the increasing excellence of our s.dentific indexing and abstracting services that we are well on the way to
solving the documentation cris.is.. This. is by
no means. the case. The bes.t possible indexing
and abstracting of scientific literature is. not
enough to make this verifiable information
about ourselves. and our environment available for appropriate application in bringing
us the s.atis.faction we want. No matter how
18
s.killfully indexed and abs.tracted scientific
information may be, it still has. to be analyzed,
diges.ted, correlated, and prepared in s.uch a
way that it can become the basis. for wis.e
individual public policy decis.ions..
REVIEWS
From my own pers.onal experience in trying
to handle scientific information, I am becoming more and more convinced that we need
to put particular effort and emphasis. on critical reviewing. I have had experience as a
librarian of medical s.chool libraries., as. an
abstracter, and as a bench and working s.cientis.t. Over the cours.e of many years., I have
found that I can orient mys.elf best into 'a field
of s.cience in which I am working, or am
expecting to work, by reading carefully s.uch
critical reviews. on the s.ubject as. may be available. I've had experience in attempting critical
reviews. als.o, and I can tes.tify directly to the
very s.trenuous. effort involved.
We have critical reviews. in s.uch fields. as
phys.ics., chemis.try, biology, phys.iology, and
pharmacology. Many of our better medical
journals. also include carefully analyzed reviews. on practical medical problems.. Many
of thes.e reviews. become clas.s.ic s.tarting posts
for the most effective kind of s.dentific endeavor. We need many more critical reviews.
It is. interesting that Annual Reviews is
aware of the necessity of critical analysis and
correlation. It is not enough merely to list
each year the important contributions that the
reviewer thinks may have been made in a
particular subject. Annual reviews. in any
subject must he critical, analytical and correlative, in order to be helpful. The great success of Annual Reviews, and of other special
review journals., again is direct testimony to
the long range value and importance of this
type of documentation activity.
My own feeling is that the s.dentific research team is incomplete without a competent
library scientist. Such a person s.hould have
equal status with the bench and working
scientists. Such a scientist requires special
training and background.
The library scientist should be familiar
with basic scientific prindples and concepts.
Library scientists s.hould know how to judge
the s.cientific validity of s.cientific reports..
They s.hould know how to apprais.e data, and
they should have wise judgment in regard to
the soundness of conclusions drawn from
scientific data. Furthermore, they should be
able to show significant correlations between
individual scientific reports, so that trends
and developing concepts may promptly be
recognized and utilized.
Library scientists would need much training in judgment. This comes best from
intensive study in the humanities, including
languages, literature, and history, including
the history of the fine arts as well as the history of science. They should have a broad
acquaintance with general philosophical concepts. Furthermore, library scientists should
be able to write easily and succinctly.
It is clear from this outline of the appro-
priate qualifications for the library scientist
in a research endeavor that the position is a
challenging and worthy one, and that it deserves full status, salary and recognition. If
library scientists really do have the kind of
judgment which I have indicated, they might
well become the scientific reporters for the
scientific research effort. Some improvement
thus might follow in the literary quality ()-/'
our scientific publications.
If satisfactory reviewing of scientific contributions can be developed, it might diminish
the pressure to extend our abstra.cting services. ActuaUy, I think our abstracting services
would be wise to include review articles as a
regular feature of their respective periodical
publications.
Library scientists could well become those
who are particularly familiar with scientific
literature and the ways and means by which
it may most effectively be handled. It would
be their responsibility to tell the working
scientists in the group what information may
be pertinent to the problem at .hand, and what
actual data may already be available on it.
It is well recognized that teamwork has
become an essential aspect of our scientific
research endeavor. It is not likely any more
that any single individual will be able to
make a major scientific contribution alone.
Certainly under economic and national pressures, the teamwork approach will expand,
and it will be necessary for the scientific research team to meet together, perhaps daily,
to go over the work as it proceeds, to coordinate it, and to continue to supply that
enthusiasm and curiosity that is essential for
success in scientific effort. In this kind of
increasing teamwork in scientific research,
the library scientists may become among the
most valued and respected members of the
team.
If library scientists have the appropriate
judgment for correlating isolated bits of
scientific information and determining trends
and significance, they may be well qualified to
act as interpreters of scientific advance for
the people at large. Library scientists could
well serve as the point of contact with professional science writers, so that all aspects of
the scientific research program could be presented in such a way that intelligent people
could understand what it's about.
I think there is a great and challenging
opportunity for library scientists in our future scientific advance. If we can successfully
develop such a group of qualified workers, in
the increasingly necessary division of· labor
in scientific research, we may provide a
background for aiding materially in our documentation crisis. Critical reviews can certainly help. They can point clearly to the
status of the scientific development in any
particular field at that particular moment.
They can indicate the gaps to be filled. They
can show the inadequacies of previous approaches. They can stimulate to fulfill the
promise of indicated trends.
IN .PROSPECT
In his interesting Science Since Babylon,
Derek de Solla Price has a chapter devoted to
"Diseases of Science." The most significant
disease of modern science seems to be an
exponential increase in number of publications. This is reflected in the logarithmic
growth of scientific journals, which may approach 1,000,000 in another half century. The
data so vividly arranged by Doctor Price are
frightening.
Many proposals. have been made about collecting scientific data in regional repositories,
with brief abstracts appearing in scientific
journals, so that scientists may know what
information is accumulating. This device,
even though it might cut down the number of
publications in scientific literature, would
still require effective indexing and reviewing
in order to be useful.
19
Whatever device may develop to reduce the
number of scientific publications would be a
welcome respite, provided that we could be
certain that the scientific information accumulating would not be lost.
All indications point to the increasing significance of library scientists in handling our
documentation crisis. It may be that the crisis
will yield to the development of an effective
technique for critical reviewing by specially
trained library scientists. Library scientists
may be increasingly expected to become a
significant part of research team effort. Library scientists might just as well be trained
and encouraged to undertake critical reviewing, so that the general status of an area of
science can quickly be estimated.
Here, as in any other aspect of science,
much depends on the responsibility which can
be developed by those who are concerned with
any phase of handling scientific literature.
Library scientists, of the type I have been
advocating, would have a heavy burden of
responsibility in accurate, non-biased, and
comprehensive appraisal of the scientific information accumulating in their fields of
interest. Further, the working scientists
would have to repose a great deal of trust and
confidence in library scientists, in order that
the mutual division of labor could proceed
smoothly. This all takes a high level of individual self-discipline oriented toward the
collective social good.
Various aspects of automation will aid in
the documentation crisis. We will be certain
to get more satisfactory and comprehensive
indexes. We will develop indexing in depth,
20
so that the conventional abstracting service
may again be challenged. The current crisis
in scientific documentation may give qQite a
bit, without bursting, if we can develop a
successful method of critical reviewing, with
competent review scientists, who would have
status, and in whose efforts both working
scientists and the general public would have
confidence.
With international concern over the crisis
in science documentation there is much talk
and some progress. A well attended international congress on scientific information was
held in Washington, D.C. in November, 1958,
and a huge volume of discussion resulted.
Jesse H. Shera is editing for International
Publishers of New York and London an important serial on Advances in Documentation
and Library Science. H. P. Yockey edited a
helpful Sympo'sium on Information Theory in
Biology (Pergamon Press, London, 1958, 418
pp.). Communication theory is developing
with elaborate mathematical analyses, and
cybernetics is an established scientific discipline with vast applications in computer
technique. The American Association for the
Advancement of Science established at its
Denver 1961 meeting a Section on Scientific
Information and Communication. Very significantly, the Committee on Government
Operations of the United States Senate has a
keen staff studying and reporting on documentation, indexing, and retrieval of scientific information. With these many pronged
attacks upon it, it would seem that there is
certain to be some yielding in our science
documentation crisis.
By Julius N. Cahn
Staff, U. S. Senate Committee on
Government Operations
3. A System of Information Systems
RECENTLY THERE HAS BEEN increasing discussiO'n throughout the natiO'n O'n the subject O'f
national goals. More and mO're O'bservers have
exp'ressed the view that, in this perilO'us age,
the United States cannO't afford to drift in any
phase O'f science and technolO'gy. These observers have suggested that we need to select our
national O'bjectives for the years ahead. Only
then, they cO'ntend, can we:
(a) establish the priorities O'f the respective goals,
(b) set a time-table for accO'mplishment,
(c) allocate our limited 'f'esources accordingly, and
(d) monitor our progress.
It is the purpose O'f this statement to urge
that a natiO'nal goal be established fO'r a Sys~
tern O'f InformatiO'n Systems. This goal WO'uld
consist O'f two complementary objectives:
(1) A Federal System of InfO'rmatiO'n Systems linking the information activities O'f
agencies O'f the United States government.
(2) A national System of InformatiO'n
Systems linking, to' the extent which they may
voluntarily regard as feasible and desirable,
as many O'f the non-governmental infO'rmatiO'n
resources as possible.
Attainment of both objectives requires a
greater degree of voluntary collaboration than
has ever before been achieved. This includes
collaboration within government, among governmental and private groups, and among
private grO'ups themselves. Both objectives
can be fulfilled; indeed, they must be fulfilled,
if other natiO'nal goals are to be achieved.
AN ORGANIC UNITY OF
INTERRELATED PARTS
But fulfillment requires, first, understanding. A system, as it is used in this regard,
implies an organic unity of inter-dependent
operations. It implies effective O'rganization
for optimal input and output; a IO'gical flow
of work, step by step; feedback from part to
part; fulfillment O'f separate purposes by each
part and of common purpose by all parts.
A System of Information Systems is required because informatiO'n is, in effect, a
natiO'nal resO'urce. The resO'urce can be maximized O'nly if it can be made more than a
jumble O'f fragments. InfO'rmatiO'n must be
assembled fO'r the goO'd O'f the republic. (Certainly, enemies are determined to'· use O'ur
pool of infO'rmatiO'n for the detriment O'f this
nation; we can hardly affO'rd to' ignore their
diligent use O'r to' under-O'rganize O'r underutilize our own infO'rmation.)
All types O'f infO'rmatiO'n in O'ur natiO'n cannO't be generally utilized; sO'me infO'rmatiO'n
must necessarily be withheld O'n national
security or prO'prietary grO'unds. But taxsUPPO'rted, nO'n-classified, nO'n-prO'prietary informatiO'n is the prO'perty O'f the American
peO'ple.
A POSSIBLE INFORMATION NETWORK
As a substitute fO'r the cO'ncept of a "System
O'f Systems," some O'bservers have suggested
different analO'gies. SO'me have used the cO'ncept O'f an "Information NetwO'rk". It WO'uld
cO'nsist O'f a series O'f "stations", each O'f which
WO'uld serve a particular audience, but which
WO'uld transmit over the equivalent O'f "cO'axial
lines" O'r "microwave" the best O'f centrally
prepared material as well. Still other observers have used the cO'ncept of a "central infO'rmatiO'n switchbO'ard", such as serves independent telephO'ne cO'mpanies, O'r O'f a "grid"
such as electric utilities prO'vide, feeding
PO'wer from O'ne IO'catiO'n to anO'ther at times
of peak loads. Whatever the analO'gy, an urge
21
for togetherness can be observed, a will to
explore the possibility of common answers to
common problems.
The time appears ripe for a bold plan to
fulfill the urge, to make a System of Systems
a living reality. The times offer many "signs
and portents" that an historic opportunity is
now available.
What are these "signs and portents"? Some
are material - strong actions taken by a
variety of Federal, professional, commercial
and other organizations.
Leadership - the indispensable prerequisite for action - is being demonstrated in
deeds.
Much of the evidence, however, consists
merely of comments by key individuals, indicating that they are in a mood for cooperative
action. Many leaders of scientific and engineering societies, for example, now seem to
recognize that present and foreseeable circumstances have made information needs
acute; remedial programs which they might
have viewed with quick misgivings only a
few years ago would apparently now receive
more sympathetic consideration.
Such recognition is not universal and it is
not uniform; but it does exist in many places.
Here are a few of the straws in the wind:
INTEREST AT HIGHEST
FEDERAL POLICY LEVEL
The most important single sign is the deep
interest which has been manifest at the highest policy level within the executive branch of
the United States government.
The Office of the Special Assistant to the
President for Science and Technology has
been giving intensive attention to the information problem. The Special Assistant, Dr.
Jerome Wiesner, has empaneled an expert
group on science information in the President's Science Advisory Committee, under
the Chairmanship of Dr. Alvin Weinberg,
Director of the Oak Ridge National
Laboratory.
THRUST FROM FEDERAL RESEARCH
AND DEVELOPMENT OUTLAY
The thrust of scientific and engineering
research and development on the part of the
United States government has inevitably
brought about a quickening of interest in
information problems.
22
In the 1963 fiscal year, it is estimated that
Federal agencies will request authorization
for $12.3 billion of expenditures in research,
development, testing and evaluation. 1
Of this vast sum, $7.1 billion are being
requested by the Department of Defense for
military functions, $2.1 billion by the National Aeronautics and Space Administration,
$1.1 billion by the Atomic Energy Commission. 2
For medical and health related research
and research facilities of all types, $1 billion
are being requested in the 1963 fiscal year,
including $756 million for the Department of
Health, Education, and Welfare. Of this latter sum, $679 million are being requested for
the National Institutes of Health. 3
Perhaps the most dramatic thrust derives
from the nation's space effort. President Kennedy had proposed, and the Congress had
approved as a national goal, a ten year program of expeditions and experiments in outer
space.
Over $5.4 billion in new obligational authority are being requested for Federal space
programs in the 1963 fiscal year. Of this sum,
N.A.S.A. would receive $3.7 billion, D.O.D.
$1.5 billion, A.E.C. $193 million, the Weather
Bureau in the Department of Commerce $47
million, and the National Science Foundation
$1.7 million. 4
N.A.S.A.'s information needs probably
represent the most urgent, the most comprehensive, the most complex such needs of any
civilian agency in American history. Timely,
efficient achievement of N.A.S.A.'s 10 year
program will, in my personal judgment, require a virtual revolution in the processes of
storage, retrieval and dissemination of scientific and technical information. The implementation of N.A.S.A.'s blueprints for its
information program will have a profound
impact- on other governmental agencies, as
well as on non-governmental organizations.
In turn, the effectiveness of N.A.S.A.'s programs depends upon the fullest cooperation
of information resources outside the agency.
IDENTIFICATION OF PART OF
INFORMATION EXPENDITURES
A third sign is the Federal government's
identification, for the first time, of at least
part of the total of government~wide expendi-
tures for scientific and technical information.
It is estimated that for the 1962 fiscal year,
at least $98.6 million will be obligated for
Federal support of certain scientific and technical information programs. This estimate
includes what the National Science Found':ltion regards as "reasonably sound"
determination of obligations for intra-mural
performance.
But, N.S.F. carefully notes, the estimated
data relating to extra-mural scientific and
technical information "exclude.[sJ obligations
for such informational activities performed
as a supplemental service under a contract or
grant primarily for research and development
or scientific education and training."
N.S.F. states, too, that "Authorities in the
scientific information field have estimated that
this amount may exceed $100 million for fiscal
year 1962."5
Now that these first estimates have been
made, we may expect increased legislative and
executive branch attention to the efficiency
and economy of in-house and extra-mural
information outlay. Particular attention is
likely to be given to that vast portion of the
latter area of spending for which accounting
is not now available.
MECHANIZATION OF
INFORMATION SERVICE
The trend toward mechanization continues
throughout major Federal agencies. This
trend makes possible, indeed it renders essential, efforts to assure reasonable compatibility
betweel1 the machine s.ystems.
Each of the three great national libraries
has mechanization very much in mind. The
National Library of Medicine is scheduled to
bring into operation two years hence its precedent-making "MEDLARS" (Medical Literature Automatic Retrieval Service.)
The Library of the United States Department of Agriculture will be part of a Task
Force on Automation to consider the problems
of all of U.S.D.A.'s documentation. The task
force may work under the auspices of the
Office of Management Appraisal and Systems
Development.
The Library of Congress has had the benefit
of the services of a team of experts in computer technology, data processing systems,
operations analysis and information storage
and retrieval, through a grant from the
Council on Library Resources.
To this list could be added similar studies,
surveys and plans for new or strengthened
mechanization on the part of other Federal
agencies. Federal officials' are naturally noting
similar trends abroad, particularly on the part
of Soviet scientific and engineering installations, including activities in machine translation. 6 The pace in the international information race continues to accelerate.
DIRECT CALLS FOR COORDINATION
The clearest evidence of what may be the
"shape of things. to come" may be seen in the
direct calls for more coordinated governmental and Federal-private programs.
Lt. General Arthur Trudeau, Chief of Research and Development for the Department
of the Army has., for example, called for: 1
"S.A.T.LC. - a Scientific and Technological Information Center - a national clearing
house for the most complete and comprehensive acquisition, translation and exchange of
information that we can get from all segments
of our nation and from the rest of the free
world."
An engineering journal has stated: 8
"Above all, in this country we need a cen~
tral agency to coordinate the overall information effort; winnow the total production; and
pick up the· loose ends like foreign publications, research bulletins, graduate theses,
government reports and papers delivered at
conferences. Organizing and operating such
an institution would be an effort no bigger
than others we have made, no more ponderous
than others the government has shouldered
and carried home. It would deserve to be
calIed investment rather than extravagance."
It should be recalIed that four years ago,
experts at the School of Library Science,
Western Reserve University, Jesse H. Shera,
AlIen Kent and James W. Perry, had proposed
that :9
" ... a national center for the coordination
of scientific and technical information be
created, which will achieve an effective balance among centralized processing of the
world's published literature, cooperative
activity with other processing centers, and
provision of direct or indirect services to
23
individuals, research groups, or specialized
information centers."
For additional and more recent evidence,
we may turn to a specialized area of communication - within medical science. Charles
E. Lyght, M.D., noted an appeal by Senator
Hubert H. Humphrey for improved information management and stated :10
"Sentiments of this kind deserve applause.
Any knowledgeable medical research worker
knows how slow and faulty communication
lines can be at embarrassingly crucia
moments . . .
"Perhaps the best response is to encourage
Senator Humphrey and his colleagues to survey the problem immediately, thoroughly,
and objectively. Then they might consider
approaching the host of independent information-gathering facilities throughout the
country with an invitation to pool as much of
their know-how as feasible and to pour as
much of their own data as permissible into a
flexible collaborative storehouse and exchange
system. Such cooperation could prove mutually rewarding and would be justified as a
wise investment of both public and private
funds.
"Through a series of compartmented conferenc~s sponsored by the government, the
needs of biomedical researchers, medical
writers, editors, abstracters, librarians, electronic specialists, and others could be determined and their invaluable experience
utilized toward converting 'medical communications' into more than a high-sounding
phrase."
There is by no means agreement on the
modus operandi for achieving increased coordination. Particularly moot is the issue of
what the ultimate Federal role should be.
Many observers express deep concern lest a
monolithic Federal information operation be
attempted. But the urgent need for increased
Federal and for public-private teamwork is
now conceded virtually everywhere.
POLICYMAKERS' INCREASED
UNDERST ANDING
Within the administrative echelons of government, some policy makers now recognize
the varied nature of scientific and technical
information and its overall significance.
24
They appreciate that comprehensive, up-tothe-minute information of all types is crucial
for administrative decision making, for budgeting, and for bench purposes.
Many recognize, too, that the inter-agency
nature of many Federal research and development programs requires pooling of
inter-agency administrative and scientific information to a greater.extent than heretofore.
It is probable that these informed policy
makers are still in the minority. Most officials
still tend to view information in a less significant light and largely in an intra-agency
context. They see it as an "auxiliary" service,
as a part of "overhead". As such, the operation of a library or an information center
may, if necessary, be curtailed - relatively
casually like electric utility or other
"house-keeping" service. Officials with this
view do not yet regard information as an
integral part of research and development.
This view of many administrators is not
shared by an increasing number of scientists
and engineers. They, by contrast, are articulating a more and more insistent demand for
improved Federal and non-Federal information services. The demand is far from universal; it is far from uniform; but it is rising.
RECOGNITION OF NEED FOR
NATIONAL REFERRAL SERVICE
The need for a national referral service is
increasingly recognized. Fortunately, the need
has long been recognized by the National
Science Foundation, Office of Science Information Service, and plans for such a service
to be conducted by the Library of
Congress - are now in the making.
At present, no Federal source has the
responsibility of maintaining central information as to all Federal and Federally-supported information resources. The result is
that no matter how conscientious and helpful
anyone information office or library or center
may be, it is not in a position to give judgment outside its field of competence and responsibility to an inquirer as to which other
sources he might profitably contact.
It has long been felt by expert observers
that the nation's scientific and engineering
community should be enabled to tap into the
total Federal information system at anyone
or more points and get the benefit of leads to
information available throughout the entire
system.
It simply cannot be assumed that even the
best informed inquirer, in or outside of government, will be familiar offhand with all the
many resources which might be contacted on
any given subject.
INFORMATION RESOURCES
INVENTORY
Only now does the Federal government
itself possess a definitive inventory of the
nation's scientific information resources.
A landmark study 11 sponsored by the Office
of Science Information Service and conducted
by the Battelle Memorial Institute, has identified 427 different organizations or projects
throughout the nation. Criteria for the listing
were carefully applied. All of the listed resources specialize in some subject area; all
provide some type of information service beyond the publication of a periodical. All are
currently operational and are available to
some segment of the scientific community,
provided security, proprietary, society membership or other described restrictions are
met. Only services indicating some degree of
service beyond their immediate organization
have been listed.
The 427 groups represent a winnowing from
an original potential list of some 10,000 organizations which were identified by response
to N.S.F. press releases or by selection from
directories, association lists, registers or
reference files.
The U. S. Air Force has also recently published a helpful guide 12 to organizations within
its area of interest.
With the help of these and other inventories,
it is now possible to take many steps which
were not previously possible or practical.
These steps include:
(a) comprehensively informing specialized
segments of the science and engineering communities of all the services
within their particular field of interest;
and
(b) encouraging individual services within
a given area of interest to open up or
strengthen discussion among themselves so as to increase cooperation.
Illustration of the types of cooperation
would be possible use of a common
thesaurus of "key words" or "descriptors" as guides to the information
systems, possible efforts to minimize
unintended overlapping, etc.
DESIRE FOR ALL TYPES
OF INFORMATION
Another phase of increased understanding
is a change from the traditional view that the
scientist or engineer is allegedly interested
only, or essentially, in one type of information, i.e., in the contents of professional journals, in serial articles, or in published pro'ceedings or books, together with their abstracts and index guides.
It is increasingly recognized that many an
inquirer would like to have access to much or
all of the total of information, whatever its
nature or format, which may be available on
any given subject or subtopic.
"Access" does not, of course, 'mean that the
inquirer wants or could possibly use or digest
all the information; but it indicates he wants
the assurance that he knows where the information generally is, what it may comprise,
and how he or his librarian can get· it if he
so desires.
In effect, the inquirer may, in individual
circumstances, want the published, the unpublished and the prepublished.
He may want:
- summ~ries of Federally supported research and development projects which are
still in progress;
- technical progress reports on such
projects;
- completed reports;
- graduate these, in process or completed;
- translations of foreign articles, monographs and books;
- bibliographies, prepared by Federal or
non-Federal sources, particularly if annotated;
- reviews, preferably of an evaluative,
not merely descriptive nature;
- audio-visuals;
- or any of a variety of other types of
information.
SOCIETIES' PROPOSED STUDY OF
NATIONAL PLAN
Officers of many of the maj or scientific and
engineering societies have demonstrated that
25
they feel the time is appropriate to plan for
a national system.
The National Federation of Science Abstracting and Indexing Services has made
application to the National Science Foundation, Office of Science Information Service,
for support of a management-type survey of
the needs and opportunities for a national
system. Many or all of the member societies
of the federation are well aware of the enormous costs which implementation of such a
national system might entail. Even at present,
the principal societies are hard pressed to
provide publishing, abstracting and indexing
services for their clienteles. But the societies
seem more willing today than ever before to
come to grips more decisively with the challenge now confronting them and their colleagues and looming in the years ahead.
These, then, are a few of the signs and portents on the national information scene and
on the national horizon.
FORTHCOMING SUBCOMMITTEE
STAFF REPORT
Other propitious circumstances could be
enumerated by way of confirming that now
is the right time to initiate vigorous followthrough.
The circumstances are set forth in a staff
report scheduled for publication by the Subcommittee on Reorganization and International Organizations of the Senate Committee
on Government Operations. The tentative title
of the report is "The Crisis and Opportunity
in Scientific and Technical Information".
This report will summarize progress in
scientific and technical information since the
committee and the subcommittee began their
review of the problem more than 3% years
ago. The report will pick up from where four
previous committee and subcommittee reports
and prints left off.
Thus, it will supplement:
(a) the two reports 13,14 on "hardware" and
"software" in agency information systems, and
(b) the two reports l5 ,16 on management of
information on current research and
development (as distinguished from
information on completed projects).
It is not now possible to preview in detail
all of the findings in the staff report. But
26
Senator Humphrey has been anxious to help
the scientific and engineering communities
move ahead as rapidly as possible in their own
plans. Therefore, I am in a. position at this
time to submit at least a tentative listing of
a few of the possible staff recommendations.
1. National goal.
The objective of a System of Information
Systems should be established as a national
goal.
Based on inter-agency agreement and on
the consensus of non-governmental organization, there should be formulated, to the extent
that it is procedurally and fiscally practical,
a national program, national time-table and
national allocation of resources.
2. High level unit.
Progress toward a Federal System of Information Systems should be monitored at the
highest level within the executive branch.
An organization unit specializing on information should be established within the
Executive Office of the President. It may be
recalled that the Subcommittee on National
Policy Machinery, of the Senate Government
Operations Committee, has proposed an Office
of Science and Technology be set up there.
If such an office is established, it should
have an information unit. On the advice of
the Director, U. S. Bureau of the Budget;
Director, Office of Science Information Service, National Science Foundation, and other
authorities, this unit could serve as an influential catalyst toward improvement in Federal
information systems.
Simultaneously, there should be set up as a
standing committee of the Federal Council
for Science and Technology, a Committee on
Information.
3. Inter-library plan.
The three great national libraries should
prepare a memorandum of agreement. It
would replace present relatively haphazard
cooperation with a plan for optimal division of
labor in serving the scientific and engineering
communities.
Needless, unintended overlap could thereby
be minimized in acquisition, cataloging, bibliographic and other reader services.
Each of the three national libraries is preeminent in its field; each has greater resources
than are now being utilized; each can help the
other, and the national information community, to a greater extent than heretofore.
4. Inter-agency compatibility
Through the technical counsel of the N ational Bureau of Standards' Data Processing
Center, efforts should be strengthened to assure optimal compatibility between the major
inter-agency and intra-agency systems.
Illustrations of inter-agency systems are:
O.T.S., Department of Commerce; Science
Information Exchange; etc. Illustrations of
major intra-agency systems are: D.O.D.;
A.E.C.; N.A.S.A.; N.I.H.; etc.
Research should be expanded as to how far
individual systems - manual or mechanized
- may vary one from the other (as they do
vary at present) while still offering reasonable opportunity for compatibility.
5. Inter-agency modernization.
Each of the principal Federal agencies supporting research and development should
review its information policy, structure and
procedure. To the extent which may be necessary, each should make necessary revisions
so as to serve more efficiently in-house and
extra-mural administrative, scientific and
engineering needs, as well as inter-agency
needs.
With but few exceptions, Federal agencies'
information activities are, at present, neither
sufficiently coordinated within departments,
nor sufficiently modern and user oriented.
Among the notable exceptions in the U. S.
Atomic Energy Commission; it possesses perhaps the most integrated, technical information program of any Federal agency. Building
on experience under the A.E.C.'s program,
the National Aeronautics and Space Administration has devised a well coordinated
organization and blueprints for advanced
information activity.
But other Federal agencies have not as yet
placed themselves in a position where they
can adequately serve their own or others'
needs. Many of the principal departments are,
by nature, conglomerations of virtually separate agencies; the result is that their respective information activities tend to be isolated,
one from the other, with virtually no organic
unity. Illustrations of this circumstance may
be found in the Department of Defense, De-
partment of Commerce, and the Department
of Health, Education and Welfare.
Remedial action is essential. It must be recalled that in the final analysis, the proposed
Federal system of inter-agency information
systems will be no stronger than its principal
intra-agency components.
6. Funding information programs.
"Lirie items·" for information services
should be established in Federal appropriation
laws for each of the major research and development agencies of the United States
Government.
At present, there is virtually no regular
accountability to the Congress efficiency and
economy in most Federal or Federally-supported information services. Only a handful
of Congressional hearings annually take up
the issue of the adequacy of scientific and
technical information. Among these exceptions are the hearings on appropriations for
each of the three national libraries, for the
Office of Technical Services of the Department
of Commerce, and for a few others. If a "line
item" is written in, there will be assured an
annual "dialogue" between information experts and the appropriations subcommittees
most directly concerned.
Such a line item should be established, for
example, in the Department of Defense appropriation, as an integral part of the funding of
research, development, testing and evaluation;
for the Science Information Exchange, in the
Smithsonian Institution; for the information
services of the National Aeronautics and
Space Administration; and for the U. S.
Public Health Service, as well as for other key
agency programs.
7. Department of Defense planning
and policy.
The Department of Defense should establish central monitoring of its farflung information programs at the highest policy level.
D.O.D. should be given a CongressiomilExecutive mandate to do more than merely
respond to information requests from its
contractors.
The most logical location for central information surveillance would be in the Office of
Director of Defense Research Engin~ering.
Here, policies, plans and programs for the
information activities of the respective serv-
27
ices should be reviewed, coordinated, and
strengthened.
D.O.D. should be given the green light to
take the initiative in disseminating its vast
reservoir of scientific and engineering information. It should be authorized and directed
to change its relatively passive policy.
8. Str'engthening A.S.T.I.A.
The great asset represented by the Armed
Services Technical Information Agency should
be capitalized upon to a much greater extent
than heretofore.
A.S.T.LA. represents a powerful instrumentality for the nation's defense effort, as
well as for civilian science and technology. At
present, however, A.S.T.LA. does not have the
resources to do its existing, limited job, much
lessto perform the tasks which it wants to and
should perform, consistent with its mission.
9. National Science Foundation.
The Office of Science Information Service
should be given the means to strengthen and
accelerate its services for the Federal government and for the nation's scientific and engineering community.
O.S.LS. has probably served as the most
influential single force for progress in scientific and technical information throughout the
world. Although its information resources are
less in magnitude than those of the Department of Defense, it has been able, from its
central vantage point and professional interest, to allocate seed money for the broadest
variety of information programs.
If the state of the art in information and
communications is to advance at the desired
pace within the next decade, O.S.LS. will require more resources. It will, moreover, have
to devote resources to support bolder experiments in innovative information systems.
10. "Institutes of Communication Science."
Basic research in documentation and communications problems should be provided for
by a series of university based "Institutes of
Information (or Communication) Science."
An informal suggestion 17 for such institutes, prepared by a group of experts acting
as individuals, deserves serious consideration.
Fundamental problems in linguistics, and in a
variety of other disciplines, represent formidable obstacles to the future of documentation
28
and communication science. A few of these
problems are at present under study by some
Federal, commercial and not-for-profit private
organizations. Systematic, Federally supported study of these problems at a few leading
universities might have a profound and beneficial effect. Such study should be part and
parcel of a program of postgraduate education, leading to the granting of a degree.
11. Inter-disciplinary training.
The supply of manpower, skilled on an
inter-disciplinary basis in information, documentation and communication science, should
be greatly expanded. Plans of the National
Science Foundation for training grants to
schools of library science should be implemented at the earliest possible date. The Federal government should establish a career
corps of information specialists. The corps
should be trained not only in communication
science, but also in the information needs of
major scientific disciplines, so as to be
of maximum service to principal Federal
research missions. The members of the corps
should be systematically rotated among the
agencies as well as in university and ether
settings. Adequate salary, tenure and other
incentives should be provided.
12. Evaluation criteria.
Refined standards must be established for
purposes of evaluating the performance of
information systems. An important contribution toward this goal has been made by the
Stanford Research Institute,18 under the
auspices of the National Science Foundation.
Much additional work needs to be done,
however, in order to determine sound costbenefit ratios for centers, systems and services.
It cannot be taken for granted that under
all or most circumstances highly advanced
computer systems will necessarily pay their
way in terms of benefits to end users. Careful
scrutiny must be made of actual use to which
systems are put, by whom, for what purpose,
when, under what circumstances, with what
results, as compared with alternative systems.
13. Author abstracting and indexing.
Editors of professional journals should consider a policy of requesting abstracting and
indexing at source.
An increasing number of organizations
have reported successful experience with the
procedure of having the author of proposed
journal articles prepare an abstract and a list
of key index words at the time he completes
his paper and forwards it. Many sources feel
that whatever the limitations of amateur
abstracting and indexing at source, it may
still offer' an important time saving. Professional abstracters and indexers can thereafter
review and polish the author's version much
more readily than if they had to originate
the effort. The Engineering Joint Council is
now actively planning to apply to all of engineering the successful experience of the
American Institute of Chemical Engineers 19
in abstracting and indexing at source. A master thesaurus is c0'ntemplated which W0'uld
serve f0'r all of engineering as the Chemical
Engineer's Thesaurus n0'W serves that specialized profession. A training pr0'gram for
edit0'rial staffs for S0'urce indexing is also
planned.
14. Management in/ormation.
Information systems must be strengthened
t0' provide cust0'mized services for purp0'ses
of advanced management 0'f research and
development. PERT (Program Evaluati0'n
and Review Technique), which has made S0'
vital a c0'ntribution to the success of the
Polaris fleet ballistic missile system, should
be adapted in other administrative areas as
rapidly as possible. 20
ADDITIONAL SUGGESTIONS BEAR
CONSIDERATION INTERNATIONALLY
Many other suggestions will bear consideration, particularly in the sphere 0'f international
cooperation. A world-wide System of Information Systems should clearly be the goal
of international scientific and engineering
organizations.
As its name implies, the Subcommittee on
Reorganization and Internati0'nal Organizations is deeply interested in international
action, particularly on the part of intergovernmental organizations.
We have concentrated on Federal and
national needs, because the parent committee
is the Committee on Government Operations.
But committee members are deeply cognizant
of the primary role of non-governmental
groups both at home and abroad.
The willingness 0'f private groups to co-
operate with one another in the common
interest will be put to the greatest test in the
years ahead.
W. T. Knox, now Chairman of the Science
Information Council, and Manager, Technical
Information Division, Esso Research and
Engineering Company, has stated :21
"The maj or national problems in this field
stem from the enormous complexity of the
existing system and the reluctance of any
organization to give up some of the known
benefits of autonomy for the hoped-for
advantages of a cooperative effort.
"One [problem] is the urgent need for the
professional scientific and engineering societies to take greatly increased responsibility
for the development of a unified United States
informati0'n system. It is the consensus of
m0'st groups who have studied the problem
that the basic control of the scientific and
engineering information system should remain with the scientists and engineers. Yet
it is painfully obvious that a number of professi0'nal societies are not fulfilling their
obligations."
THE CHOICES BEFORE THE NATION
The United States is at a turning P0'int in
its program of scientific and engineering
informati0'n.
We have reached this point by virtue of the
dedicated labors of small numbers of pioneers
in information and communications. They
have laid a firm foundation for future action.
But the action must be coherent, not chaotic;
planned, not haphazard.
The coherent plan can only come from the
cooperative efforts of a vast number of individuals responsible for information offices,
libraries, centers, systems and services. Each
month, these information resources tend to
pr0'liferate further, and the task of linking
them grows more difficult.
Inevitably, each would view the problem 0'f
a System of Systems from the standpoint of
his own mission. Each would view proposed
changes, particularly collab0'rative efforts,
against the criterion of what his group stands
t0' gain and what it may lose in terms of
independence or increased cost in striving for
compatibility.
These are perfectly understandable criteria;
they are far from insuperable. A free society
29
is ultimately based on the "engineering of
voluntary consent."
The future is rich with opportunity. It is
also fraught with hazard if efforts prove too
little and too late.
Mere progress in information systems is
not enough. Knowledge, particularly at the
inter-disciplinary frontiers of science and
technology, is growing at an exponential rate.
The capacity and flexibility in systems must
be strengthened at a tempo and to a degree
which keeps pace with, or preferably, surpasses the growth tempo of knowledge itself.
The decision to move ahead with all proper
dispatch toward a national information goal
cannot be made by only one source, even by
the President of the United States.
It can only be made, and be implemented,
by the scientists and engineers of the nation,
acting voluntarily through their chosen
instruments.
The hour is late. The need is great.
The choice is: to drift or to steer; to break
through to a higher level of efficiency' or to
plod along with antiquated ways; to be content with pro-forma cooperation or to make
information teamwork a living reality.
NOTE:- The views presented herein represent individual judgment; .theyare not intended as an official
statement of the policies of the Senate Committee
on Government Operations or of the Subcommittee
on Reorganization and International Organizations,
except where specifically stated otherwise.
REFERENCES
'U. S. Bureau of the Budget, "The Budget of the
United States Government for the Fiscal Year
Ending 1963," 1962, Special Analysis G-1, p. 327.
'Ibid., pp. 332, 334, 335.
3
Ibid., p. 330.
4Ibid., p. 329.
5National Science Foundation, "Federal Funds for
Science," Surveys of Science Resource Series, December 1961, NSF 61-82, p. 44.
6The attention which the U.S.S.R. accords, in turn,
to the problem may be seen in a U. S. translation
of five reports published by the U.S.S.R. Institute
of Scientific Information, Academy of Sciences. The
translation appears under the title "Foreign Developments in Machine Translation and Information
Processing," No. 30, distributed by the Department
of Commerce, Office of Technical Services, OTS:
61-31, 465, June 2, 1961.
30
7
Address, "Industry's Contribution to Military Research and Engineering Programs," American
Management Association, New York, N.Y., Oct. 6,
1961.
8Editorial, "What's the Price of Information Unlimited?" The Trend in Engineering at the University of Washington, Oct. 1960, p. 2.
9"Information Resources, A Challenge to American
Science and Industry," based upon the Proceedings
of a Special Meeting of the Council on Documentation Research, Western Reserve University, Cleveland, Ohio, Feb. 3-4, 1958, p. 7.
10American Medical Writer's Association, Quarterly
Bulletin, Feb., 1962, pp. 1-2.
II"Specialized Scientific Information Services in the
United States, A Directory of Selected Specialized
,Information Services in the Physical and Biological
Sciences," National Science Foundation, NSF 61-68,
Nov., 1961.
12United States Air F-orce, DCS/Plans and Operations, Headquarters, Office of Aerospace Research,
Washington 25, D.C., "Directory of Rand D Information Systems," A Listing of Centers, Services,
Sources and Systems Engaged in Collecting, Storing
and Disseminating Scientific Data and Information
Applicable to Aerospace Research and Technology,
Aug. 1961.
I3"Documentation, Indexing and Retrieval of Scientific Information," A study of federal and nonfederal
scientific information processing and retrieval programs, Senate Document 113, 86th Congress, Second
Session, June 23, 1960.
14"Documentation, Indexing and Retrieval of Scientific Information, Addendum to Senate Document
No 113," Senate Document No. 15, 87th Congress,
First Session, Mar. 9, 1961.
""Coordination of Information on Current Scientific
Research"and Development Supported by the United
States Government," Senate Report 263, 87th Congress, First Session, May 18, 1961.
16"Coordination of Information of Current Federal
Research and Development Projects in the Field of
Electronics," Sept. 20, 1961.
17 Joseph Becker, Robert M. Hayes, Klaus Liebhold,
Melville J. Ruggles, Roger Sisson, Proposal for an
Institute of Information Sciences, Mar. 9, 1961,
revised July 16,1961.
IBC. P. Bourne, G. D. Peterson, B. Lefkowitz, D. Ford,
"Requirements, Criteria, and Measures of Performance of Information Storage and Retrieval Systems," SRI Project No. 9741, Dec., 1961.
19B. E. Holm, "Information Retrieval - A Solution,"
Chemical Engineering Progress, Vol. 57, No.6, June
1961, pp.73-76.
Address, "Managing the Fleet Ballistic Missile Program," National Security Industrial Association;
Jan. 11, 1962. Excerpted in Aviation Week and
Space Technology, Jan. 29, 1962, p. 21.
20
""The Technical Information Crisis," Industrial Research Institute, Pittsburgh, Pa., Oct. 16, 1961, p. 6.
By Paul W. Howerton
Information for Industry, Inc.
4. Status of Technical Information Centers
IF ONE MAY ADAPT A CLASSIC PHRASE from
that great practitioner of the English language, Churchill, it can be said that: "Never
in the history of human communication has
so much been written by so many which is of
use to so few." The question before us is:
"Why are these writings of use to so few?"
My contribution concerns itself with the
status of technical information centers with
the word "status" to be understood to mean
both "prestige" and "hierarchal position".
I shall attempt to avoid confusion of Utopian
ideal with reality by discussing status already
achieved, but I shall reserve the right to make
forecasts at the end of this paper concerning
certain pragmatic developments in the handling of technical information in both the
short and long range. These forecasts will be
made in the general context of the status of
the technical information center.
In the course of this paper, a recurrent
theme of service to the user will be evident,
and the principal thrust of my argument will
be that the term "status" is semantically
equivalent to "se.rvice." The two generic types
of technical information centers to be discussed are: 1) the internal technical information center which serves its own parent
organization exclusively, and 2) the center
which makes its service available to the
scientific and technical community without restriction beyond normal administrative rules.
STATUS OF THE USER
The popular assumption that information
center users are, by definition, both industrious and well informed, not only about their
personal subject interests but also about how
the information is retrieved, needs to be dispelled. A symbiosis of user and proprietor of
technical information centers results in an
advantage to both parties. The user who com-
municates his total problem to the center's
information specialists gains status by virtue
of his acceptance of the specialists' talents
in solving his problem completely and
expeditiously.
There is, however, a more important status
gain for the user who understands the capabilities and limitations of the center. He tends
to improve the capabilities and reduce the
limitations of the service by his active participation with the information specialists in
the identificaton of the association of units
of information.
On the other side of this partnership, let
the technical information center proprietor
volunteer his expertise to the scientist. Traditional library practice has been reluctant to
allow the client direct access to the stacks and
to data or information files. There were good
reasons for both of these actions, but in today's world, information services can still
defend the integrity of collections while guiding the user through the intricacies of the
sophisticated organization of the system.
In other words, the only purpose of a technical information center is to serve its
clientele. The improvement of the status of
the user by responsive action in the center
will be reflected in a concomitant improv~
ment in the status of the center.
THE INTERNAL TECHNICAL
INFORMATION CENTER
'When a company, a government agency or
a component of either operates a technical
information center, it assumes a responsibility to supply adequate and useful service
to its customers. This responsibility is frequently not met because of lack of status
witl1in the organization which is to be served.
As the former Deputy Assistant Director for
Central Reference, I know that the real suc-
31
cess of the central reference facility of the
Central Intelligence Agency is largely due to
the expertness and ingenuity of its staff; but
the fact that Mr. Allen W. Dulles and General
C. P. Cabell, as the Director and Deputy
Director respedively of the CIA, actively
supported the efforts of the staff, guaranteed
the status of the facility. If someone dismissed
the central reference service as inadequate
or incompetent, he knew that the justification
for his statements might have to be presented
at the highest level of the agency.
The status of any internal technical information center is directly proportional to the
support given the effort by top management.
It is this statement that must be the keystone
of the center. This position leads one into the
question of what should be the orientation of
a technical information center in an organization. Should there be a horizontal administrative r~lationship with customer components;
that IS, should each division head have
authority jointly or independently to determine policy for the center? Or should there
be vertical administrative responsibility to
a top official to whom customer component
heads also report? I have no doubt but what
the latter course is the wiser, both from the
point of view of the center and its clientele.
Let me now refer to my earlier points about
user status. I have found that the most effective internal technical information centers
had their reference staffs substantively
trained in the fields of science of greatest
interest to the people using the center, and
secondarily trained in library techniques. The
so called literature scientist falls into this
pattern of preparation for exereising his ingenuity to induce new levels of knowledge
from his base of experience, both with the
tools of the library and his discipline. Because
these internal technical information centers
are usually, but with some noteworthy exceptions, defining their scope fairly narrowly,
the 'qualifications I suggest are not difficult to attain. Customer acceptance is much
more likely to be easily established in an
environment of peers.
THE PUBLIC TECHNICAL
INFORMATION CENTER
There has been a deploring of the situation,
a wringing of hands and wailing about the
32
inadequacy of technical information centers
available to the general public; Long and
tedious hearings have been held by the Congress on the subject. The press, both technical
and general, has had feature articles or editorial comments on the subject. But when a
rigorous, rather than emotional, analysis is
made of the use of existing and competent
aids to technical documentation, such an
apathetic view is held by so many scientists
that one is startled by their vocal concern for
technical information centers. The findings of
surveys made under National Science Foundation auspices support the contention that
the present scientific and technical literature
of the world is not being used by the scientific
community as it should be. Why does this
condition exist?
I believe the reason lies in the status of
these centers and tools in the scientific community. Those who view this situation with
alarm do not realize that documentation and
documentalists have now reached a professional level commensurate with the other
members of the total research team. The
establishment of technical information centers is regarded as deflecting· funds which
could be directed to support of laboratory
science.
There are many who believe this technical
information center problem belongs to government alone. I do not count myself in this
group. Private enterprise, which must provide
a satisfactory product or fail, has a strong
motivation toward the efficient and useful
operation of information centers, just as it
does in other marketing enterprises. I have
urged that when machine translation becomes
an operational capability,the production of
translations be turned over to private
enterprise.
The task oriented service bureau to which
industry, government or education may turn
for comprehensive coverage of the literature
of the task on a retainer or fee basis, must
deliver the goods or fall by the wayside. The
status of these privately organized and administered centers will be measured by the
service they perform in response to the
interests of their clientele.
Small companies, whose budgets cannot
justify the establishment of· a technical library facility, can command the services of
substantively competent documentalists in the
centers set up by private enterprise. The integrity of the staff of the center will protect
the proprietary interests of the client in relationship to other clients in the same field.
Without this assurance, the center's prestige
and status will fall into oblivion. For a fee
considerably less than the cost of setting up
its own information center, a company can
have linguistic, substantive and documentation talents applied to its problem which, in
many cases, even the largest internal technical
information centers cannot assemble.
The compatibility of an information service,
such as an abstracting or current-awareness
service, with the technological advancement
in the mechanical linguistics of indexing, the
hardware of storage and the techniques of
reproduction, must take into account the interdisciplinary flavor of modern research. The
tremendous corpus of technical literature
already in being must be tied logically to the
generations of research results of the future.
The scientific method prescribes that from a
data base of the past, extrapolations into the
unknown are made followed by the collection
of evidential support to change the hypotheses
into scientific fact. This retrospective search
of the literature of the past has to make a
compatible link-up with the handling techniques of the newer reports of today and the
yet unwritten material of the future. To do
this thing breeds status.
CONCLUSION
The burden of my discussion can be summarized as follows: Status is a function of
service; service depends on alertness to the
changing interests of the clientele; and alertness depends on positive efforts to attain the
best communication possible between the technical information center user and proprietor.
The support of the internal center by top
management is essential to the achievement
of status and recognition. The so called public
technical information center or service, operated by private enterprise, has a dimension
of incentive to service not possessed by efforts
within the government structure - that of
satisfaction of the customer or failure.
Up to now there has been little real effort
to let the rules of supply and demand apply
to the handling of technical information centers. I believe the time has come when the
economics of the literature problem should
enter the picture. The adage that things are
usually worth what you pay for them, or will
quickly adjust to that condition, would be the
term of reference for establishing a new
approach to our concern for the increase in
the technical literature problem.
33
By John Sherrod
library of Congress
5. Functions of a Technical Information Center
THREE ASPECTS OF THE modern day technical
information center are sufficiently important
to single out for genera.} discussion. These are
the informational materials serviced, the
clientele served, and the operating procedures
employed to bring information and user
together.
ASPECTS OF A TECHNICAL
INFORMATION CENTER
The principal function of an information
center, and indeed its only excuse for existence, is to serve the needs of its clientele. In
many ways it is similar to the special library.
The materials serviced are generally specialized in that they are narrow in subject breadth
but comprehensive in depth. The clientele are
more often than not sophisticated in the subject matter of concern to the information
center, though not necessarily well informed
of their own information needs or of the
problems inherent in information storage and
retrieval. Finally, the methods of information
handling are likely to be classed as nonconventional because they usually employ
advanced documentation techniques, including
mechanization, which are designed to ensure
subject analysis in depth for information storage and maximum efficiency in information
retrieval.
The unique functions of a technical information center spring primarily from operating
proc.edures based on a policy ·of aggressive
information collection and dissemination. In
contrast with normal library service which is
more passive in nature, e.g., information may
be supplied only on request, an information
center can be expected to be continually aware
of the present, as well as, anticipate future
needs of its clientele and to actively disseminate information to meet these needs. Further,
the information center will undertake to edu-
34
cate present and potential users to their need
for, and the value of, information assistance.
As the store of information becomes both
more voluminous and specialized, there is a
growing requirement for bigger collections
and better trained subject specialists to serve
more adequately the user's needs. To store and
subsequently retrieve all existing information
on any subject is impossible. A maximum degree of completeness in a subject search
requires considerable resources. The expense
involved often can be justified only in exceptional cases, such as the national defense.
FEDERAL GOVERNMENT ROLE
Since it is certain that considerable and
ever increasing amounts of money will be
required to support information centers, the
question naturally arises as to the role the
Federal government should play in this effort.
One of the key questions being debated currently involves the degree to which centralization and/or coordination of technical
information centers is desirable or possible.
There are many reasons for not concentrating all or a major portion of this country's
activities into a single technical information
center. Among these are included the enormous cost of such an undertaking, the effective
resistance of some existing centers and other
entrenched interests, and the lack of any
positive assurance that such a centralized
effort would be successful. In fact, even the
most critical observers of existing systems
here will admit that in most cases a reasonably good job is being done at present.
In spite of these arguments, a gradual increase in the degree of centralization and
amount of Federal coordination appears to
be inevitable. For one thing, it will never be
possible to ascertain that a more centralized
effort would not produce better results unless
additional steps in this direction are tried.
Secondly, one might argue on a purely intuitive basis that there are certain inherent
advantages (as well as disadvantages) to
bigness. Finally, it is true, certainly, that
other countries are doing it. The French and
Russians have, to a considerable degree,
centralized their technical information efforts; the British are planning an extensive
centralized effort.
Against this background of controversy,
and lacking positive evidence as to the exact
amount of centralization or decentralization
that is desirable, existing technical information centers within the Federal complex are
developing progressive programs to serve
both the interests of their own individua.l
agencies and those of the country as a whole.
It is the purpose of the remainder of this
paper to trace briefly the program development of one of these information centers.
SCIENCE AND TECHNOLOGY
AT THE LIBRARY OF CONGRESS
The activities of the Library of Congress
are much broader in scope than a strict interpretation of its name would imply, although
service to the Congress is its primary responsibility. It has become, in fact, a national
library serving the whole Federal government
and, by means direct and indirect, the entire
nation.
The Congressional Act of April 24, 1800,
which established the Library of Congress
provided simply, "for the purchase of such
books as may be necessary for the use of Congress at the said city of Washington and for
fitting up an apartment for containing them."
The first significant material of a scientific
nature came with the purchase by Congress
in 1815 of the 7,000 volume personal library
of Thomas Jefferson, which was the finest
private library in the New World and one
assembled with a special eye to "natural philosophy". The greatest stimulus to the collection of scientific and technical literature by
the Library came in 1866 when, by act of
Congress, the Smithsonian Institution's "complete collection of the memoirs and transactions of learned societies throughout the world
and an entire series of the most important
scientific and literary periodicals" were transferred to the Library of Congress in order
that they might be housed and serviced more
adequately.
Today, the Library of Congress contains
the largest, most comprehensive, single collection of published and unpublished materials
in science and technology the world has ever
known. Such a statement might be difficult to
prove but equally difficult to disprove. It
would be safe to state, though somewhat less
dramatic, that the collections are this nation's
largest.
More than 25 percent of the Library's total
classed collection is in science, technology, and
closely related areas of knowledge, or nearly
two million volumes. Some 15,000 journal
titles in science and technology are received
currently. Nearly 300,000 technical reports
and some 600,000 manuscripts related to
science are included in the Library's holdings.
The size and scope of the Library's present
holdings attest to the vigor with which acquisitions in these fields have been pursued. In
general terms, it is the policy of the Library
of Congress to collect extensively the current
publications of the world in all fields of science
and technology except for clinical medicine
and technical agriculture. Acquisition and information service related to these latter two
fields are excepted because of the extensive
programs of the Library of the Department
of Agriculture and the National Library of
Medicine.
What has the Library done to ensure utilization of these vast and unique resources?
The focal point of subject competence within
the Library in all fields of science and technology is the Science and Technology Division,
which has primary responsibility for reference and bibliographical services and for advising on acquisition of all pertinent scientific
and technological materials. Established in
1949 with a small staff to plan and develop
a science program, the division's program has
matured to a point where, in many respects,
it compares favorably to that of any national
technical information center.
Space does not permit spelling out in detail
all of the program, but some highlights may
be helpful. Scientific and technical inquiries
are received and answered from every state
in the Union and from nearly all the principal
countries in the world at an annual rate of
nearly 20,000 inquiries. Exhaustive bibliog-
35
raphies in mDre than 20 scientific fields are
under preparatiDn at any given time, with
the resulting publications made available to'
the general public. The majority of staff members respDnsible for the preparation of these
literature surveys and for answering scientific inquiries haye had graduate training in
science or technology, many at the dDctoral
level. The division, as a technical information
center, is constantly studying the application
of improved and advanced documentation
techniques in order to better cope with the
increasing volume of informatiDn.
The latest significant development in the
division's information prDgram has come in
response to many requests by industrial concerns and private researchers for broader access to government held collections of scientific and technical literature. TO' meet these
growing demands, the Science and TechnolDgy
Division in cooperation with the Office of
Technical Services, U. S. Department of Commerce, in July, 1961, initiated a program designed to prDvide an inexpensive method by
which a large segment of the public with special, often urgent, needs for research information can have almost immediate access to
the science collections of the Library of Congress, as well as to the specialized holdings
of the Office of Technical Services.
36
This new information program provides
for two kinds of literature service. The first
is a "current awareness" bibliography, designed to keep subscribers abreast of the
latest developments in their field Df interest
as these develDpments Dccur, or on a periDdic
basis. The second is a retrDspective bibliDgraphy service which provides for a literature
survey of the information available at the
time the request is made. Built in flexibility
of the service accommodates pin-point research on individual subjects while providing
at the same time for continuDus, in-depth
research within a broad related field. The fee
for this tailor-made service is eight dollars
an hour.
It is hoped that this brief report will serve
to highlight the rather unique services which
have come to be associated with today's technical information centers. This, together with
the description of government interest and activity, particularly the program Df the Science
and TechnolDgy Division of the Library Df
Congress, should serve as an optimistic note
for those interested in the management of
technical information centers. Much remains
to be done, for sure, but that many good things
are being done is certain.
By W. C. Asbury and J. E. Moise
Esso Research and Engineering Company
6. Technical Information Services
in an Industrial Organization
ENOUGH HAS ALREADY BEEN SAID here and in
numerous papers at other meetings about the
information crisis.', 2 One statistical reference
should suffice to provide the proper perspective for this discussion. Three hundred thousand technical articles are now being published each year in recognized periodicals,
with the combined reading time for all of
them in the neighborhood of 465 man years.
And it is likely that this rate of publication
will double in a handful of years.
As regards the general areas we propose
to discuss, we will first have something to
say about the need and justification for a technical information division in an industrial
organization. Second, we will expain how we
at Esso Research are equipped to carry out
our responsibilities, including a description of
the organization of our Technical Information
Division, which provides technical information services. Third, we will discuss our plans
for the future to maintain our effectiveness
in the face of the expanding volume of
information.
This discussion will be specific to our organization, whose technical information is in the
field of petroleum technology. However, this
does embrace quite a few disciplines, such as
chemistry, physics, and mathematics, with
others, such as biology, creeping in, plus the
practical applications of these sciences in
engineering fields.
MANAGEMENT SUPPORT
We have no magic solution to the problem
of handling the flood of technical information.
There is no one answer, and our system, although adequate for the moment, can stand
a lot of improvement. What we can do is to
assure you that the practice of science and
engineering has never been more dependent
on competent technical information services,
and that - unless Esso Research and Engineering Company is an exception - management people will strongly support efforts to
screen, process, store, and retrieve this avalanche of literature for the tailor-made benefit
of the user.
Such management support doesn't always
come easily, of course. Those of you who are
just setting up technical information programs will certainly encounter at least a lack
of enthusiasm from some quarters, if not
downright resistance. There are those who
state rather firmly that the technical man
should read everything in his field. Mr. Howerton's paper at this meeting also emphasized
the importance of management support.
One of us (Mr. Asbury) recently asked his
son-in-law whether and how he managed to
keep up with his professional reading. This
son-in-law is an engineer working in the field
of acoustics. He told Mr. Asbury that he didn't
have any trouble, that he was able to rely on
Physics Abstracts to keep abreast of what was
being published. In a highly specialized field
where good abstract services are available, it
is quite possible for a technical man to fend
for himself. In less specialized fields, the story
is different. No man can do all of the reading.
That's where technical information groups
enter the picture, and it's on that basis that
they must earn management's endorsement.
Now, to go further into why we have technical information services, let us first consider
the four principal objectives of technical management in industry. The first is to maintain
the company's capital and operating costs at
37
the lowest possible effective level. Second, to
help the company obtain new high quality
products, efficient processes, and other tech~
nical assistance at a level that will improve
or at least maintain its competitive position.
Third, to provide the company with new opportunities for attractive capital investments
and new markets for new and existing products. And fourth, to achieve the first three
with a minimum of research expenditures,
including those for technical information.
certain special information research services
that we'll touch on later. Our bulletinswhich generaIly publish abstracts within one
month after the article appears in the literature or a patent copy is received - are our
principal organs for achieving current awareness. This is not to say that researchers do
not read magazines. Most of our researchers
read certain magazines on a regular basis,
but secure references to other articles of interest to them through our abstracts buIletin.
These objectives can be difficult to achieve.
You can't schedule creativity or inventiveness
as you can manufacturing output. There has
to be a reasonable degree of investigative freedom. On the other hand, it's hardly feasible
to create technology at an outlay of time and
money that makes it economically useless. So
management must take the attitude that industrial research is a business, to be managed
as a business. And. be assured, a vigorous,
well working technical information agency is
essential to the businesslike conduct of industrial research and engineering.
The current-awareness items of today are
the retrospective searching items of tomorrow, which brings up the second phase, retrospective searching. This area involves indexing, storing and retrieving information. As
you will see later, we include in searching not
only storage and retrieval, but also analysis
and interpretation.
It's management's responsibility, for example, to decide when work should begin on a
research project, and when it should terminate. In most industrial laboratories, there are
always far more ideas being generated than
there are people to carry them out.
In all of our experience, we can't recall anyone's ever scratching his head and saying: "I
wonder what we ought to do next." Technical
literature is one of management's most useful
guides in making the decision as to which
project should get going and which should
wait or be scrapped entirely.
JUSTIFICATION OF
INFORMATION SERVICES
So far, we have jumped back and forth between two phases or categories of technical
information which constitute its principal
uses - current awareness and retrospective
searching. The first of these, of course, involves keeping our research people informed
of developments in their specialized fields in
order to minimize the number of periodicals,
patents, government reports, etc., which they
would otherwise have to try to read for themselves. Current-awareness activities involve
acquisition, selection, abstracting, publication
of bulletins, circulation of periodicals, and
38
Let us illustrate by example how a technical
information service can help laboratory research. Our people had under consideration
an exploratory program in the field of productive research; the need for it seemed beyond
question. It would take a month, perhaps
longer, to find the required answers, and we
were about to give the go ahead when one of
our information researchers came up with a
document that made the whole project unnecessary. In another case, an analysis of more
than 100 chemical compounds was stopped,
just before starting laboratory work, when a
literature search turned up the fact that
someone else had already done the work.
Let us look at a quote from a talk by P. L.
Saltzberg of duPont: "It might be that the
term R&D is becoming a cliche and no longer
means research and development. Perhaps it
means 'repeat and develop' or possibly 'reconnoiter and duplicate.' What an abysmal waste
of technical talent! What a dull existence."3
Duplication of scientific work can truly be
a sinful waste of time, money, and brainpower, and no research organization that I
know of has an oversupply of any of these
commodities. Of the three, time is perhaps the
most important. It alone is the irretrievable
factor, and it takes on an added significance
when the research effort is in any way tied
to our national defense. In the competition
between communism and the free world, the
Russians are very much aware of the v'alue
of published information in scentific work.
Every Friday night, a plane leaves this country for Moscow with a cargo of technical reports and trade publications. Every day, a
half dozen of our better newspapers go under
the microfilm camera for transmittal to Russia. We can afford to be no less aware than
they.
There are exceptions, of course. There are
times when it is more practical to re-research
something because of the difficulties involved
in obtaining the literature, or because new
techniques will provide more complete, more
meaningful results. Some of the points we can
consider before plunging into a project are
listed below:
1. Can we determine when further extension of the literature search will cost
more than the research wo:rk necessary
to get the same information?
2. At what point will it become uneconomical to delay the actual initiation of a
program in order to. discover further
pertinent literature references?
These first two points are rather specific,
but what are the more general benefits from
a retrospective search? They can be numerous, including such things are serendipity for
a browsing exploratory worker, direct leads
for further research studies, or complete aban.donment of a project, as shown in an earlier
example.
The evaluation of technical information on
the productivity of the worker is very illusive
and cannot easily be defined. Similarly, the
eventual effect that the information might
have as an over all benefit to the company is
difficult to evaluate, but the potential benefits
can be enormous.
INFORMATION SERVICE
AT ESSO RESEARCH
When we speak about technical management, of course, we can speak with authori.ty
only where our own company is concerned. It
might be appropriate, therefore, to teIl you a
little about how our technical information
services are carried out.
To begin with, Esso Research and Engineering Company is the principal technical
affiliate of Standard Oil Company (New Jers.ey) and its world-wide manufacturing and
marketing affiliates. Esso Research is made
up of nearly 3,000 men and women, about half
of them professional people. Our main facilities are in New Jersey, in Linden and Florham Park. Our responsibilities to our affiliates
include coordinated research and engineering
programs concerning petroleum and petrochemical products and processes. In addition
to the work of our own people, we also are
responsible for related activities at affiliated
research installations elsewhere in the United
States, as weIl as in Canada, England, France,
and Germany. In terms of money, this effort
comes to $78 million a year.
Our company came into existence as a technical department of Jersey Standard, back in
1919. From the beginning, efforts were made
Esso Research and Engineering Company's
Technical Information Division
I
Manager
I
Library and
Systems
Research
Specia Ii zed
Information
Centers
Information
Processing
and Meetings
Information
Research
EXHIBIT 1
39
to support our technical people with the literature services they needed, including a well
stocked library. Our abstracting services of
patents. and literature started in 1920, but it
was not until 1957 that we expanded the scope
of our information services to give this activity full-fledged divisional status. This was
brought about by an internal survey of our
technical people in 1956.
We are somewhat chagrined to find from
this survey that our information program
wasn't as effective as we'd thought. Many of
our technical staff members felt that they
weren't as up to date in their own fields as
they could be. In some cases, our research
people didn't even consider themselves adequately informed about the company's own
work. The result was the creation of our
Technical Information Division, five years
ago, to strengthen and expand the job of
gathering technical information and making
it more easily available.
Exhibit 1 shows an organizational plan of the
Esso Research and Engineering Company's
Technical Information Division as it pres.ently
exists. As shown, the four principa.I areas are
the library, information processing, specialized information groups, and information
research.
The library includes the functions of acquisition, cataloging, reference services, and
maintenance of keys to retrospective searching. The information systems research shown
here involves development of improved information handling techniques for most of TID's
operations, both from a cost standpoint and
for obtaining satisfied customers.
Information processing includes such current-awareness activities as the publishing of
five bulletins of technical abstracts and a
monthly index to the company's technical reports. The abstract bulletins are given wide
distribution to scientists and engineers
throughout the Jersey organization. Each bulletin is tailored to meet the requirements of
readers in a specific field. Carefully defined
section headings, listed in tables of contents,
make it easy for readers to find abstracts in
their areas of special interest. Speed of reading is enhanced by use of topical-sentencefirst abstracts and by "highlights" pages that
discuss the most important abstracts included.
In addition to the monthly index to company
40
reports, an annual index is prepared for reference use, and a machine based index is being
developed. Other activities in this area include
centralization of reports distribution to affiliates, management of technical meetings, and
clearance of technical publications.
The specialized information centers include
a crude oil information service which not only
accumulates all assays of company interest,
but also coordinates necessary action to get
new assays that are needed and to distribute
and interpret them. Another central activity
is on physical properties, primarily of hydrocarbons; this includes deriving correlations
for the prediction of physical properties.
Another new specialized center which the
Technical Information Division helped to organize and staff is the Engineering Information Center, located at our facilities at Florham Park. This Engineering Information
Center 4 (EIC) was set up to supply specialized needs of the Esso Engineering divisions.
This center indexes in depth all correspondence and other internal documents having
technical information of permanent value.
The entire system is arranged around a machine based indexing system. Searches can be
made manually for simple questions or through
machine retrieval procedures for more complex questions. Since the indexing is done in
depth, this system is felt to be capable of not
only rapid but fairly complete and selective
retrieval of documents.
The last area in the organization - information research - provides dynamic liaison
between old and new information from all
sources and current or planned research projects. This is the chief new activity of our
Technical Information Division. The information research function has been discussed
at length in a number of previous. papers,5,6
so we will only briefly summarize its requirements and activities:
1. Know current company interests.
2. Evaluate current technical literature in
fields of interest.
3. Direct pertinent information to right
people.
4. Guide abstracters in selection of material to reflect the company's current
research activities.
5. Prepare comprehensive reviews of lit-
erature and internal reports in specific
fields.
6. Analyze and interpret review information and suggest areas for further research studies.
Since we have now briefly described our
organization, it may be fitting to present a
few figures for our reports room and library
holdings that show the magnitude of our operations. Company reports on file number some
10,000; our present input is about 1,500 per
year. Information research reports issued by
the Technical Information Division, some 50
per year, are few in comparison, but, as stated
earlier, they are comprehensive reviews in
specialized fields. The library's files contain
some 80,000 patents, over 18,000 acquired in
1961. Books and bound volumes of periodicals
number 40,000; additions are now 2,900 per
year. Abstract cards are being added at the
rate of 100,000 per year, with the files now
holding about 2,000,000 cards representing
over 500,00 abstracts. Abstracts published in
our bulletins during 1961 numbered over
25,000.
About 600 journal titles are acquired, many
with multiple subscriptions, and some 2,000
books are cataloged each year (gradually increasing). The circulation of these journals
and books in 1961 was in answer to 43,000
requests, about half of which were based on
items in our abstracts bulletins. Library reference requests totaled 2,500 during 1961.
Of course, numerous other services are provided by our library staff, and these rather
broad and general type of data serve only to
orient you to the scale of our operations,
small compared to some, but large compared
to many other information service operations.
STAFFING
From the earliest days we have followed
the philosophy that technical information,
being one of our major resources, warrants
management's full support .and attention. We
staffed the new division with good technical
men carefully selected from our research divisions, so that our program would be properly
oriented in terms of our technical work. We
had no intention of setting up a rest home for
tired scientists and engineers. The men we
picked were expert in their fields. Some have
rotated back into research work, taking with
them a broader outlook than they had prior
to their experience in the information field.
We also found that it was difficult to attract
high quality scientists and engineers from
experimental research into the information
field, and this may be true in most countries.
Listen to this complaint from the U.S.S.R.:
"Regrettably, heads of industrial enterprises
often staff the technical information agencies
with workers made available by chance and
who are difficult to employ in other divisions."7
We have also, of course, supplemented our
staff with people having the specialized skills
necessary to an effective information program, i.e., abstracters, literature searchers,
information systems men, etc. Our staff in the
Technical Information Division now numbers
60, over half of them technical. The budget for
the operation is over one million dollars a year.
PRINCIPLES
At Esso Research, we feel there are some
fairly basic principles that cannot be ignored
in operating a successful technical information
service:
1. The information service must dedicate
itself to the active exploitation of recorded knowledge. It's not nearly enough
to maintain a storehouse of learning.
Information is worthwhile only if it can
be put into the active stream of daily
work. If this information can't be made
available in the most useful and usable
form - when it's needed, where it's
needed, and in the amount that it's
needed - then the program adds up to
little or nothing.
2. The information service should not operate in an isolation booth apart from the
men at the bench, but should work in
close liaison with project teams from
the preliminary planning stages on.
3. The information service is not an end
in itself, but is a device by which its
users get information to help solve their
daily problems. The user isn't interested
in the mechanics or details of the operation. He doesn't care how carefully the
whole program haS' been designed and
constructed. Does it work for him?
That's his concern.
4. As with all services, there is need for a
reasonable amount of research on new
41
information techniques. The evolution
of office equipment has spilled over
into the information service field, with
punched card, punched tape, rapid copying, microfilm, and computing equipment being used in ever increasing numbers. However, the idea of a mechanical,
fully automatic data handling and retrieval system has not yet been achieved.
More progress can be made in this area.
5. The service ought to be measured from
time to time to make certain it is fulfilling its role properly, effectively, and
at reasonable cost. Those best qualified
to judge the effectiveness of the service
are its users.
The last two items touch on planning for
the future. Some of the areas through which
we hope to maintain and improve the effectiveness of our organization as the volume of
technical literature, internal and external,
steadily increases include:
1. The application of new procedures or
techniques. As you know, numerous new
techniques are constantly being proposed
and developed for indexing, cataloging,
storing, and retrieving information. We
plan not only to work on our own ideas
in this field, but also to keep in touch
with others so that we can secure the
best system adaptable to our needs.
2. The use of automated systems. Here, we
are referring to the use of machines,
whether fully automated or not. This
could be considered part of the first
item in applying new procedures and
techniques but, since a differential can
be made, we list it separately. In addition, some of the machines necessary are
too expensive to justify. In our case the
expensive machines are already available,
and it is only a case of securing time
on them.
3. Cooperative efforts with other groups.
We consider these quite important, and
will make a few more remarks about
them directly.
4. Information systems research. Here,
specifically, we are thinking of orienting
our systems to the users. Our plans include additional surveys to determine
user requirements and just how effective our communications systems are.
42
5. Evaluation of the information systems
in terms of their costs, as compared to
the over all benefits to the users and to
the company. We should like to determine what are the real benefits of disseminated technical information, just
what use the user makes of the information after he gets it. Is there a way
of putting a dollars and cents figure on
the value to him and to his research
efforts?
ONE EVALUATION
It may be pertinent here to describe a simple exploratory study that we made to determine if we could evaluate the worth of technical information services to our research and
engineering organization. An industrial research library or technical information service, unlike a public library, must result in
economic benefit to the supporting organization. While there is ample evidence that
scientists and engineers use technical information services,8,9 there are no available data
relating operating costs for a technical information service to the economic benefits to
the organization.
This brief study was carried out to determine if a relatively simple survey technique
could measure the value of reference services
provided by a special technical library. Scientists and engineers. who had actually used the
library reference services were queried to
determine the use to which the reference information had been put. Twenty reference
questions from 18 different scientists and engineers were thus analyzed.
The following questions were asked:
1. Did the library provide you with an
answer to your reference question?
2. If so, how did you use this information,
and of what value was it to the company?
3. What would you have done if you had
not obtained this information from the
library and, in case you did not get the
information, what did you do?
The following results were obtained:
1. Answers to reference questions:
Answered completely
Answered partially
Questions not answered by requested information
Answer unevaluated by requester
Answer not received by requester
12
3
2
2
1
2. Use of information received:
Background on R&D projects
Specific markets data for products
Specific data used in calculations, patent
application, etc.
Information given to company customers
Preparation of talks, book
5
3
4
2
3
3. Consequences if information had not
been provided:
Repeat previous laboratory work
Discuss problem with expert
Potentially profitable lead would not
have been uncovered
Translation would have been ordered
Decision would have been more difficult
Existing, less efficient procedures would still
be employed
6
2
3
1
3
2
Many interesting observations can be made,
even from this limited survey, regarding the
role and functioning of a technical information service in an industrial research organization. Of special interest are the 60 percent
score (12/20) for complete answering of
queries, the 10 percent of cases in which the
questioner asked for information not germane
to his problem (this percentage may represent
those in our company who do not know
exactly what they are looking for), and the 10
percent of cases in which the answer still lay
un evaluated by the questioner.
More important for the current study are
the indicated consequences if the requested
information had not been provided. In 35 percent of the cases (6/17), previous laboratory
work would have been repeated. It proved
difficult to measure the probable costs of the
laboratory work made unnecessary by the
reference services. However, in one instance
the research chemist estimated that two weeks
of laboratory work had been saved. At current R & Dcosts, some one to two thousand
dollars of laboratory work was avoided by a
reference question costing about 15 dollars to
answer. In another instance, a U. S. patent
corresponding to a foreign language patent
was located, thus obviating the need for a
35 dollar translation.
The results of this exploratory study indicate that a user survey to determine tangible
benefits derived from specific information
services may provide a satisfactory means for
determining the economic justification for
technical information services.
So far, we have talked mainly about what
one company is trying to do. We would like
to say, however, that what anyone company
can achieve in this whole complex field of
information service will never be nearly
enough. The problem is big, and it goes far
beyond the borders of any single organization.
Thus, we believe we have an obligation to
support similar efforts by industry groups,
professional societies, and the government, in
any way that we can.
COOPERATION
For instance, we have been working with
the American Petroleum Institute to improve
and expand the abstracting and indexing programs of its Central Abstracting Service.
This industry group already abstracts over
140 journals. and the patents of 10 countries
on an "express service" basis, providing member companies with a service which would be
far more costly if each one had to produce its
own. The API's output of 25,000 abstracts a
year (a third of them in the petrochemical
field) has for the past several years been classified on cards ·for later use. The number of
cards gets unwieldy, and traditional methods
of indexing, such as the card index now in
use, have serious limitations. As a consequence
the API is now setting up a machine bas~d
retrieval system which we're proud to say was
derived from a system we developed at Esso
Research. This is one example of our belief
in the free exchange of information processing know-how. Another example, directed toward the national welfare, is the service of
one of our men as chairman of the Science
Information Council of the National Science
Foundation. Here, the contribution is based on
our know-how regarding needs and techniques.
A well organized information service, given
full and continuing management support,
needs one additional ingredient before it can
function effectively. It needs the cooperation
of its customers - the individual technical
men. Unless they're willing to use your services and know-how you're in trouble. In our
company, we make it a point to tell new technical people about the Technical Information
Division, and to encourage them to make full
use of its services.
One thing we need to do is sell the value
of our services to people who can't, or won't,
appreciate what can be done for them. One of
the arguments that you hear over and over
again in this regard, we're told, is that "no-
43
body can do my browsing for me!' This view
is generally more common among people in
exploratory research than it is among people
in development work. The latter know just
what they're looking for; the former don't
know. But what information people do for
the exploratory researcher is to supplement
his browsing, not substitute for it. Who can
tell when a relatively obscure piece of information in a remote field will catalyze the work
of a man who simply doesn't have time to read
everything? If you can convince your customers that you're not in business to spoon
feed them, you will have gone a long way toward encouraging them to make intelligent
use of your support.
INFORMATION INPUT
We also feel obliged to comment briefly on
the source of the information problem today
- the input aspect.
The exchange of technical information is
vital to the advancement of technology, but
are we publishing too much that isn't. really
essential to the forward momentum of science
and engineering?
There is a growing suspicion that we .can
no lon'ger afford the luxury of the adage,
"publish or perish."
We would like to see technical organizations, including universities, set up higher
standards for publication so that their people
publish wisely, not wantonly. We would also
like to see journals crack down more on marginal papers that don't say anything of moment, or which take six thousand words to
say w:h.at could be said in the confines of a
note or a letter to the editor.
We are also concerned that really worthwhile data finding its way into print often is
unavoidably a year to a year and a half behind
the actual state of. the art. More information
transfer must take the form of direct contact
between those who need the information and
those who are creating it. One step toward
encouraging this is the newly formed Science
Information Exchange at the Smithsonian
Institution, where a number of government
agencies are cooperating under National
Science Foundation sponsorship in coordinating information as to where, and by whom,
research is being conducted.
44
A thousand years ago, the accumulated
knowledge of the world was entrusted to a
small band of monks. Laboriously, they copied
their previous fund of recorded intelligence
so that it could be handed down to future
generations.
A hundred years from now, and perhaps
even sooner, your most modern methods of
handling information might seem as slow and
as crude. Revolutionary changes will come.
They'll have to if chaos is to be avoided - if
man is to avoid strangling on his own surfeit
of information.
In conclusion we would like to repeat what
we consider to be the responsibilities of the
Technical Information Division of the Esso
Research and Engineering Company from the
management standpoint. Broadly, the first
responsibility is to disseminate current technical information and to facilitate retrospective searching. The second responsibility is to
operate effectively and efficiently. The third
is to develop or procure the tools necessary to
maintain a high degree of effectiveness. The
fourth responsibility is to make the necessary
system research studies to orient the system
to meet the routine and the complex requirements of the users.
REFERENCES
1
Walter Sullivan, New York Times, Articles and Editorial, December 25-27, 1961.
2W. T. Knox, "The Technical Information Crisis,"
Meeting of Industrial Research Institute, Pittsburgh,
Pennsylvania, October 16, 1961.
3
P. L. Saltzberg, duPont Patent Anniversary Dinner,
November 14, 1961.
4"No Rest for Know-How," Chemical Week, p. 61,
January 27, 1962.
5G. H. Cloud and W. T. Knox, "Information Research
-A New Tool for the Petroleum Industry," 1959
Fifth World Petroleum Cong~·ess.
6W. T. Knox, "Information, P~ease," 53rd Annual
Meeting, American Institute of Chemical Engineers,
Washington, D. C., December 5, 1961.
7
Aran S, Melik-Sha}l:hrozarov, "Technical Information in the U.S.S.R.," Moscow, 1960.
8
S. Herner"Information-Gathering Habits of Workers
in Pure and Applied Science," Ind. Eng. Chem., 46,
p. 228-36, (1954).
9
E. Tornudd, "Study of the Use of Scientific Literature and Reference Services by Scandinavian Scientists and Engineers Engaged in Research," Preprints of Papers for the International Conference
on Scientific Information, Washington, D.C., November 16-21, 1958, Area I, pp. 9-66.
By DeWitt O. Myatt
Science Communications, Inc.
7. Designing an Information Center
to Meet a Real -System Requirement
My ASSIGNMENT is to address myself to the
matter of the working problems that come
when one attempts to translate information
science into a successful information center
operation. It may be termed the "engineerdesigner" function. This is a pre~llmptuous
position for anybody to assume in the present
state of ignorance: to link our adolescence in
information science to the human user, the
human institutions, and the patterns of human effort that ultimately bring to a point of
actual value any particular increment of
information in our recorded storehouse of
knowledge.
Among the underlying reasons behind the
great upsurge of concern about making more
effective use of scientific knowledge are dramatic illustrations from the past fift~en or
twenty years of ways this technique of science
and research can be utilized for the purposes
of man. Society has discovered science. Fifty
years ago, most of science was a matter of
interest and concern for men of intellect.
Twenty or thirty years ago, through such
industries as the chemical industry, it was
being exploited for economic affairs and gains.
Today, there is a very significant additional
use that is being made of it: the use of science
for social purposes. And I think that some of
our current concern with it, the feeling of
need to do the best we can with it, has been
stimulated by this largest and grandest recent
overtone associated with scientific effort.
This paper is not concerped with codes or
machines, and their technology or techniques.
Nor is it concerned with the library, mechanized or not, although it is a quite probable
component of the information c~nter.
In most technical fields today, the thought-
ful scientist views his obligation to remain
well informed with a sense of acute personal
distress. Almost certainly, the annual output
in his own specialty has continued to increase. But beyond that, our great triumphs in
merging science at the base have unblocked
new areas, previously the specialty of another, in which he can also be competent. Self
interest, if nothing more, presses him to command them as well, or accept consignment to
a relatively narrow and barren field.
A second major influence on the scientist
has been the way our society has altered its
expectations of him in the past generation.
In fact, we have only to go back a few years
to Sputnik I to see quite significant changes,
although the historic inflection point probably
should date from the OSRD activities of
World War II. Before then, the scientist really
lived in a comparatively simple and undemanding world. He roamed pretty much where
he pleased, and the terrain of a single discipline usually proved fertile enough to keep
him happy and rewarded. Society pretty much
looked on his useful discoveries as windfalls.
Since it had not yet grasped the full implications of creative research, it issued no
socially motived directives and established
no quotas for our scientists and technologists.
Today, the individual scientist usually finds
himself a part of large and complex technological endeavors. These endeavors reflect the
rapid industrialization of science and the
scientist. And, more and more often, they are
addressed to the meeting of a social desire or
need, which may be the conquest of a major
disease, the creation of a,.new tool of national
defense, or the enhancement of an economic
resource. The scientist and his science bear
45
the sobering responsibilities of cultural
maturity.
In this new environment, the scientist's
information need cannot be dismissed as a
uniquely personal concern. Increasingly it has
become viewed as a responsibility his employer
should share, most particularly with informational services that conserve the ever more
sought talents, and ever more costly man
hours, of available technical manpower. This
is the shift, we believe, that has led to the appearance of the modern technical information
center.
The scientist's very endorsement of the
information center concept suggests strongly
that new forces of actual historical magnitude have come into play. Especially significant is the scientist's philosophical acceptance
of the judgment of another in sifting subject
matter through a storage-retrieval mechanism that he does not ever expect to operate
personaIly. It is true, of. course, that he still
is free to use other channels that give him
direct contact with original material. But the
fact that research scientists will even concede
value to techniques requiring a third party
participant indicates the magnitude of the
influences they feel.
Because the information center is a young
and still vulnerable institution (perhaps, in
fact, the first undisputed child of an information surplus technological culture), we believe those who would design and operate
them should make a real effort, however
scanty the current firm knowledge, to understand underlying forces, the practical limits
they imply, and the special innovations they
encourage. This is the time for intellectual
rigor, so far as it can be applied.
DEFINING THE TECHNICAL
INFORMATION CENTER
In order to draw broad lines between
what the library and information center
can do and naturally tend to do, let us accept
for purposes of definition that the "pure"
library does not have peer competence in the
technical matter dealt with. As our growing
understanding of the tools of technical communication have advanced, the alert technical
librarian will begin to think in terms of
mechanizing the library, through the development of codes and through the choice of
46
machinery that allows an orderly search to
be made of items that have been indexed far
more deeply than was considered useful or
feasible previously. There is, I think, a general
tendency for the librarian to assume that
there is no way of knowing what kind of
question is going to be thrown at him or her,
and that therefore this encoding and preparation for mechanization ideally should be
comprehensive, including the total input to
the library and additional services that can be
subscribed. In general, the picture that is
evolved is one of a resource of value but one
that depends. upon outside request or demand
for it to function effectively. There is a passive service stance if one ignores the fact that
there are things like accession lists that do
represent initiative on the part of the
librarian. The orientation tends to be toward
documents rather than toward subject matter.
In contrast, in the information center there
is more of a tendency to ask, "What value is
this for whom?" and to inform th~ scientist
that here is something worthwhile and interesting. Instead of the librarian's reaction of
comprehensiveness, there is an assumption or
an arrogation that the information center
specialist knows what the scientist wants and
passes information to him on his own initiative. Rather than subscription to a service
that is of interest to the using organization,
although it had to be prepared to serve at
least 100 other organizations to be economic,
the information center staff person will make
selective extractions from material as it
comes in.
Any realistic definition of a technical information center must relate it to a larger
functional system. This larger system contains, in addition to our center, information
users and (since the use of knowledge almost
invariably creates new knowledge) it contains
information sources. FinaIly, the larger system contains a policy or managerial element.
The executive element decides the users to be
served and the sort of information and service
the center is expected to provide.
Let us formalize an assertion from these
comments:
"The proper informational mission of a
technical information center is a function
of the technical mission of the population it
serves." (ASSERTION I)
This assertion does not differentiate the
technical library from the information center.
We shall contend that the staff of the information center is professionally qualified in
the technical subjects handled, whereas the
special expertise of the technical library is the
management of documents. In any real situation, we rarely find either information centers
or technical libraries whose staffs are entirely
unqualified in the other's domain. But if we
attempt to fix the ultimate distinction between
prime competencies that lead to totaUy different roles within the organization, we believe
the test is the presence or absence of professional capability (i.e., capacity to function as
a technical expert) in the subject matter.
In talking hereafter of a technical information center, we mean an organization that
satisfies the following definition:
"A group serving a technical organization
or field by collecting and supplying pertinent and specialized information to other
specified groups and individuals, and qualified and functioning in this role as a professional peer of these groups and individual&."
(DEFINITION I)
The clause within this definition that reads
" ... functioning in this role as a professional
peer ... ," calls for examination of one other
prime question. The question concerns the
resources the center must command to collect,
store, search, retrieve, and communicate at
peer level. To account for them, we suggest:
"The information center's proper technical
knowledgeability and subject range are determined by the technical competence of the
persons being served and by the nature of
the problems in which they are eng[,ged.
Its proper information processing and supply practices are determined by the knowledge patterns traditional in the special fields
represented among those it is serving."
(ASSERTION II)
It will be seen that a center satisfying
assertion II does not require those it serves
to learn anything about its information processing operations. This is possible because it
has peer competence in the technical subject
matter, and thus can "translate" its raw
retrieval output from "information system"
language to, say, coUoid chemist language. In
this respect, the center can make life distinctly easier for the person served than the
"pure" library we defined previously.
Assertion II also implies that working
knowledge of specialized information processing techniques is not an inherent responsibility of the bench scientist. As the opposite
side of the coin, it implies that there is a
distinctive field of expertise pertinent to information and communication processes. (We
have heard "information scientist" with
increasing prevalence, so we should perhaps
call it "information science" - and its applied
art companion, "communication technology".)
We believe that the rapidly advancing research
in machine and non-machine storage-searchretrieval methods already has presented substantial justification for this very important
claim. Machine data processing, for example,
may be a small part of the total information
art, but it is important in special areas and
certainly has demanded specialized training
far beyond the tolerance of the person who is
not a machine specialist.
The logical consequences of this line of
reasoning produce guide lines that we think
are of primary importance when one actually
gets down to the matter of designing or operating an information center. They include:
1. The designer should investigate the field
of knowledge patterns, competency levels,
personal attitudes, and working objectives of
the group he proposes to serve. From these,
he can derive the information service range
and techniques that make connections with
real customers at the other end of the line and
are neither over nor under designed to serve
them.
2. His information output should be purged
of any characteristics imposed by the center's
internal storage-retrieval techniques. (We
will concede one departure from this ideal,
when purging costs are greater than the combined effects of the educational costs and
communication losses imposed on those he is
serving.)
3. His staff must cbntain two professional
competencies: in specified technical subjects,
and in information science and communication technology. Key m~mbers of his staff
must be knowledgeable in both fields.
4. He should view his staff as prime terminal vehicles for conveying information to
the customer, and for accepting it initially to
the center. As information carriers, face to
47
face discussions (or ear to ear telephone conversation) are impressively distinctive from
the accession list, or even the individual
memorand urn. He should be most particularly
aware of the novel ways in which the human
can function as a component link of an
information service system.
HUMANISTIC FEATURES OF THE
TECHNICAL INFORMATION CENTER
'When "science" and "man" are mentioned
in the same breath, the first quick comparison
that comes to mind - particula.rly to the mind
of a scientist possessing the true faith - is
the relative fallibility of the human when
compared with the ordered beauty and reassuring reproducibility of scientific law.
Shouldn't we be doing our best to build
information services that exclude the human,
rather than making him the pivot component?
And isn't the added suggestion that the
service be derived from the human interests
of the group to be served really a sort of
ultimate confession that there is no discipline
or substance to this subject at all?
We believe neither is so.
For our first advocate of the human ingredient in the technical information service,
we offer the French philosopher-scientist
du Nouy. Du Nouy advances the essential
argument that "science" is inherently shot
through with humanism. "Science," he says,
"is only that portion of nature that humans
have been able to make sense of through a
process of rational ordering."2 If one follo.ws
du Nouy's basic contention to its application
here, the human is the most fully compatible
vessel for contending with scientific information, being its creator. Mechanisms (like
index cards or computers) may aid the human
but only rarely can carry value discrimination
beyond the moron level. And professional
scientists have only limited use for moronic
services.
For our second fundamental argument in
defense of the human ingredient, we offer
R. E. Gibson, the scientist-executive-philosopher who heads the Applied Physics Laboratory of the Johns Hopkins University. APL
originated and currently operates one of the
first modern technical information centers,
the Solid Propellent Information Agency.
Gibson asserts:
48
"Knowledge, the distillate of human experience, is stored in three types of banks: (a)
in the human mind and memory; (b) in the
literature - periodicals, books, reports, and
so forth; (c) in the products of technology
and culture, commodities, tools, services, and
organizations. Of these the only bank which
pays interest and offers capital gains is the
mind. The growth of knowledge is a function
of the capacity and the number of educated
minds engaged in its cultivation."3
So much for natural congruence of art and
discipline with the human mind: what about
human fallibility? To examine this question,
we propose to review the conditions and consequences one might reasonably anticipate if
he set up an information service.
Let us suppose you have been charged with
the responsibility of designing a technical
information service. Let us suppose further
that a real organization, or field of interest,
staffed with real scientists and engineers,
with real laboratories - even real libraries
and librarians - exists already, to use and
be used by this service you have been asked
to design.
If you subscribe, as the author does, to the
belief that an optimal technical information
service is a derivative of the population
it serves, the logical outgrowth of your
efforts should possess several gratifying
characteristics:
1. Your recommended design should relate
closely to the expressed interest of the population being served - which makes them
happy (at least before the actual service is
launched; it also helps in getting executive
authorization) .
2. In your design, you will find yourself
virtually forced to take account of the firm
data you were actually able to develop. This
stricture reduces the likelihood that the
"service" will really be an experimental
vehicle for one person's untested theories. By
no means, however, does it eliminate the
opportunity for imaginative or unconventional
techniques. It is more likely to stimulate innovations that have the notable added virtue of
utility.
3. Even if your survey technique is creditable, there is usually enough "slop" in the
returns that you can sneak a few of your pet
techniques into less critical areas of the de-
sign. This usually means that the exercise
leaves you happy.
So we finally get our service established,
through steps we know include human fallibility. What happens then?
The normal design errors in such a system
do not compound to absurdity or catastrophe
like a runaway computer. For the characteristic strength of the humanistic element now
comes into play. The staff members of a real
technical information center, being human
and able to recognize the technical shortcomings of initial operation, quite predictably
will react to, compensate for, and ultimately
correct the imperfections. Our basic strategy
is therefore quite sound. This organic vitality
of the modern technical information service,
manned by professional scientists or engineers, permits- indeed, counsels-acceptance
of design directives from a population that is
not sophisticated in information technique
per se. Just breathe life into the service by
endowing it with a well trained, service
oriented technical staff, and it will gravitate
toward its optimal role. It will do more yet:
it will follow the changing and evolving needs
of the population it serves as the whole
endeavor proceeds..
At first glance, one might conclude that an
information service· built on specifications laid
down by the user does not allow us to make
use of the more sophisticated information
processing techniques. This is not so. It does,
however, illuminate for us the sharp distinction that should be made between the methods
used within the service group and the techniques it employs in communicating with its
"clients." Internally, the search-retrieval system may be technologically formidable. Externally~ the output should be delivered in a
. format acceptable to the recipient. The
recipient (even though he is a scientist) turns
out to have human fallibility, too, and woe
betide the system that fails to accommodate
to him.
We might have a sense of despair on reaching the conclusion that all of science is really
just that slim caricature of all human experiences that we have found organizable. But we
might also view the emergence of the humancentered technical information center as
heartening evidence that we have at last
accumulated enough science that we can no
longer subject our precious hoard to the intellectual crudities of non-humanistic mass
processing techniques and still retain our grip
on it. Its growing richness has just begun to
make demands for something better - to be
specific, those store-search-retrieval schemes
that employ more effectively the subtleties and
discriminating powers of the human mind.
BENEFITS OF OPINION QUESTIONS
IN PRE-DESIGN SURVEYS
This leads us to a justification for soliciting
opinion as well as fact when one sets about
designing the technical information center.
I believe we should not take too much for
granted at this primitive tintype stage of
technical communication and the assistance
of human effort through institutional devices
such as the technical information center.
Therefore, I would suggest that you cannot
beat the technique of going to and talking to
the people who are working in the particular
field the center is to serve.
If one predicates that (1) the person interviewed understands what you are asking him;
(2) he is asked to answer only on behalf of
himself, or a small group with which he is
intimately associated; (3) he is made reasonably convinced that his personal practices,
preferences, and opinions are inherently
"right" answers to your questions; and (4)
a sufficient sample of the population concerned is polled to eliminate statistical uncertainties; we believe that a more complete and
reliable specification for the proper infor-·
mation service has been established than
specifications one can develop through other
means, and most particularly through means
restricted to non-opinion measurements. We
believe this because the working success of
the service ("center", etc.) depends so intimately upon subjective considerations, including such "unscientific" matters as the college
degrees held by the center's employees, the
official name of the service, and mass attitudes
of the group being served. One should make
positive efforts to draw out these subjective
factors, so there is some reasonable chance
of dealing with them effectively in the design.
These precepts underlie the specific techniques utilized in a recent study conducted for
the Office of Naval Research. A summary of
the survey objectives, the approach used, and
49
the service derived through the investigation
are given.
A CASE HISTORY: DESIGN OF A
TECHNICAL INFORMATION CENTER
SERVING THE UPPER ATMOSPHERE
TECHNICAL COMMUNITY4
Objectives of project
The ARIES project had the broad purpose
of determining whether a technical information service would contribute to current
scientific and engineering programs concerned
with knowledge of the upper atmosphere (the
altitude range 30 to 300 km). To accomplish
this purpose, it was necessary to determine:
1. The merit of a service
2. The optimal technical coverage
3. The service techniques and staff requirements, and
4. The estimated annual operating cost.
The survey technique
The investigation method employed had to:
1. Identify the members of the population.
2. Determine the technical information
usage of an adequate sample of the
population.
3. Determine the most acceptable service
techniques for serving it.
4. Ascertain the desire of the population
for a centralized technical information
service.
Depth personal interviews, coupled with
"~ross check" mail questionnaires on key
points of the survey, were employed to obtain
this information.
So far as possible, we chose persons at the
"group head" level for detailed interview
discussions, so a single interrogation would
produce an informed appraisal of the involvement of all persons in the group with
upper atmosphere data, and would yield a
working-knowledge opinion of the technical
information need.
Persons interviewed were .asked to answer
only on behalf of their own group, as we in
effect synthesized a total population answer
in our compilation of individual replies.
Size of population sample surveyed
The typical group head supervised three to
10 professionals. A single interview thus reflected the explicit practices and requirements
50
of perhaps five to 10 persons, including technicians as well as professionals.
The survey returns indicated that the following population sample was represented:
Sent
Interview check list
Mail questionnaire
Total
No.
No. Professionals
Respondents Represented
80
80
38
39
160
575
-735
77
In addition, relatively significant discussions
were held with approximately 35 persons under circumstances not resulting in a filled-out
questionnaire. (The distribution of their attitudes regarding a service appeared consistent
with the more formally obtained replies.)
Analysis of returns
Statistical tests were employed to determine the significance of the replies obtained
from the sample population, and to see
whether sub-groups in the population varied
in their information usage and their desire
for an information service.
The "do you want a service" answers in
the interview sample and the mail questionnaire sample were cross checked to see
whether these different survey techniques indicated significantly different sample populations. (The test showed a very high probability that the populations were identical.)
To facilitate the cross correlation process,
a Keysort card transcript of all mail questionnaire and interview check lists was prepared.
Its model was the brief mail questionnaire. Tests of more detailed questions were
made by hand compilation from the interview
questionnaire.
DERIVATION OF
OPTIMAL SERVICE DESIGN
Desire for an information service
The questionnaire data were analyzed to
determine the desire of the technical population for an information service. The results
were:
Desire for Service, % Desire
Highly More Less Rating"
---- --
Interviewees (33)
55
24
21
Mail questionnaires (34) ~ ~ ~
Total (67)
'Composite of multiple-choice replies rating
"'By ~ test.
Statistical
Significance""
1.67
Yes
1.65
Yes
1.66
YeS
from -1 to +3.
Subject mnge of the service
The data usage of the population was em-
ployed to establish the optimal subject coverage for the service. The results (10 subjects
with the highest use ratings) are given, in
rank order in the following table.
Inspection suggested a cut-off value of 0.8
in defining the initial subject coverage of the
service.
Subject Field
Use rating'
Statistically
Significant"
Density
1.84
Yes
Solar Radiation
1.4
Yes
Temperature
1.37
Yes
Winds
1.2
Yes
Pressure
1.1
Ionization
1.05
Yes
0.8 ••••••••••••••••••••••••••••••••••••••••.••••••
Water
0.63
Oxygen
0.63
Ozone
0.47
Nitrogen Oxides
0.52
'Composite of multiple-choice replies rating from -1 to +3.
"By t test.
Technical skills required
The subjects ranking above 0.8 in usage
value then were inspected to determine the
technical skills desirable in the service staff.
They indicate principal needs for professional
knowledge of:
1. Physics (especially radiation
phenomena)
2. Meteorology
3. Instrumentation (electronic and
mechanical)
Service techniques desired
Interview check list returns next were analyzed with respect to preferences expressed
for different information service techniques.
The results for the 10 service functions with
the highest ratings follow, in rank order:
Technique
Desire rating'
Statistically
Significant"
Yes
Loose leaf' manual
2.2
Data collection-compilation
1.9
Yes
Yes
Accession list
1.5
Abstract bulletin
1.5
Technical extract-reports
1.4
Newsletter
1.2
Report and document collection 1.2
"Man Friday"
1.1
1.0 '••••••••••••••••••••••••••••••••••••••••••••••
Language translations
0.85
Handbooks
0.67
'Numerical weighted average for population answering the
questions.
"By t test.
Inspection of the spread of desire ratings
suggested a cut-off value of 1.0.
These service preferences show a predominance of function requiring technical knowledge, and editorial rather than library skills.
The broad picture evoked by these preference rankings provides a rather explicit set
of directives for the personal and information
technique skills, staff activity patterns, and
physical resources of a service that will satisfy
the demand. They are itemized as follows:
THE SERVICE SHOULD PROVIDE:
1. A data collection-compilation activity,
with the compilations disseminated in
the form of "perpetual" manuals updated as appropriate with loose leaf
inserts.
2. Periodic announcements of new subject
knowledge sources acquired by the
service. Periodic accession lists- were
recommended.
3. A newsletter. (We recommended that
the newsletter function be provided as a
"Progress Highlights" opening section
of the accession list.)
4. The document collection necessary to
conduct these activities.
5. Technical reportorial coverage of meetings and conferences.
6. "Man Friday" information services by
staff members.
7. Technical editorial skills and personnel
with the professional capacity to deal
effectively on a person to person basis
with project scientists and engineers.
8. An appreciable travel activity.
THE SERVICE DOES NOT
NEED TO PROVIDE:
1. Conventional library services, such as
document loans, book purchases, reprint
ordering, etc.
2. Special correlations or mathematical
services.
3. Administrative support in organizing
meetings.
The operating service organization
An organization chart for an information
center performing these functions is shown in
Exhibit 1.
51
This pattern was chosen to provide a maximum assignment of sp~cific work responsibilities to individuals, yet retain necessary
flexibility for unusual work loads. Total staff
includes four senior and one junior professionals (p) plus a supporting clerical-secretarial
(s) force of four.
2. The population proved quite uniform in
its attitudes regarding a service, thus
producing key answers possessing high
statistical assurance from relatively
small samples.
3. This consistency also produced substantial concordance regarding the design
for certain services and subject coverages. In fact, almost no interpretive
judgments by the survey group were
required to reconcile the replies with a
set of realistic recommendations.
EFFECT OF CASE HISTORY
SPECIFICS ON GENERAL METHOD
We suspect the ARIES case history was inherently kinder to the surveyor than many
of the situations one might
be called on to
explore. To enumerate:
1. The sharp definition of the subject field,
namely, geophysical attributes of the
atmosphere and rocketsonde technology,
considerably eased the conception and
preparation of the "morphology list",
which we feel contributed considerably
in gaining professional acceptance for
the interviewer and technical consistency during interview discussion.
4. The service desires expressed were readily accommodated in an organization of
logicaI design and realistic size, manned
by persons with conventional discipline
training and career experience.
~
We may, of course, be leaning over backwards in this generalized judgment, in view
of our introductory emphasis of the humanistic underpinnings of scientific fields. But we
can readily visualize patterns of survey re-
Director
(p)
Information Technologist
(Preferably with background
in geo- and astrophysics)
(s) Secretary
~\
I
I
\\
\\
\\
\\
\
\
Document Operations Group
(p) Information Technologist
(skilled. in ~echnical indexing I--data storage-retrieval)
\
,
\
\
\
\
\
,
",
\
(p) Technical Information
Specialist
(meteorologist)
\
\
"-l:';;O~
,,
,
,,
,
&
,,
,,
(p) Technical Information
Specialist
(I nstru mentation and
Electronics)
,
,
SERVICE'
IL-----.J
EXHIBIT 1
File & mail-room clerk
Publications Group
(Data Manual, Accession list,
Newsletter)
~,
\
52
-1
,
(s)
(p) Technical Editor
(p) Compilation Assistant
(s)
Typist
----
turns that would be much more difficult to
cope with than ARIES proved to be.
Should we apply this basic survey technique
to a more refractory problem, we do have
available a few more tricks in our kit bag that
might spell the difference between a derived
design and blind guessing. For example:
1. Fields of interest might be approached
by asking more than one leading member of the survey population to outline
it for us, and then looking to see where
their replies coincide and diverge. In
some instances, a field is really a number
of "schools", loosely bound together by
a common purpose, perhaps, but with
distinctively different technical patterns. Where such characteristics are
found, some of the service functions may
properly be provided to certain parts of
the field, while others may serve across
the board.
2. Within limits, certain statistical assurances can be obtained by increasing the
survey sample size. Should an initial
sample show uncertain concurrence, we
can examine the return, and by risking
an assumption or two, get a reasonable
indication of the additional surveying
needed to clarify the ambiguities.
3. At times a particular service requested
is explicit and technically justified, but
the specialized training required and
the fractional manpower called for make
it clearly uneconomic to provide a properly qualified specialist in the centralized
service. In such an instance, we might
recommend that part time information
service duties be assigned to an appropriate specialist in the group requiring
the service. In general, peer competence
in the information organization facili-
tates practical solutions that are not
strictly observant of organization chart
lines.
Finally, we wish to urge this key recommendation on any person assigned the task
of designing an information service: begin
by just looking at the situation with the greatest perceptivity you can bring to bear. At the
beginning, try to avoid any preconception as
to the methods you will use to develop firm
guide lines for your design. For just as certainly as a sound design study can avoid unproductive costs in the operation of an information center, the initial natural disposition
of persons, attitudes, and activities usually
contains hints regarding particularly apt
design solutions. If you can see them, they
usually suggest singularly productive approaches to data collection, and stimulate the
development of imaginative and unusually well
fitting designs.
REFERENCES
1
Portions of this paper are reprinted from DeWitt O.
Myatt and Thompson E. Upham, "A Quantitative
Technique for Designing the Technical Information
Center", Journal of Chemical Documentation 1, 1824, 1961, by permission of the copyright owner, The
American Chemical Society.
'Lecomte du Nouy, Human Destiny, The New American Library of World Literature, New York, N.Y.,
1949.
3
4
R. E. Gibson, "Impact of Government Programs on
Growth of Knowledge", presented before the 134th
ACS National Meeting, Sept. 9, 1958, Chicago, Ill.
Project ARIES (Atmospheric Research Information Exchange Study), A Study to Determine the
Merits of an Upper Atmosphere Technical Information Service. Contract 3071-00, Office of Naval Research. (Grateful acknowledgment is made to the
Office of Naval Research, Code 416, for permission
to utilize the results of that study for this paper.
Information. and opinions presented here do not
necessarily carry the endorsement or represent the
viewpoint of the U. S. Navy.) Check lists and questionnaire responses of this survey are available from
the author on request.
53
By James Hillier
RCA Laboratories
8. Management's Evaluation of Information Services
My POSITION IS AS "devil's advocate" in evaluating information services from management's point of view. I represent the manager
who must pay the bill the best he can, watch
out for empire builders, and separate fact
from fad. We have to look at the forest instead of the trees. In giving management's
view on information service, technical information, and information retrieval, perhaps
I can help you in dealings with your managers
on similar problems in your operations.
First, let me express a note of caution. We
ought not rush ahead without giving more
thought to the overall systems we are concerned about. We need to look at some of the
critical problems. I am worried that, despite
all the advances in this field, we may inadvertently be building limitations into the
information system. These constraints may
limit creativity which is so important in a
research laboratory. I direct my remarks to
the system as used in a research organization
where, we must always remember, it is necessary to obtain results. This is an extremely
complex problem; I want to re-emphasize this
and to give a. somewhat different perspective
on the problem, from the management point
of view.
There are two basic aims in a technical
information system. One aim is to deal with
the information explosion facing the technical
staff. Management's problem here is critical:
determining the optimum distribution of manpower to meet the need. The second aim is to
provide management with the technical information it needs for decision making. Many
decisions have to be made on technical
grounds,both in government and industry.
This is an unfamiliar territory for many high
level managers. It, too, is related to the general problem of information services. Let us
consider both aims in more detail.
54
TECHNICAL INFORMATION FOR
THE TECHNICAL STAFF
The problem of providing technical information to the technical staff is the one most
people in the field are interested in; it's the
one I'm most interested in. Let us consider
the parameters we deal with in this problem
and how management looks at this. The basic
problem in almost any research lab is trying
to get productivity out of the lab, where productivity means new ideas, new developments,
and progress in a scientific sense. The manager of a research lab generally has a fixed
budget. Thus, optimizing the system ( and information services are very obviously an
exceedingly important part of the system)
really means: how do you distribute this
rather fixed manpower or dollars? (I shall use
manpower and dollars interchangeably because in a research lab they are interchangeable.)
When you optimize, you have to have a
measure. How do you measure productivity
or creativity in a research lab? There are subjective judgments that are surprisingly good
in measuring creativity, even though there
appears to be no way to put a quantitative
measure on it. If I select a group of people
that several of my lab directors know very
well and ask the lab directors, "Layout in
rank order the creative ability of these individuals," I will get a list from each lab
director that is almost identical one to the
other. There may be one or two people that
are out of place by one or two marks on this
ranking list. These lab directors somehow or
other know how to measure creativity in an
individual. Another parameter is indicated by
the fact that everyone of you probably has a
different idea of what is meant by "research".
Perhaps it is radically different from my ideas.
Here, too, we are dealing with a spectrum.
Creativity in a laboratory doing basic, exploratory research, as my own lab, needs
quite a different sort of measure than might
be applied to development laboratories at the
other end of the spectrum, where costs and
schedules, etc., are much more important.
Looking more specifically at information
services, one question is, "What kind of serve.
ice are you going to provide?" Again this
presents a whole new spectrum of problems.
One extreme is simply a traditional, limited
library that does nothing but provide references on call. The middle range is information
retrieval according to subject matter with all
its attendant problems. The other extreme of
service is provided by a highly technical group
that is going to feed chosen or specified information to your bench workers. You've got to
measure the situation; and then somehow or
other you've got to decide within this spectrum what is the best, what is the optimum
service to provide.
In the laboratories themselves there are
important variables related to information
services. What type of lab is it? Is it one doing
very basic research, applied research, or is it
the basic almost engineering type of laboratory? You have this spectrum to deal with.
There is another spectrum in the technical
sophistication of the group.
There are also two considerations that are
quite important from the management point
of view. One is the size of the lab. Obviously,
the budget that might be allocated to an in~
formation service is going to depend considerably on the size of the laboratory. A
second and perhaps more important variable
is the homogeneity of the lab. It is possible to
have a laboratory that is quite large, but in
an essentially narrow discipline; that is, a lot
of people doing closely related work in, say,
chemistry, or a limited field within chemistry.
At the other extreme is the very heterogeneous
lab where there are many different disciplines.
If you want to consider the limit, such a
laboratory would have individual technical
workers, each one doing a completely unrelated problem in different fields and different
disciplines.
Thus, we have a complicated system which
the mathematician would say is represented
by about a five dimensional matrix. Somehow,
the manager has to thread his way through
that matrix and find the optimum situation
for the particular laboratory he manages.
This is really an operations research job made
all the more. difficult because, while it is fairly
easy to describe these problems in qualitative
terms, there are no quantitative measures for
these variables. It's one thing to talk about
the size of a lab, but try to put a measure on
the heterogeneity of a lab. Try to define precisely in what part of the spectrum of
research any particular lab exists; within the
lab itself you'll find the spectrum. It's not an
easy problem.
When I go into the literature on the subject
of technical information services, I am concerned because I find a tendency for people to
pick out only a part of the problem and· make
a very intensive and very intelligent study
of it. There is nothing wrong with that, provided that you know that the problem studied
is only a specified part of the larger problem.
When I try to find discussions of this larger
problem, I don't find them very often. I have
tried to resolve this overall problem; to see
what the problem is, with what we are dealing, and what are some of the critical elements.
The matrix approach is too complicated. However, there are some boundary conditions,
some ways of describing the system, that will
give an insight into what the problem really is.
I feel quite strongly that it certainly is not
just satisfying the user's need for information. I'm not sure that the user's need for
information is what the user asked for. This
is quite important. In talking to my own
people, I get very vague answers as to what
they really need in the way of information.
In many cases, they havep't analyzed what
they need. Their information needs seem to
be determined by tradition, by custom and by
habit. I often wonder how many people have
really analyzed what they need. What form
would the information come in?
As has been indicated, I have the problem
of trying to relate the expenditure of
manpower/dollars optimally in this multidimensional matrix. I will go through the
parameters already mentioned in some detail.
When I place my laboratory in this multidimensional matrix, it is what I would call
a worst case. In system design, worst cases
are the extreme cases. I say that not because
I feel particularly sorry for myself, but be-
55
cause it just 'happens to come out that way.
First, we have a very large laboratory with
450 professionals which, with all the services
that go along with it, totals about 1,200 people.
It is a very heterogeneous laboratory. I find
that I have about 18 to 20 separate, isolated
fields represented in the laboratory; this runs
a span from metallurgy, to chemistry, to
meteorology, and so forth, including several
facets of physics which are essentially individual fields today. There are 150 clearly
identifiable projects; thus, there is an average
of only three people per project.
It is a very basic and exploratory laboratory. We do no development work; as soon as
we establish a principle, other parts of the
RCA organization do the development work.
Since our 450 people represent only five or six
percent of the professional staff of RCA, we in
this central laboratory are a very small element of the company. Yet, the level of
technical sophistication is very high. Unless
you get into some very basic or nuclear
physics, I doubt if there is a lab that has a
higher technical level. Obviously in this environment, I'm looking for new ideas, I'm
looking for new inventions - really inventions of materials or components rather than
inventions of systems and apparatus. My
basic criterion of evaluation has to be creativity. Of course, the decision I have to make
in this environment is the magnitude and the
nature of the service that I should provide
to this group.
means of a transmitter. The signal is transmitted over wires, radio, etc. At the other end
there is another transducer that converts the
electrical form back to sound waves, which the
ear can appreciate. This is the receiver. The
language that's used, of course, is a code. The
code goes into the brain, where the message
is disentangled. The expectation is that the
message that was disentangled in the receiving brain was the sarrle that was started out
in the other one. Quite often it is not.
CREATIVITY AND INFORMATION
But if it's a small number, and there is
essentially an infinity of information potentially available, obviously there must be some
selection process in the information system.
And selection is important when a human
being can take in information at only these
few tens of bits per second. There are only
twenty-four hours in a day, and humans can't
live in parallel; thus, everything feeds through
that narrow channel. Further, that's the rate
for the total information that he needs for
living, not just for the work he's doing,
Everything he does has to feed through that
channel.
However, before this can be done I must
tie creativity and information together. If
you don't have an understanding of this relationship, then this paper is meaningless. For
a more detailed discussion of this I would
recommend a paper that I wrote which gives,
I think, a rather new approach to the relationship between creativity and information
.flow.' I will now review this relationship between information and creativity, because it's
quite important relative to what people should
be . doing in providing information services.
Let us consider the analogy between a
human communication system and an engineered communication system. You're all
familiar with the telephone. The system starts
as a message in a brain. An encoder system
then puts the message into electrical form by
56
CHANNEL CAPACITY
However, the most important characteristic of any communication system is what a
technical man would call signal to noise ratio.
This in turn, through Shannon's theories, can
be translated into the channel capacity. In
other words, how much information can you
put through this system? It turns out that
any communication channel has a certain
capacity. You can put a certain number of
bits per second through that channel. The
human system is no different. It has a limit,
and that limit is surprisingly small. Various
people measure and estimate it, and it always
comes out in measures of a few tens of bits
per second. Information experts tell me that
in the usual presentation of a paper, allowing
for the redundancy of the languages, etc., information is conveyed at the rate of only ten
bits per second. The important thing is that
it's a very small number. It really doesn't
matter what the number is.
There are a lot of people writing, and still,
when an individual reads, he's only one individual. Even though it takes longer to write
than it does to read, it doesn't take long before
you have more people writing than one man
can read. That, in very simple terms, is the
information problem.
SELECTION PROCESS
When the amount of information gets very
large, there is another problem involving
probability which is the key to the whole picture. There's a certain element of chance that
the scientist will get certain pieces of information and not get others. Long before he
became a scientist or before he became a customer of an information center, the scientist
has learned through experience the most
probable channels that will get information of
value to him. The value of the information
is usually judged on how well it satisfies his
personal desires.
Remember that it is the individual who
controls the selection process. He has learned
it through his whole lifetime of experience,
and he controls it. You can't do anything
about changing it no matter what you try to
do. If you don't provide a higher probability
of value to that man, he's not going to pay any
attention to you.
In response to an overload, the human or
biological information system departs somewhat from the engineered system. When overloaded, the engineered system starts making
mistakes and the information is garbled. The
human, or biological, system is very accurate,
almost on a linear scale up to and near the
channel limit of the system. However, when
it is overloaded, it collapses. It quits. If you
were ever in an automobile accident, where
information started coming too fast, you'll
know what I mean by the information channel
collapsing. Perhaps nervous breakdowns occur
when we've exceeded the information capacity. I'm quite sure I exceed this capacity at
times, and I have to get back quickly on the
track and limit my input.
The important thing is that all of us at all
times are controlling this input. We've learned
to stay sane, if you will, by the simple process
of not overloading our channel capacity. But
we're always doing it by selection. This is
selection on both the output channels and the
input channels. You have to recognize that
you're dealing with this selection whenever
you start trying to provide any kind of information for an individual.
Let us return to the process of creativity.
Everybody seems to agree that a creative idea
comes out of one mind only when the right
pieces of information somehow coincide in
time in that mind. The "flash of genius" is
what the patent attorney calls it. I think any
of you who have been in creative pursuits
have recognized how things have been hazy,
and all of a sudden the right bits of information trigger off the idea and everything falls
into place. Thus, it is necessary to have the
right pieces of information in the mind, and
those have got to come into the mind through
this selection process, keeping the channel
from overloading.
This is the heart of the problem, because
you've got to get the right pieces of information into this creative mind in order to get the
creative act. But you don't know what the
right pieces of information are; you don't
know what information the man already has
in his mind; you don't know at what time
these will come together in the right relationship. Thus, you are back at the roulette table
and dealing purely with chance.
This is very important because the management of a research laboratory is always trying
to raise the level of creativity of the people
in the laboratory. You have to find some
means of raising the probability of a creative
act or a creative bit in that man's mind. But
as ~ou are trying to raise the probability, you
do not know, a priori, what information is the
right information, what time or how to give
it to him. Somehow you've got to find ways of
raising this probability, and it's not an easy
problem. If you read the paper referred to,
you will see a few ideas that we have been
trying out which have been working quite
wen.
The important point is.that this is a probability situation. In this type of situation, if
you superimpose one probability on another
the ultimate probability is always smaller.
It's impossible to have a probability greater
than one. This is one of the things that has
bothered me about the particular type of information service which selects or feeds the
so-called "right" information to the individual. I tend to look on such a service as
having a probability less than one of
transmitting the right information to the
individual. When you superimpose that probability on his probability of getting the right
channels, you have lowered the over-all
57
probability of the situation. There still is this
great mass of information potentially available and somehow you have to be able to pull
together the right bits which result in creative
ideas.
However, all this does not answer the
problem of how to' do this, and I cannot provide such an answer. It is important that the
existence of this probability be kept in mind
as various information systems are developed.
If this is considered as the key point in the
multi-dimensional matrix referred to, then
there is a chance of being able to work toward
a useful informatiO'n service.
TYPES OF INFORMATION SERVICES
Let us return to the needs to be satisfied.
I consider that there are five types of information needs in the kind of lab that we have
considered. To repeat, this is a large, heterogeneous lab, doing basic exploratory work,
where creativity is the important measure.
These five needs will be considered in order
O'f increasing importance.
PREVENTION OF DUPLICATION
The first is the need for avoiding excessive
duplicatiO'n of research, particularly in those
experiments which are destined to give negative results. We don't like to waste money on
projects that aren't going to go anywhere. If
people would publish the negative results as
well as the PO'sitive ones, we CQuld save the
country, the company and everybody a lot of
time and money. This is a conventional information retrieval problem. When you have a
specific project or experiment, it is a fairly
straightforward information retrieval effort
to find out what else has been done on this
specific experiment.
We do very often have the problem of
vocabulary changes. In electronics the vocabulary has been changing so fast that information gets lost because one doesn't know where
to IQok for it. This is a familiar problem.
Whether or not a specific information service
is needed to handle this is a question. You
certainly need the basic element, which is a
reference search system. Incidentally, at our
laboratory we have a good traditiO'nal library
which I consider to be the reference look-up
type.
58
SPECIFIC INFORMATION
OR DATA NEED
You also need to provide specific information that the technical staff requires. The
great mass of technical literature and patent
data contains the specific piece of information
that the man wants - it might be the thermal
conductivity of some special alloy - which
can usually be defined quite accurately. If it
exists in literature, it should be rather a routine job to find it. If he knows it's in a certain
paper, this simply requires reference 10O'k-up;
if not, then conventional information retrieval
may help.
CATCHING UP INFORMATION
The third most important item is "catching
up" information. We have a lot of prO'blems
today in research where the individual stays
current in a field fQr a relatively short length
of time. But if he's going to stay in research,
he has to catch up with sO'mething new. It may
not be often - every five or 10 years, perhaps.
However, when you need to do this it must
be done with the maximum efficiency. Again,
it seems like a fairly rO'utine, straightforward,
simple information retrieval job. It may even
be simpler than the previous information
retrieval jO'bs, because most of the time the
catching up information comes in textbooks
and review articles. These are useful until
you get almost up to date. I dO' not think an
information service group could help this
prO'blem unless you could build into them the
skill to select exactly the right material for
this indivdual who has a very specific back.ground and very specific experience. I'm not
sure this can be done.
CURRENT AWARENESS
The fO'urth item cO'ncerns an information
service to provide an efficient means for enabling the member of the technical staff to
keep current; this is the problem most people
are concerned about. What is gO'ing on in the
field today? What is cQming out in the literature? This is one that has always presented a
problem. It's one of the reasons for the abstracting services, but they, too, are becoming
too vO'luminous and too unwieldy.
To meet this fourth need there is a possibility of having an information service which,
at least in brO'ad categO'ries, is able to' select
and at least notify bench workers that certain
information is available. Just how big you are
willing to make this group becomes a complex
problem. I have made some very coarse estimates in my own situation. If you assume that
it takes maybe two or three people to be thoroughly acquainted with the field or all the
literature in the field, and allowing for some
overlap between fields, I estimate that I would
need to add 150 people in this information
service activity. This along with my staff of
450 would be 600 people. But I don't have
enough money to supply 600 people, I have to
stay with 450. This means I've got to cut back.
Of course, I could have 115 in my information
service group and only 335 on the bench in
the research activity. However, let us return
to this point later.
INFORMATION TO STIMULATE
CREATIVITY
This fifth need gets right down to the crux
of the information problem at a research
activity. There is need for an information
flow which will stimulate creative thought,
and you must provide this flow in a way that
will maximize the probability of having some
creative ideas. This is the basic reason for
doing true research, no matter where it is.
In a way, you could think of the four other
needs as being a part of this one - but this
last one is the crux of the matter. In considering my situation with 115 people in a technical information group, with 335 on the bench,
but where I have 150 projects, my question is:
"Is it better to separate this group out, or do
I leave them where they are?" I recognize
that if there is a separate group I must put
highly technical people in it. They have to be
just as good technically as the bench workers,
otherwise this isn't going to work. On the
other hand, the information group already
exists in my operation; they're already in the
laboratory as a part of each one of these little
groups.
I conclude in my particular situation that
it would be much better if I left them where
they are. The reason for this is that in being
closely associated with the work, the probability of their being able to add something that
wiII make that group more creative is much
higher if they were in a centralized group.
This is particularly important when there is
wide diversification. as in our laboratory. The
decision could be quite different in a different
situation. There are information systems in
other types of laborato~ies that differ from
ours. Yet they are probably good systems for
these groups. You have to look at the individual situation, and you have to think about
the probabilities.
There is another consideration very pertinent to having a centralized information
activity in a research laboratory. If I were to
pick out certain individuals from all these different small technical groups and put them
together in a technical information group, I
had better be very careful how I do that or
I am going to make second rate citizens out
of them. When these information people are
a part of the individual research groups, they
are on the same status as the technical worker.
If you get a research man of sufficiently high
level to do this technical job, he's as good a
man as the research man. But in a separate
group in a service role, our social pattern is
such that this information specialist may not
be as acceptable as the research worker. This
question of status is important, especially
when you need good technical people in the
information function.
SELECTIVITY OF CREATIVE PEOPLE
I picked out some of my most creative staff
and had a roundtable discussion one night on
this subject. I got some rather surprising
results. They didn't agree with any of the
points I have made but they said this is all
academic, because, they said, "We don't read
the literature," and, "We're so involved with
the work, we know where the good work is
going on."
In other words, these men have already
selected their communication channels on the
basis of the probable usefulness of the work
of various individuals in the field. They know
that one person is going to get good results,
and that another won't get anywhere. So they
watch the first man, and they don't bother
with the second one. They watch the good one
and when they see results coming, they communicate, they go to meetings, and they find
out what the results are long before th~se
results get into the literature. Even when they
go into the literature, they see what the good
worker has written, and they don't bother
with what the other man has written. There is
something to be said for this, when you have
59
creative people, and my lab is built around the
very creative people. However, I recognize
that the world isn't full of such people, and
you still have to provide for the people who
are going to read the literature.
TECHNICAL INFORMATION
FOR MANAGEMENT
When I examine the problem of technical
information for management, it appears to
be a separate problem from that of providing
technical information to technical people. The
management information problem has two
aspects. First, the president of any corporation is no longer making all his judgments
solely on a financial basis. He has to make a
great many of these decisions on the basis of
scientific data. This also applies to decisions
made in the government.
However, there is an interesting conflict in
this situation. The traditional system of a
management hierarchy depends upon information flow. It was made possible by the consolidation of information. The sales data for
a local field and local markets are consolidated; the production costs in the different
shops are consolidated. Other data are consolidated piece by piece as you climb up the
company management ladder. The information is thus condensed into a form that can
fit into the communication channel of the
manager; otherwise, if you gave him all the
bits and pieces of data, you would overload
his communication system. The whole system
has been built up because the information was
susceptible to consolidation.
But how do you consolidate technical information? At present, it just doesn't seem
possible. An example will illustrate this point.
Somebody goes out and finds that Joe Doakes
at 58 Main St., at Podunk Hollow, doesn't like
the color of the cabinet on his TV set, and he
reports this to the president of the company.
The president obviously can't run his business
on the basis of this individual fact. The president has some national market surveys made
so he knows in general what the population
likes and dislikes in colors. He also gets from
60
inside the company a distribution of the
colors that are in inventory. If this inventory
doesn't fit with the market survey, somebody
down the line is not a good manager,' and
corrective action is taken.
Although I have simplified the situation for
the example, this is basically the way the
president runs his company. It does him no
good at all to have the information that Joe
Doakes doesn't like tne color of his set. But
as far as technical information is concerned,
this is almost exactly what is being done.
Information retrieval systems are built so that
the president of the corporation can push a
button and get out an individual fact. But
how can he use that fact any more than he can
use the fact that Joe Doakes didn't like the
color of his television set? This inability to
consolidate technical information in a form
useful to management is a nasty problem.
There is a second aspect to this whole problem. The type of hierarchical company with
which I am concerned, with its pyramid of
management, is a selling-distribution system.
But the national government is a concentrated
customer; it works in reverse. With regard to
this matter of technical decisions, I am concerned about the inverted hierarchy which
seems to have developed. This is something
that the information service people should
give some real thought to. There is no question in my mind that a great deal of the future
of this world, this country and all the companies in this country is going to depend upon
major top level decisions which in the future
perhaps fifteen to twenty years from
now - are going to have to be based on
psychological matters and on technological
facts. We have to be able to get the facts to
these individuals in a consolidated form that
doesn't require that they have 10 years of
physics, 10 years of chemistry, or 10 years
of rocketry to understand them. This is a
major chaIIenge for all of us.
1
REFERENCES
"The Theory of Communications in a Research Laboratory," Industrial Research Institute, Vol. 3, No.4,
1960, Winter ed.
By Bernard K. Dennis
General Electric Company
9. Financing a Technical Information Center
INCREASING AWARENESS OF THE technical information problem and a better understanding
of the problem's nature and implications have
stimulated the development of a Technical
Information Center, in General Electric's
Flight Propulsion Division, tailored to the
needs of the business. Although originally
based on traditional library methods, experience has shown that these must be improved,
modified and added to in the light of newly
available techniques and concepts. Also, implementing an operating philosophy of self
initiated selective dissemination has required
a thorough, continuing analysis of the center's
functions, personnel requirements, costs and
p,erformance as related to both short and long
range needs of the business.
A brief review of the center's operational
philosophy, organization, personnel requirements, tools and services will help establish
the perspective necessary for understanding
the financial philosophy and procedures discussed in Part II of this paper.
I. THE FPD TECHNICAL
INFORMATION CENTER
The Technical Information Center provides
much of the technical information required by
the 2,500 to 3,000 scientists, engineers and
management personnel at the Evendale, Ohio
plant of General Electric's Flight Propulsion
Division. The center also extends its service to
several hundred other G. E. personnel in other
locations. Most of the center's users are engaged in applied research and development,
covering many technologies related to the
division's aerospace propulsion products with
major emphasis on materials.
Since 1957, a continuing effort has been
conducted to reorient, reorganize and restaff
the Technical Information Center to assure
its users more complete, efficient and reliable
access into the available information pertaining to their technical activities. Efforts to
establish an aggressive information program
have resulted in several new information tools
and services. Implementing the center's operational philosophy of self initiated selective
dissemination has required the initiation of a
broad range of services to meet specific needs.
AUTOMATIC INFORMATION
RETRIEV AL SYSTEM
The center's Automatic Information Retrieval System (AIRS) was designed and put
into operation in 1958 using an IBM 704 electronic computer and associated data processing equipment. During 1961, AIRS was
transferred to an IBM 7090-1401 computer
system and its programs modified to take
advantage of the 7090's greater capabilities.
Also, several important improvements were
made to the system's programs during 1961.
The Automatic Information Retrieval System is based on an inverted coordinate index
for searching, and features magnetic tape
storage of abstracts with bibliographic and
security data. The system can handle up to
1,300 multi term machine questions simultaneously. Maximum asked on a single run to
date is 265. There are now over 65,000 documents in the file, with approximately 1,000
new items added each month. Average depth
of indexing is estimated at between 20 and
25, with about 7,000 items in the system's
vocabulary. A run of 100 or more machine
questions may require two to five minutes of
mainframe time to perform all coordinations
and deliver access numbers. Or, if desired,
the system will automatically look up all abstracts with bibliographic and security data
in an additional six to nine minutes of 7090
61
mainframe time. The relationship of machine
search customer, literature searcher, machine
and center is illustrated in Exhibit 1. Typical
customer service cycle is 24 hours.
When AIRS was designed, the emphasis
was on information retrieval. However, as
the center learned more about its machine
system's large simultaneous search capacity, it
became apparent that selective dissemination
would be its primary role, with retrospective
searching a by-product.
physical sciences, fluid· mechanics, mathematics and computation, propulsion systems, management and administration, and
materials and processes.
At the center's present level of operation,
about 5,000 of these items are reviewed, and
out of these, 1,000 are selected and indexed
into the Automatic Information Retrieval
System for dissemination and retrieval. By
defining an individual's information needs
and setting up his interest profile expressed in
the retrieval system's language (key words),
he is periodically given a very selective look
at the center's indexed input of 1,000 documents per month, and automatically receives
abstracts of new material entered into the
file since he was last contacted.
Thus, in less than an hour's reading time
per month, a Current Awareness Service
client, in effect, scans 5,000 new documents
of potential value to him, carefully reviews
the contents of 1,000 of them and then focuses
his attention on only a few for detailed study.
CURRENT AWARENESS SERVICE
A current awareness program conducted
by the center attempts to keep specific individuals abreast of world-wide technology in
subject areas of immediate importance to
them. Other participants in the program are
attempting to keep informed on a broader,
longer range basis. The Current Awareness
Service principle is illustrated in Exhibit 2.
The Technical Information Center's monthly
input of potentially useful information is
estimated to be between 15,000 and 20,000
items in the form of reports, books, memoranda, technical society papers, journal
articles, patents, abstracts, etc. This input is
in more than 20 broad subject areas, such as
AEROSPACE MATERIALS
INFORMAT10N PROJECT
At the beginning of 1961, through close cooperation with the division's materials groups,
Exhibit 1
INFORMATION
OUTPUT
CUSTOMER
PERTINENT
ABSTRACTS
ONLY
LITERATURE
SEA.RCHER
TECHNICAL INFORMATION
CENTER
IBM
7090
62
LITERATURE
SEARCHER
the Aerospace Materials Information Project
was established within the center to better
meet the information needs of more than 125
materials scientists and engineers in six
materials laboratory groups in the Evendale
complex. The AMI project was staffed with
an experienced metallurgical engineer and a
chemical assistant. The project's. functions
are illustrated in Exhibit 3. To accomplish its
purposes, the project serves as a continuing
technical liaison between its materials clients
and the center. The AMI project has brought
to the center a far better understanding of
the division's materials information needs
than ever before possible. This is resulting in
a significant upgrading of the center's acquisition, indexing, dissemination and retrieval
of materials information.
Two major dissemination programs have
been conducted by the Aerospace Materials
Information Project for the past year. One of
these is a continuing alkali metals technology
survey conducted by a local consulting chemist
under the AMI project engineer's direction.
Results of the survey are indexed and selectively disseminated by the project to severaf
interested scientists and engineers in the
division.
The other continuing dissemination effort
conducted by the Aerospace Materials Information Project involves both the American
Society for Metals information searching
service and the Technical Information Center's Automatic Information Retrieval System
as seen in Exhibit 4. By correlating the major
information needs of its clients, the AMI
project has established a bi-weekly input of
abstracts from ASM. As they are received,
the ASM abstracts are screened and distributed to appropriate materials clients. On a
monthly basis, the Automatic Information
Retrieval System is used to selectively disseminate its current input in the same subject
areas to the same materials clients.
Exhibit 2
CURRENT AWARENESS SERVICE
TECHNICAL INFORMATION CENTER
ENGINEER B
SPECIALlS-T C
63
Exhibit 3
AEROSPACE MATERIALS INFORMATION PROJECT
DOCUMENTS
~
CONVENTIONAL LIBRARY SERVICES
UPON REQUEST
.................. '..
:
~.~~.~~.~~.~~.~~~~.ttrrttrt~~ ~.
,..·.·.·.·I·N·
:·:·:fE~HfJjt·AL·::::·
..··F··O····R·M····A····T-··I·~~·;.·.:.·.:.·.
::::::::::::::::.......
...• ...... •........ •· .. I...o...-SPECIAL
•••••
V 1'1
••••••••••
DOCUMENTS
.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:. ~
:....................
':':;';';':'C''-;;-~-;i''-'::::::'':::
':::::.
. . .: :. .:
. . . .:
1;1'1
........................ .
ER'::::"::
••
t :.:.:.:............: ••, •••: .............:.:.:.:
UPON REQUEST
•..AEROSPACE:·:·
••: :•• : : : : : : : :••: . .
.tt~~~~ttt~~~~~~ttt~I~~~
:........ AEROSPACE ':":.
SERVICES, - - ........................ .
-..
,.,.................
:':':':'::$U'~~i:i'~r:':':':':':':
:.:.:.:.:.:MATERIALS :.:.:.:.:
....................... ..
:·:·:·LABORATORIES .:.:
• • • ;0 • • • • • • • • • • • • • • • • • • • • •
:::~¥.f.:T.~~.It..~~ .::::
SElECT - 01 RECT -
:::::::::::::::::::::::::::::::::::::::::::::::::::
.....
...::::::.
GUIDANCE
(NF(;)~MA.'tioN r--.
~ ·......:.MANAGERS·:..: ..
:::::::F.O:~~:ti~hi~·:::::::::
:.:.:.:;·P...R·.XJ...E·.f.:T·:.....·..
CONNECT INFORM. :.:.:.:.:...:::..::::..: ....:::;.:.:.:.:
.:........,...........: ..:::........: •••:
I::,':.•..v ....'r ...:.:.:.:
;:::;·~·~~O~MAtl~:~::::: -T.I.c. PROGRAMS - -+ ::::::::::::::::::::::::::::::::::::::::
IN PUTS
.:.:.::.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
:::::::::::'5C if NTisTS ":::::::::
::::::::::::'~'i~t~'M~":::::':::::
SOURCES - NEEDS INTERNAL INFORM. -
.::::........: ..::::......: ......:::..
.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
PLAN N ING
.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:
.....
: ..: ......::..::............
: ..:...
•••••••••••••••••••••••••••••••••••
:..........
:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:
::::::::::::.:'.':.::.:::.:'.':.::.::.::.::'.':.:::.::.:
::.::.:'.• .•:'.': .:~.': .: .:..' •. ..:'.'::.::.::.::.::.::.::.::.:'.'::.::.::
•••.•.••••••••••••...
~A,PPLICATIONI--.-..
•••••••••••••••••• r
--'
.. ---..- ........................
••••••••.•••••••••.•••••••.
.........................:::........
EX PLOIT ATION
•.........•: ....................: •.•.•.: ....:.
........................
SERVICES - PLANS -
:::::::::::~.~~!t-!~.~~.~.::::::::'
:::::::::::::::::::::::::::::::::::::::
':':':':':':':':':':':':':':':':':':':'.
:::::::::::::::::::::::::::::::::::;::::
...........: ....: .............::::...
:::::::::::::::::::::::::::::::::::::::::::::::
POLICIES - ETC.
.:::::::::::::::::::::::::::::::::::::::::::::::
:::::::::::::::::::::::::::::::::::::::::::::::
................... .
:'
...•••••. '1
:':'
.
::::::::::::::::::::::::::::::::::::::::::::::::::
USE OF CONVENTIONAL
•••••••••• ••••• ••••••••••••••••.•.•
::::::.~5?.~.~.~.~I~t!~~.:::::
•.: ...•::..••••••: ..•.•.•.: ....••: ••.•: .•..
·lIBRARY FACILITIES
Exhibit 4
AEROSPACE MATERIALS INFORMATION PROJECT
T.I.C. INDEXED
INPUT 1000
DOC./MO.
SElECT -
DIRECT -
CONNECT
HOT ITEMS
1
AMI
AIRS
ASM
PROJECT
300 ABST RACTS
I--PER MO NTH
200 ABSTRACTS
INFORMATION
PER MONTH- SEARCHING
COMPOSITE INTEREST PROFILES
® ® ©
®
@
....
SERVICE
....
......
............
......
. . . . ....t
............
64
......
......
OTHER MAJOR SERVICES
1. Conventional library services
The center maintains two libraries in Evendale, with appropriate self service literature
research facilities including manual uniterm
coordinate indexes, conventional card catalogs, and 30 indexing and abstracting services.
The center's technical report collections contain over 100,000 reports, with security classification through secret R.D. The center
subscribes to 375 technical periodicals and its
book collection numbers over 20,000. The center's collections also include technical society
papers, U. S. and foreign patents and
translations.
2. Auxiliary document service
Several supporting services are provided by
the center to make its documents more available to requestors and potential users. For
example, through use of Xerox and microfilm
reproducing equipment, prompt, low cost,
appropriate photoduplication service is provided. Other auxiliary services include coordi-
nation and maintenance of the division's
internal report system, translations, and
special acquisitions.
3. Announcement bulletin
TIPS, the center's weekly bulletin, is distributed on a subscription basis. Issued 50
times each year for the past five years, TIPS
has consistently appeared at its subscribers'
desks on schedule. TIPS announces over 200
new items per week and accounts for 75 percent of the requests for specific documents
handled by the center's libraries.
As shown in Exhibit 5, TIPS is one of four
products obtained from the same punched
cards. This integrated approach to automation
has improved accuracy and has speeded the
publication of TIPS. In addition to substantial
cost savings, clerical time is freed for other
documentation work.
4. Special purpose retrieval systems
The Technical Information Center develops
and maintains small scale information retrieval systems to meet special needs. These
Exhibit 5
65
systems are physically located in the immediate areas of their users, along with the
special collections into which the systems provide access. Since documents indexed into the
small scale systems are also input for AIRS,
they are thereby available to all Technical
Information Center clients.
5. Information searching
Although the Automatic Information Retrieval System is the center's major IR tool,
it provides reasonably complete coverage of
only the division's internal reports. For more
exhaustfve, comprehensive searches, AIRS
serves as the first step. Information searching
is provided by the center's two engineers who
specialize in this field. Both specialists have
been in the division for 10 years in various
engineering activities and, as a consequence,
are well acquainted with the division's
products and key technical personnel.
6. FIND-X
A manual version of the Automatic Information Information Retrieval System, FINDX is. presently contained in four large "search"
books and 25 abstract binders. This manual
search tool gives immediate but limited access
into the more than 65,000 documents now
H OW
contained in AIRS. For those questions too
complex or time consuming to handle satisfactorily manually, a machine run can be
made. With FIND-X already in use in the
Small Aircraft Engine Department's library
in Lynn, Massachusetts, we expect to install
FIND-X in several other G.E. locations.
7. AutoCom
An experimental technical information dissemination tool under development by the
center has been dubbed "AutoCom", for automatic communication. The system utilizes an
electronic telephone answering device to collect input for the Automatic Information
Retrieval System. By using his phone (as
shown in Exhibit 6) - inside or away from the
plant; day or night - a scientist or engineer
may call in information he thinks may be of
interest to others, or that he may wish to
refer to at some future date. In response to
instructions he receives over the phone, the
information contributor identifies himself and
gives bibliographic data, an abstract and his
choice of index terms for each item he calls in.
The key to Automatic Communication lies
in the fact that when information is fed into
the retrieval system, the contributor's identity
A U TO COM
W 0 RKS
FROM OFFICE . . .
OR HOME . . .
INPUT
ELECTRONIC SECRETARY
FPD TECHNICAL INFORMATION CENTER
66
is maintained in association with his contribution. Thus, when that information appears
as machine output during either a retrospective search or a current awareness run, the
recipient of that particular abstract bearing
the contributor's name, title and extension
may be led to a person of related interest and
experience or, perhaps, to one who even is
actively engaged in related or overlapping
research. Hence, AutoCom may correlate
people as well as recorded information. Not
to be overlooked is the information selection
function served by such a device. With several
hundred scientists and engineers using AutoCom, the Technical InfQrmation Center's
normal acquisition efforts may be greatly
extended.
system work are needed to operate the center's systems and to perform research and
development to advance the state of the
information handling art for the center. A
more detailed view of the three capabilities
required by the center is given in Exhibit 7.
ORGANIZATION
The Technical Information Center's relationship to its primary user groups is seen in
Exhibit 8. By being administratively part of
the Applied Research Operation, the center
maintains close and continuing contact with
the longer range, more sophisticated information requirements of the division.
How the center has organized its functions
is seen in Exhibit 9. Ratio of professional to
supporting personnel is approximately one to
one.
PERSONNEL REQUIREMENTS
To provide the wide coverage of services
reviewed above, we have found that three different types of capability are essential. First,
to provide conventional library and auxiliary
services requires personnel trained in the
handling of technical. documents. Next, tQ
select useful information and to process it
requires personnel with competence in the
subject content being handled. Third, personnel trained in documentation and information
An information group with the nature and
scope of the one just described requires a
considerably higher level of financial support
by the user organization than that usually
accorded a traditional technical library providing passive, non-specialized technical document services. In Part II of this paper, the
FPD Technical Information Center's approach
to achieving financial support is presented.
Exhibit 7
FPD TECHNICAL INFORMATION CENTER
3
,
TECHNICAL
DOCUMENTS
• ACQUISITION
• PROCESSING
• INTERNAL
REPORTS
CAPABILITY
AREAS
,
MANAGER
,
,
I
I
I
TECHNICAL
LIBRARIES
INFORMATION
SYSTEMS
AEROSPACE
MAT'LS INFO.
TECH. INFO.
APPLICATION
CONSULTANTS
• SPECIALIZED
INFORMATION
SUPPORT
• LIAISON
• MGM'T
• EVALUATION
• TECHNICAL
SUBJECT
AREAS
• LIBRARY
SERVICES
• LIBRARY
FACILITIES
• AUXILIARY
SERVICES
• R&D
• OPERATE
• PROCESS
SElECT
INDEX
ABSTRACT
• LIAISON
• GUIDANCE
• FORECASTING
• SALES
• SPECIALIZED
SERVICES
• SEARCH
TECHNICAL
DOCUMENTS
INFORMATION
SYSTEMS
TECHNICAL SUBJECT
MATTER
67
II.
FINANCING THE FPD TECHNICAL
INFORMATION CENTER
Over the past four years, the Technical
Information Center of General Electric's
Flight Propulsion Division has attempted to
develop it method for maintaining a level of
operation realistically related to the information needs of its users, the expanding parameters of the technical information explosion
and the needs of the business. Even though a
permanent long range funding plan has not
yet been fully defined, certain basic indications have begun to emerge.
The first and most significant indication is
that the level of funding which can be expected as a direa-t charge against general
overhead or administrative costs is not likely
to exceed the requirements for a bare minimum of basic library services. Therefore, in
the long run, the more refined and specialized
technical information services must in some
way be supported by a direct charge against
the specific using group, whether or not the
using group is part of the local organization.
In other words, customers must pay for the
services.
This means, of course, that the Technical
Information Center must act to a certain
extent as a business entity, complete with
continuing problems of customer satisfaction,
sales volume and seeking out of new markets.
I t is obvious that the financial as well as the
technical success of the center depends upon
its consistently "delivering the goods".
BALANCE OF SOURCES
In its efforts to achieve a more secure financial basis upon which to operate, the Technical
Information Center has attempted to develop
three IT_ajor sources of income. We believe
that by distributing our "risk" across a
broader base, a greater degree of stability
should ensue. Whether an optimum balance
(Exhibit 10) among the sources may exist, is
not clear. If so, we have not yet found a magic
rule for achieving it.
However, in view of the continuing technical information explosion, with all of its
ramifications and implications, and the increasing complexity of the division's aerospace propulsion systems products, it seems
reasonable to expect that the center's users
Exhibit 8
FLIGHT PROPULSION DIVISION
SMALL
AIRCRAFT
ENGINE
DEPT.
AIRCRAFT
ACCESSORY
TURBINE
DEPT.
FLIGHT
PROPULSION
LABORATORY
DEPT.
LARGE
JET ENGINE
DEPT.
Serving the Division's
applied research needs
TECHNICAL
INFORMATION
CENTER
68
Serving the Division's
information needs
should provide the largest share of its support
in the form of direct payment for specialized
information services. This may tend to offset
short range fluctuations in that part of the
center's income from general overhead, thereby enabling the center to maintain a longer
range, higher level of technical effectiveness
during local business recessions.
Sources among which the Technical Information Center is attempting to achieve a
balanced support are:
1. Management subsidy
This old, traditional support for technical
information services is, to whatever extent
it may exist, usually a secure base since management generally feels the need for some sort
of library or technical information service.
However, with only this single source of support, the Technical Information Center is
subject to the individual views of a relatively
few people who, generally speaking, are
directly benefited by the center to a far lesser
extent than the scientists and engineers at
the technical work level. Thus, when funded
solely as a direct charge to overhead, the center may be particularly vulnerable during
short term business fluctuations, with the
immediate need to reduce overhead taking
precedence over the longer range benefits of
an effective technical information service.
Because of the intangible or unmeasurable
nature of the center's influence on the profit
making capability of those it serves, it becomes an early target in a management drive
toward overhead reduction. This is, at least
in part, due to the fact that a scientist or engineer will get as much information as he can
- one way or another - with or without the
assistance of an information group.
Unfortunately, immediately increased costs
and longer range penalties in duplicated
effort and reduced productivity incurred in
so doing are usually not apparent, particularly
in a large, complex organization. Therefore,
reducing the center's level of operation
may not produce any immediate, clearly
discernible ill effects.
With management subsidy as its sole source
of support, the center is handicapped in its
efforts to keep pace with the increasing scope
and complexity of the technical information
service problem. It may be virtually imp os-
Exhibit 9
FPD TECHNICAL INFORMATION CENTER
I
Secretary
MANAGER
I
I
I
TECHNICAL
DOCUMENTS
Supervisor
TECHNICAL
LIBRARIES
Supervisor
-
Acquisition Specialist
CONSULTANTS
Managemen
Operations
Research
{
--Chemist
other Division technologies as req'd.
I
I
I
INFORMATION
SYSTEMS
Supervisor
AEROSPACE
MATER'L INFO.
Project Engr.
TECH. INFO.
APPLICATIONS
Specialist
librarian
-
Stenographer
IClerk
r-- librarian
-
Information Systems Engineer
il-Clerk
~
-
Indexing and Vocabulary Specialist
f--
I
Clerk
'-- Engineering Assistant
Keypunch Operator
--
r- T.I.S. Reports Clerk
f--
Technical Information Specialist
-
i--
Technical Information Specialist
I--
General Reports Clerk
Computer Programmer
69
sible to negotiate the funding required for
advancing the state of the art in tools and
services. This is because of the severely limiting traditional ratio of information service
expenditures to all activities of the business
components it serves.
Thus, although management subsidy is
generally a secure source of financial support
in that it is seldom completely withdrawn, it
can also be widely fluctuating and severely
restrictive.
2. Direct sales - internal
We have attempted to initiate the more
refined and specialized information services
on a "will it sell?" basis. This means that
once a new service has been launched, it must
provide visible value compatible with cost in
terms of specific results to the buyer.
There are distinct disadvantages to selling
technical information service. For example,
more effort is required to provide service on
a direct sales basis. There is a fragmentation
of effort, since effort which might have been
spent in providing the service is expended on
selling, billing and customer relations. Also,
funding is piecemeal and not guaranteed. This
generates a certain element of uncertainty in
planning manpower and expense budgets.
As might be expected, making a direct
charge for a specific service reduces its utilization. This may be economically desirable,
since level of the service will more likely be a
function of real need or benefits. However,
this means that the center will need to provide
a greater range of services and that it must be
prepared to adjust quickly to keep its services
more nearly tailored to the requirements of
its users.
Exhibit 10
FINANCIAL
SOURCES: Managem't
Subsidy
BALANCE
Internal
Sales
External
Sales
1.
33%
33%
33%
?
2.
50
25
25
?
3.
60
30
10
?
4.
?
?
?
DISTRIBUTE RISK -
70
SUPPORT
-
INCREASE STABILITY
There is a major advantage to providing
information services on a direct sales basis.
By broadening the center's market - if the
services sell - a more balanced income is
attained. The center has a hedge against
economic fluctuations since it is unlikely that
all components buying its services will drastically reduce their purchases at the same time.
3. Direct sales - external
A third source of revenue available to the
Technical Information Center is the sale of
services or work to entities, industrial or nonindustrial, outside General Electric. Although
the center is primarily an "in house" engineering service, it may provide certain
services and development work for others to
the extent that in so doing, it strengthens its
capabilities (or at least does not diminish
them) to perform its internal mission.
Contracts with organizations outside General Electric may involve the development,
installation and operation of information
handling systems, information systems studies
and experiments,and routine services such as
information searching or current awareness.
Due to several factors, there is a relatively
limited external market available to the
Technical Information Center. Geography
may tend to restrict the center's activities to
a significant extent, since only a small portion
of the aerospace or related industries are
within a reasonable service radius. Problems
associated with handling proprietary and
security classified information also tend to
restrict the center's freedom to operate in an
external market. And, of course, our first
consideration is to General Electric. This has
not yet been a deterrent factor in external
efforts, but in certain instances it could
become one.
A natural hazard in developing an external
market is the probability of major fluctuations
in demand. This confronts the center with
problems related to building up for a large
contract and letting down after its completion.
Our progress toward achieving a broader
funding base is indicated in Exhibit 11. We
believe our rate of development has been
slower than it should have been. Yet, we have
made significant progress. Barring unforeseen
occurrences over which the center has little
or no control, we expect to make more rapid
progress toward our goal of financial stability.
Exhibit 11
INCOME
100 -
~
80 PERCENT
TOTAL
INCOME
DISTRIBUTION
TREND
~ANAGEME~T
SUBSIDY
~
--- -----
60 DIRECT
SA\ES
40 -
INITIATED
SALES
CONCEPT
20 -
~
\
~
-----
~
0-
1956
57
58
How to achieve balance in activities as well
as a stable financial base is an extension of
the balance problem. In other words, how can
the center expend the effort and funds necessary to advance the information handling art
in ways that will prove useful to our customers, while providing services only to the
extent that they are self supporting? It is the
old "which came first - the chicken or the
egg?" story.
There is no predetermined formula to
achieve balance in either support or activities.
At its present stage of development, the FPD
Technical Information Center has not yet
achieved the established status which would
permit a high degree of administrative, long
range planning. In lieu of this, the modus operandi which appears nearest to being effective
consists of keeping fully aware of all of the
above factors, coupled with a good degree of
flexibility and rapid response to meet short
term requirements and to take advantage of
opportunities as they are recognized. This
does not prevent us from having long range
goals as to professional standards in the
technical information field in keeping with
economic justification.
59
60
61
62
OPERATING COSTS
In arriving at its annual operating expense
forecast, the Technical Information Center
attempts to determine what is needed to
operate in terms of its current workload. In
arriving at· a distribution of costs, full consideration is given to the "axiom" that people
without tools are hardly more useful than
tools without people. Thus, every attempt is
made to arrive at the most reasonable fund
allocation in the center's budget (Exhibit 12).
Exhibit 12
Technical Information Center
OPERATING COST BREAKDOWN -
1961
SALARIES (lNCL. BENEFITS) .. 55.7%
ACQUISITIONS ........... .
DATA PROCESSING
5.6
11.5
OFFICE SERVICES
3.6
PRODUCT INFORMATION....
1.5
TRAVEL, LIVING ......
1.2
TELEPHONE, TELEGRAPH.
1.1
ALL OTHER VARIABLE. . .
5.8
RENT, DEPR., AMORT.
PERCENT
OF TOTAL
COST
12.2
71
Establishing a budget and securing its
acceptance by higher management are two
different things. We have found no secret path
to successful budget negotiations. To the contrary, as shown in Exhibit 13, we have followed
a rocky road in our efforts to finance the center. We have learned to accept unexpected
and sometimes drastic deviations in support
as a normal operating hazard of a Technical
Information Center serving a defense contractor - particularly when a disproportionately large share of the center's income is in
the form of management subsidy.
The center's continued progress toward a
higher level of operation is due mainly to two
factors. First, customers have proved willing
to pay for information services, thereby providing the center with income to help offset
drops in management subsidy. Second, exploitation of automation has raised the center's
capabilities to provide more specialized services for a larger market. In fact, the center is
now headed toward the point where it would
be impossible to provide its services manually
without adding people.
Exhibit 13
TREND IN FUNDING (EXPENSE)
+50%-
~~
J
40
PERCENT
VARIATION
-
A~
/ \
DIRECT
SALES,
30
The trend in automation expense incurred
by the center is shown in Exhibit 14. It should
be noted that 1961's increase was due primarily to a heavy programming effort required
to transfer the center's Automatic Information
Retrieval System from an IBM 704 to an IBM
7090-1401 computer system, and to modify its
programs to take advantage of the 7090's
greater capabilities. Also, i!\everal significant
improvements were programmed into the
system. Effects of these improvements are
reflected in the lower 1962 computer expense
forecast. In spite of the rapidly growing
AIRS file and increased use of the system,
1962 total computations expense is expected
to be 10 percent or less of the center's total
expenditures.
A major financial problem faced by the
Technical Information Center is how to finance
advancements while operating on a tight
budget. The implications of this with respect
to management subsidy are obvious. If management doesn't See the need for advancing
the state of the art of information handling,
then development of systems is definitely
If~
~
~
/
Y \
V
\ /
/ \ \ 17
20
~i\
10
1956 BASE
10
-
20
-
~
\
V
J
V
MANAGEMENT
SUBSIDY
30
-
40
-
\~ /
-50%-
1956
72
57
58
59
60
61
62
slowed. Individual users are unwilling to
gamble on the development of new services.
They are only interested in paying for ones
already established and proven.
INCOME
Thus far, the Technical Information Center
has achieved only limited success in its efforts
to establish a broader financial base. In order
to achieve what we believe to be a more stable
balance, two major jobs need to be done:
1. We need to secure more outside business, both of a development nature and
services.
2. We need to increase internal direct sales,
primarily in the more advanced or specialized service areas.
The center's direct sales of services in 1961,
as shown in Exhibit 15, were substantial but
below target in most areas. There were several
mitigating factors, the chief one being an 11
percent reduction in management subsidy, at
the end of the first quarter, which was necessitated by certain defense contract reductions.
Because of reduced funding and a highly unstable internal market during 1961, the center
held up staffing at 20 instead of moving ahead
to a forecast manpower level of 24. This, of
course, restricted services to some extent.
During the latter half of 1961, the center
began actively seeking business outside General Electric. We believe we have a unique
and marketable situation consisting of an
experienced, well staffed Technical Information Center productively operating in a live,
industrial R&D environment, using a large
general purpose computing system as a major
information handling tool and with its well
developed Automatic Information Retrieval
System in practical operation. We see this as
a practical development and proving ground
for technical information systems. It is our
plan to develop outside markets for both development effort and services. How rapidly
we can accomplish this - we wish we knew!
SELLING AND PROMOTION
Thus far, we have spent insufficient effort
on selling. We have been mainly occupied with
building up the center's staff, and with developing our tools and services while gaining
sufficient acceptance by management to move
ahead. We have done considerable selling, but
not enough.
In selling our technical information capabilities and services, we have attempted to
develop sales and promotional plans much as
is done for any business, except for two major
Exhibit 14
AUTOMATION
EXPENSE
TREND
12 -
11PERCENT
OF TOTAL
T.I.e.
EXPENSE
10 -
9-
/
8-
Includes: AIRS fperation, mainterance,
programmirg, spec. documerts
c~rculation
Plus: library
records, TIPS,
catalog an~ loan cards, misf"
7-
I
1958
'59
'60
'61
62's
FORECAST
73
distinctions. First, the unusual nature of our
merchandise, technical information, must be
kept in mind. Unlike a hard goods product,
benefits usually cannot be readily identified or
measured. Need for and use of the product,
technical information, is a very personal matter which may make clear recognition of
benefits impossible.
Also to be kept in mind are the peculiar
professional inhibitions of our te.chnical customers. They do not like a lot of ballyhoo or
a hard sell approach in our promotion or sales
efforts. Generally speaking, they are not susceptible to normal forms of advertising and
promotion. Our product has no utility to them
as persons outside their work. Their attitude
is such that to gain their respect and acceptance, we must prove to them that information
handling has the stature of a profession. They
traditionally look down upon a library as
being a non-technical activity which can be
helpful. when asked to do so and if given
careful guidance. The concept of a Technical
Information Center staffed with technical
people and aggressively selling. ,technical
services is new to them. Thus, in attempting
to promote itself and its services, the center
is bucking many obstacles ~ not the least of
which is tradition.
III. ·FUTURE
A number of implications for the Technical
Information Center's financial future have
emerged from our experiences. thus far. First,
it is to be expected that information services
and methods for providing these services will
require continuing review and adjustment if
the center is to adequately cope with the
technical information explosion and the
changing needs of its customers. The center's
work will become increasingly technical, requiring a corresponding increase in the
capabilities of its technical staff.
This gradual upgrading of the center's staff
and services means an increase in operating
expense which, of course, will intensify the
support problem. It is extremely probable that
maintaining an adequate level of management
subsidy will always be a problem, since in the
final analysis, internal sales are an operating
Exhibit 15
% OF TOTAL SALES
1961 DIRECT SALES
%
DEVIATION
*GOAL
**ACTUAL
28
24
14
AEROSPACE MATERIALS INFO.
22
17
23
INFORMATION SEARCHING
25
24
>4
INTERNAL REPORTS
12
11
8
SERVICES:
SPECIAL ACQUISITIONS
5
..
20
TIP S
3
2
33
PHOTOREPRODUCTION
1
1
0
INDEX CARDS
4
4
a
100
87
13
PENALTIES (Mat'l Lost, Overdue)
*Forecast based on T.I.C. staff of 24
**Accomplished with a staff of 20
74
expense to the business. The solution lies in
achieving and maintaining a high level of
management understanding and acceptance,
so that its level of support will be more closely
related to the information needs of the business and less affected by opinion as to what
the business can afford to pay.
Another implication for the center's future
is concerned with changes in the requirements
for personnel working in the technical information service field. Today, most information
workers fall into two extremes. At one end is
the theoretician and at the other is the conventional library functionary. Both are
essential to the field. In addition, however, we
need people· in between, who might l;>e called
"information engineers" and "information
technicians", who can develop the skills required in the practical application of
information handling technology.
A fourth implication for the center's future
operation concerns planning. Due to the
nature of the Flight Propulsion Division's
business, financial planning in the Technical
Information Center is of necessity short term,
usually one year or less. This means that management subsidy must be negotiated, and
challenging but attainable goals for direct
sales must be established, on an annual basis.
The only long range planning that can be
done by the center is concerned with the
achievement of a state of readiness to take
advantage of opportunities. This infers that
the center's planning must aim toward:
1. More sophisticated and flexible systems
which are proven and ready for
expansion.
2. High caliber personnel trained in both
information systems and subject content
areas.
3. Acquiring a detailed knowledge of the
information requirements of present and
potential customers.
4. Keeping fully informed of the professional resources in applicable portions of
the information field - services, people,
technology and equipment.
IV.
CONCLUSIONS
Operating an industrial Technical Information Center is both a special profession and
a business function, and has to be treated as
both at once. To plan and direct the center's
programs and activities, one must be a professional in the technical information handling
field and a business manager at the same time.
We have found no magic formula for relating the many variables pertinent to the
operation of the Technical Information Center. The situation is complicated more than
may be generally realized, with many "new"
problems peculiar to information handling
and information services in addition to the
"old" administrative type problems common
to most businesses.
However, the center's most crucial problem
is how to maintain a stable level of support,
realistically geared to the needs of the business and obtained from a balanced variety of
sources. Progress toward solution of its many
other problems is, in one way or another, a
function of how well the center finds long
range solutions to its financial problems.
75
· By HelenL. Brownson
National Science Foundation
10. Documentation Needs
I SHOULD LIKE TO BEGIN by explaining briefly
why I agreed to consider the documentatioI).
needs of scientists. I feel that an explanation
is in order because I am not a scientist and
can say nothing; therefore, a.bout the subject
on the basis of any personal experience. Nor
can I say much at all about documentation
needs as such, for we have very little precise,
objective knowledge about them. It is because
of this lack of knowledge that I welcomed the
opportunity to discuss the importance of
learning more about scientists' documentation
needs.
Since "documentation" is a word that means
different things to different people, it may be
helpful if I explain that I use the word
broadly to include all the processes involved
in preserving, disseminating, and gaining
access to recorded information. By "documentation needs" I mean the needs of scientists with respect to effective use of recorded
scientific infohnation. Scientists have other
information needs, of course, that are met,
and perhaps could only be met, by conferences,
visits, and other means of direct communication. We have yet to determine the actual
significance of documentation in the total
communication process in the sciences. Undoubtedly we would all agree that, since
science is a cumulative endeavor, access to
recorded information is vital to scientific
advance. But for a number of reasons we must
work toward a deeper understanding of the
whole research process, of the role of documentation in that process, and of the real
documentation needs of scientists:
1. We need a more rational basis for reaching decisions as to the relative importance and utility of different types of
existing and proposed services, for
choices must be made from time to time
- in the Federal government, in the
76
of
Scientists 1
scientific societies, and elsewhere concerning the expenditure of available
funds and skilled manpower on information services.
2. We need a more rational basis also for
choosing the most productive areas for
experimentation, in order to try to improve the flow and accessibility of
information.
3. We need to be able to assess the results
of an experiment with a new information service or system in terms of the
needs of its users.
4. We need much more precise knowledge
about needs, in order to design improved
retrieval systems that will meet those
needs.
I do not mean for a moment to argue that
we must fully understand documentation
needs before we try to improve existing services or to design or try out new services. On
the contrary, one of the best ways to learn
more about needs may be to experiment with
new or modified services, such as a new means
of disseminating research results, a new type
of index, or a new searching service, in order
to find out if they do, in fact, meet a need and
prove to be truly useful.
I plan to summarize briefly the growing
interest in the problem of information needs
as exemplified in three conferences. I shall
then discuss the studies that have been made
of the way scientists now communicate and
use information. Finally, I shall mention some
relevant studies now in progress or being
planned.
CONFERENCES
The first conference that was concerned
with the information needs of scientists and
the uses they make of information, was the
Royal Society's Scientific Information Con-
ference in London in 1948. 2 In preparation
for that conference, some studies of the use of
scientific information were made; and one of
the conference recommendations invited the
Royal Society to consider the initiation of
further research into the uses of scientific
literature. I am not aware that this recommendation directly stimulated any subsequent
studies in the United Kingdom following the
conference, but both the recommendation and
the studies prior to the conference stimulated
interest in the United States in further studies
of the use of information."
The second major conference concerned
with this problem was held a decade later in
1958, namely the International Conference on
Scientific Information in Washington, sponsored by the National Academy of SciencesNational Research Council, the National
Science Foundation, and the American Documentation Institute. 3 One session of it was
devoted to the "literature and reference needs
of scientists". The papers discussed in this
area of the conference were representative of
the types of studies that had been made during
the decade since the Royal Society Conference.
They included studies of the time devoted
by chemists .to communication activities, the
use of scientific literature and reference
services in Scandinavia, the use of information in an atomic energy establishment, the
use of technical literature by industrial technologists, the information gathering habits of
medical scientists, the u;;e of scientific periodicals, the planned and unplanned communication activities of scientists in several
disciplines, the ways in which scientists learn
of work important to them, and the determination of information requirements from
reference questions asked of scientific libraries. All but one of these studies were of
information activities and uses, not of needs.
From a careful analysis of their results, one
might hope to come up with inferences or
hypotheses about needs; but these should be
tested experimentally before they are accepted
as valid. The only study dealing specifically
with requirements was the analysis of reference questions asked of libraries. From such
a study, however, one learns about needs as
they are made known to libraries and not
necessarily the ,real needs, for it is quite likely
that scientists tailor such requests to fit what
they think the libraries can do for them.
In the conference discussions there appeared to be tacit agreement that the needs
of users and the uses to which they put the
information should determine the design of
scientific information systems and services,
and that we must learn more about needs and
about factors affecting the efficiency of scientific communication. It was admitted that
study of the information requirements of
scientists is difficult, for the communication
of scientific information is extremely complex.
Needs vary with the subject field, the type of
research, the availability of information
services and source materials, geographical
situation, and differences in human abilities.
In fact, the needs of a single individual may
vary from time to time, depending on his
current interests and activities. Furthermore,
the real needs of scientists and their conscious
or expressed needs may be quite different.
It was suggested that all completed studies
of information use be carefully reviewed,
evaluated, and compared. This suggestion was
acted upon and a review was prepared. We
will return to this review later in this paper.
A number of suggestions were made concerning future studies of information uses
and needs - namely, that effort be made in
the future to obtain qualitative judgments as
to the value of information, in addition to
quantitative data on the use of it; that psychologists be consulted in planning future
work because their knowledge of human perception, motivation, capabilities and limitations is relevant to the whole communication
problem; and that oral and informal means of
communication be studied in order to make
meetings more effective.
The third conference was a small meeting
late in 1960, devoted exclusively to the subject
of information needs in the field of psychopharmacology.4 Some 30 specialists in psychology, psychiatry and pharmacology were
called together to discuss informally their
own information needs. A few documentation
experts were also invited to take part in order
to discuss technical advances in the field of
documentation and to learn at first hand about
the needs in psychopharmacology. According
to my notes, at one time or another the scientists mentioned the following needs - it is
a long list but it indicates the variety of needs
this particular group was concerned about.
These are not in order of importance, for 'no
77
priorities were mentioned:
Current information about who is doing
what
Directories of various kinds about people
in the field
Information about meetings scheduled
A centralized collection of transcriptions
of the proceedings of meetings
Higher editorial standards for primary
journals
Better access to government reports
Better access to the patent literature
Access to information on negative results,
provided they are obtained under proper
conditions
Prompt information about new books and
about reviews of them
Specialized bibliographies adequately
indexed
A central source of reprints or photocopies
of journal papers
More and better critical reviews
Up to date information about reviews in
preparation
More up to date textbooks
Exhaustive handbooks of data
Better correlation and interpretation of
data
A central service for preparing translations
on request and furnishing translators and
interpreters for international conferences
Greater specialization of libraries so that
particular libraries can build up comprehensive collections in their special fields
More trained library scientists to help researchers
A searching service to identify all publications and documents relevant to any
particular question or topic.
I feel certain that this informal discussion
of needs was helpful to the persons concerned
with planning and providing information
services for psychopharmacology. It seems to
me, however, that in order to determine the
actual needs in any area it is essential to
consider each need and each information
service as a part of the whole and to reach
some conclusions about the relative importance of each with respect to the others. I
cannot imagine that there will ever be enough
money and skilled manpower to provide all
78
the information services that might be desirable if there were no limit to our resources;
so I think we shall always be faced with the
necessity of making choices.
During this discussion, as in many others
in recent years, the scientists frequently expressed the hope that machines could be used
to advantage to handle information. But generally speaking; scientists do not seem to feel
any responsibility for helping to outline in
some detail precisely what a mechanized
information system would be expected to do
for them. To design a system that would meet
their needs, the designers should know the
sorts of questions the scientists would put to
the system, as well as, the types of responses
wanted. It is not enough to say, as they often
do, that they will want everything relevant to
any stated question without specifying the
probable types of questions.
Moreover, as the volume of recorded information grows larger, one wonders if scientists
really will want everything relevant to a
question. It may be that they will find the
results of a mechanized searching system
helpful only if some way can be found of
insuring that the system does not burden
them with a lot of material that may be
"relevant" but of little value. In a recent
paper on the state of the art of information
retrieval,S Don Swanson states: "Clearly in
a situation in which the volume of information exceeds the individual scientist's digestive
capacity, some method of assisting him to
discriminate between the important and the
unimportant literature can be counted among
the most crucial of requirements."
STUDIES OF INFORMATION USE
AND PRACTICES
I want to discuss briefly the studies that
have been made of scientists' information
gathering behavior and their present communication practices. It is extremely difficult
to summarize what has been learned thus far
from these studies. Some of them have been
conducted within individual organizations, or
for the purpose of learning about the use of
a particular publication or service. These
studies have usually produced results tliat
served as a basis for decisions about information services within the organization or about
the future of the particular publication.
For example, the study of physics abstract-
ing in 1950 was a survey of a large number
of U. S. physicists to learn more about their
use of the journal, Physics Abstracts. 6 Its
nndings resulted in some recommendations
regarding coverage and indexing, and in certain decisions about American financial support of this journal which is prepared and
published in England. It led both to the increased use of author abstracts in American
physics journals and to the acceptance by the
editors of these journals of responsibility for
the adequacy of the abstracts. These changee
helped to speed up publication of the abstracts
in the abstract journal.
Because it was restricted to a single type
of information service, however, this study
did not shed any light on the relative importance of abstracts in the whole complex of
existing and potential information services in
physics. An example of a study within an
organization was the study of scientists' use
of library materials in the United Kingdom
Atomic Energy Authority in 1957, which resulted in a number of recommendations concerning the library services available to
scientists within that organization.7
Some of the studies, however, have had the
broader purpose of gaining insight into scientists' present information practices and
problems. All the studies supported thus far
by the National Science Foundation fall into
this category, and they have all had the principal objective of experimentation with a
particular methodology. One study, conducted
by the Department of Agriculture, employed
a diary technique. 8 Its results indicated that
this technique, as it was used in this study at
least, may not produce the complete record of
information activities one would hope for,
probably because it places too great a burden
on the participating scientists who must keep
detailed diaries.
A study at Columbia University, in which
faculty members of three scientific departments participated, made use of extensive
interviews in depth covering use of the literature and informal means of communication
as wel1. 9 An attempt was made to learn about
specific instances of failure to locate existing
information that would have been useful, and
about the circumstances surrounding the
receipt of unlooked-for pieces of useful information. The principal result of this study was
a long list of questions or topics that are
believed to require systematic investigation.
Still another study, conducted by an opera-·
tions research group at Case Institute of
Technology, experimented with the use of
observational techniques for determining how
chemists spent their time and the use made of
publications by chemists and physicists. I 0
The results indicated that the average chemist
working in industry spends more time in
scientific communication (that is, in reading,
writing, listening, and talking about scientific
matters) than in all the rest of his activities
directly concerned with science. Of a total of
32.4 hours per week, both during and after
working hours, devoted to scientific activities,
16.5 hours (just over 50 percent) were spent
in scientific communication.
Such findings have reinforced our belief
that it is important to try to find ways of
making scientific communication more effificient. It appears that an observational
technique such as employed in this study
produces reliable data on what scientists do
but it must be admitted that this method alon~
tells us nothing about why they now do what
they do, or what they would do under other
circumstances. The logical next step is to
develop hypotheses as to the services that
might enable scientists to spend more productively the time they devote to communication
activities, and then to undertake controlled
experiments to test these hypotheses.
A review of the studies in this field, some
20 of them, was prepared for the National
Science Foundation by Herbert Menzel of
Columbia University in 1959. 11 It displays the
diversity of approaches that have been used
in studies of the behavior, habits, experiences,
and expressed needs of scientists with respect
to scientific information. It also serves to suggest the very large variety of topics that can
be studied. By and large, the results of the
different studies made in different places and
for a variety of purposes were simply not
comparable for a number of reasons: the
populations studied were diverse; the units of
observation or recording were different; communication media and channels were classified
differently; and few studies included careful
analysis and interpretation of the collected
data.
The reviewer stated that he believed the
time had come to devise new research
79
strategies, and he suggested some neglected
approaches that seemed to him to be both
feasible and promising, such as studies focusing attention on the functions of the scientific
communication system and the ways in which
each is being met.
There is not sufficient time to go into the
details of the comparative results of the different studies brought out in Dr. Menzel's
review. 1 shall give one illustration to show
what such a comparison produces. In five
diary studies and two interview studies,
scientists were asked how they learned about
the publications they read or found useful. 12
The results appear to add up to an indication
that routine perusal of the literature and pure
chance accounted for the largest percentage
of materials read, ranging from 14 to 70 percent, an average of 38 percent for these seven
studies. Citations in other publications was
the next highest category, averaging 11 percent. Personal recommendations averaged 10
percent; previous use of memory averaged
nine percent; abstracts or indexes, six percent; initiative of the library, five percent
and result of a search, six percent. This last
category includes the use of a library catalog,
personal index, bibliography, or other reference work, and may overlap the "abstract or
index" category.
Although I have figured these averages to
show how one would thereby rank the various
sources calling items of information to scientists' attention, I do not believe that any conclusion should be drawn from such averages.
Some of the studies covered all materials read
and others covered only materials found to be
useful or significant. Some of the studies were
restricted to different subject fields. Some
were of senior scientists only, and one was of
technologists only. So to draw any conclusions
from their composite results, it would be
essential to consider the findings with all the
obstacles to comparability in mind.
At this point, you may very well be asking
if studies of information use are worth while.
Dr. J. D. Bernal, in his paper for the 1958
conference,3 stated succinctly both the reason
for and the limitations of such studies. He
said, " . . . enquiries as to present uses of
scientific information services, though a necessary background, can by themselves tell us
little of use for improving the service. They
. tell us what people do with an admittedly very
80
imperfect service, not what they would do
with a better one (which would naturally
include proper training for its users)." I
would agree that such studies, if they produce
reliable data on the use and lack of use of
existing services, are a necessary background,
making it possible to determine later how
much the situation has been changed, and
hopefully improved, by new or modified
services.
With respect to the lack of use of existing
services, perhaps some of you have seen
"Mooers' Law", which was formulated by
Calvin Mooers, a documentation specialist. 13
He pointed out that as we furnish users with
more and better retrieval systems, we cannot
be assured that they will make greater'use of
them. He phrased his "law" this way: "An
information retrieval system will tend not to
be used whenever it is more painful and
troublesome for a customer to have information than for him not to have it."
He went on to justify his assertion as follows: "Having information is painful and
troublesome ... If you have information, you
must first read it, which is not always easy.
You must then try to understand it. To do this,
you may have to think about it. The information may require that you make decisions about
it . . . The decisions may require action in
the way of a troublesome program of work,
or trips, or painful interviews. Understanding
the information may show that your work was
wrong, or ... needless. Having information,
you must be careful not to lose it. If nothing
else, information piles up on your desk unread ... Finally, if you do try to use the
information properly, you may be accused of
puttering instead of working. Then in the end,
the incorporation of the information into the
work you do often may not be noticed or
appreciated. Work saved is seldom recognized.
Work done - even in duplication - is well
paid and rewarded." Perhaps we should also
try to develop an understanding of the effect
of this so-called "law", which, it seems to me,
is only partly facetious.
CURRENT STUDIES
As for, current studies in this field, I sh:all
mention only those of fairly general interest
and of broader scope than "use" studies alone.
The major trend at present seems to be toward broad studies of communication within
major scientific disciplines, in which studies
of information use are only a part. The American Institute of Physics for the past two
years or so has been studying a number of
publication problems and the utility of various
types of indexes in physics. At present, they
are devising techniques for determining some
of the parameters of an efficient information
retrieval system for physicists. More specifically, they are planning an extensive survey
to accumulate information about questions
physicists would put to an idealized information retrieval system for physics.
vidual had made or had had made for him questions about the length of time he could
have waited for the relevant references, the
need for current or quite recent documents,
the preferred form in which information
about relevant documents could be supplied,
and the amount of time he is willing to spend
weeding out irrelevant material. The results
indicated that a wide array of requirements
would be of some importance to the performance of a retrieval system serving the
population studied.
The American Psychological Association is
undertaking a broad analysis of communication in the field of psychology, in order to
learn more about the conduct of psychological
research and the behavior of the researcher,
especially as it relates to needs for information services. They are planning a number of
separate but related studies of psychologists'
use of information; of the cost, readership
and other aspects of psychological journals;
of the adequacy of abstracting and indexing
services in the field; and of effectiveness of
communication at meetings and conferences.
IMPORTANCE OF DETERMINING
DOCUMENTATION NEEDS
The American Institute of Biological
Sciences has established a biological sciences
communications project to study and improve
the flow of information in the biological
sciences. As a part of this project, plans are
being made for an intensive study of the
research process, and of the role of information exchange in that process, in a number of
biological laboratories.
The National Science Foundation is supporting all these current studies I have mentioned, and is encouraging other scientific
societies and groups of related societies to
take a fresh, and a critical, look at their
present communication practices and
information needs.
One other study we have supported recently
should be mentioned here. The Stanford Research Institute, in the course of an exploratory study of criteria for evaluating the
performance of information systems, has developed and tested an interview guide for
determining more specifically than has been
done heretofore, the requirements of users of
a retrieval system. A "critical incident" technique was used in which specific questions
were asked about the latest search the indi-
Even though we have as yet barely
scratched the surface in gaining real insight
into the process of communication in the
sciences and the documentation needs of
scientists, it is encouraging that a number of
scientific societies are now engaged in studies
that should add to our understanding, and
encouraging also that recognition of the need
for such understanding is growing.
In a recent, extensive survey of the "state
of the library art", Margaret Bryant, the
author of the portion on bibliographies, abstarcts and indexes,14 stated, "We can say
without fear of contradiction that we know a
good deal more about our techniques than we
do about our objectives." She urged that research to determine objectives and the needs
of users be given the highest priority.
Don Swanson, in his paper on the state of
the art of information retrievaI,S pointed out
that the studies of scientists' communication
habits have thus far been "descriptive rather
than diagnostic", and that there have apparently been no reports of "controlled experiments which exert a perturbing influence on
the scientific community." He went on to say
that "It is premature to attempt to translate
a description of present behavior into future
requirements; to discern any relationship
between the two requires considerable further
study." He indicated that he thought emphasis
should possibly now be shifted to experimental studies subject to greater control.
In a recent book entitled Science Since
Babylon,15 Derek Price, a historian of science,
devoted a chapter to the "diseases of science",
- its superabundance of literature, its manpower shortages, its increasing specialization,
81
and its tendency to deteriorate in quality.
These so-called "diseases" all stem from the
exponential rate of growth of scientific activities and publications, a growth which, he says,
is "very much more active, much vaster in its
problems, than any other sort of growth happening in the world today." Price argues that
"there should be much scope for scientific
attack on science's own internal problems, yet
curiously enough," he goes on to say, "any
attack is regarded with much skepticism, and
the men of science prefer, for the most part,
to talk as unskilled laymen about the general
organizational problems with which science is
currently beset." I imagine Price would regard
the current studies of communication I have
mentioned as encouraging signs that scientists
in some fielqs are concerned with one of
science's internal problems, namely that of
management of scientific information.
In closing, I should like simply to emphasize
that the problem area I have discussed, that
of determining scientists' documentation
needs, is wide open for new approaches
and for soundly conceived research and
experimentation.
REFERENCES
'Most of this paper was presented previously at the
International Conference on Scientific Information
in the Field of Crystallography and Solid State
Physics, Nishinomiya, Japan, October 3, 1961, and
appears in the proceedings of that conference.
2The Royal Society Scientific Information Conference, June 21 - July 2, 1948: Report and Papers
Submitted, The Royal Society, London, 1948.
3Proceedings of the International Conference on
Scientific Information, Washington, D.C., ·November
16-21, 1958 , National Academy of Sciences-National
Research Council, Washington, 1959 (two volumes)
82
'Information Service in Psychopharmacology: Proceedings of the Conference on Scientists' Needs for
Inf01'mation, November 25-26, 1960. Prepared by
Psychological Research Associates, June 1961, for
the Scientific Information Unit, Psychopharmacology Service Center, National Institute of Mental
Health, Washington 25, D.C.
sDon R. Swanson, Ramo-Wooldridge Division, Thompson-Ramo-Wooldridge, Inc., Canoga Park, California), "Inf01'mation Retrieval: State of the Art",
Presented at the Western Joint Computer Conference, May 9-11, 1961.
6Dwight E. Gray, "Physics Abstracting",. American
Journal of Physics, vol. 18, October 1950, p. 417.
7I. H. Hogg and J. Roland Smith, "Information and
Diterature Use in a Research and Development
Organizat'ion", Proceedings of the International
Conference on Scientific Information (reference 3)
pp. 131-162.
"Ralph R. Shaw, "Pilot Study on the Use of Scientific
Literature by Scientists". Mimeographed report to
the National Science Foundation, 1956.
"The Plow of Information Among Scientists: Problems, Opportunities and Research Questions. Submitted to the National Science Foundation by the
Columbia University Bureau of Applied Social
Research, May 1958.
,oAn Operations Research Study of the Scientific
Activity of Chemists, November 1958. An Operations
Research Study of the Dissemination and Use of
Recorded Scientific Information: I - Journal Reading by Physicists and Chemists; II - The Cost of
Journal Publication; and III - The Effect of Condensation on Comprehension of Journal Articles,
December 1960. Both submitted to the National
Science Foundation by the Operations Research
Group, Case Institute of Technology.
"Review of Studies in the Plow of Information
Among Scientists. Submitted to the National Science
Foundation by the Bureau of Applied Social Research, Columbia University, January 1960.
'2Ibid., Table F-20.
""Mooers' Law, or Why Some Retrieval Systems Are
Used and Others Are Not", American Documentation, vol. 11, no. 3, July 1960.
14Margaret S. Bryant, "Bibliographies, Abstracts,
and Indexes", volume 2, part 2 of the series, The
State of the Library Art, Rutgers University Press,
New Brunswick, N.J., 1960.
,sDerek J. de Solla Price, "Sc';ence Since Babylon'~.
Yale University Press, New Haven, 1961.
By William B. Kehl
University of Pittsburgh
11. Communication Between Computer
and User in Information Searching
THIS IS A REPORT of an educational experiment
at the University of Pittsburgh Computing
Center. I believe it may offer several suggestions to large organizations which have a
centralized large scale computing center, on
how to take advantage of the availability of
such a facility.
We have a tape oriented IBM 7070/1401
system at the center. The tape oriented configuration determines the flow of activity, as
shown in Exhibit 1.
The IBM 1401 computer is used only to
transfer information from cards to magnetic
tape and, at the completion of the task, from
tape to printer. Not even editing is done on the
1401. All computation is carried out on the
IBM 7070 system proper. Such a system pro-
vides very high efficiency of operation since
the 7070, which serves as the central processor, is never "waiting" to read cards or to
print results.
Such an operation is typical of most large
computing centers. The question that immediately arises is, how does the user communicate with such a system? This problem is
solved in general by the use of a monitor
system computer program. We refer to our
monitor, somewhat facetiously, as PEST, Pitt
Executive System for Tapes. It controls the
sequencing of programs, one after the other,
automatically, allowing a mixture of student
homework, graduate theses, faculty research,
and production type runs without interruption. Some of the runs use a compiler language
Exhibit 1
Flow of Data in Tape Oriented Configuration
4 SWITCHABLE
DATA
MAGNETIC TAPE DRIVES
IBM
CARDS
IBM
1401
COMPUTER
IBM 7070
COMPUTER
PRINTER
OR
PUNCH
6 MAGNETIC TAPE DRIVES
83
like Fortran, some are in machine code or an
assembly language, and some are just the
execution of a program in our computer program library. Approximately 150 different
runs are executed each day.
One expects a scientist or an engineer to
adapt to such a technique of operation. But
we also wanted our computer to be used by
schools other than the natural sciences. We
wanted to allow the professor of education to
be able to analyze the vocabulary of first
grade readers, the student of Middle English
to use computer aids, and the lawyer to make
similar use of a computer.
It was in this latter area that our programming received the greatest impetus. We have
transferred to magnetic tape over 10,000,000
words of statutory materials from selected
states. This material was prepared by Mr.
John Horty, Director of the Health Law Center,.at the university. Free or natural text was
typed using Friden Flexowriters to prepare
paper tape, which was then converted to
punched cards. There are about 10 words on
each card. There are no deletions. The sentences average approximately 18 words long.
All material was punched as complete sentences. Each statutory section was treated as
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
CREAM, BUTTERMILK, AND ALL OTHER FLUID DERIVATIVES OF MILK.
"MILK PRODUCTS" MEANS ICE CREAM, ICE CREAM MIX, CUSTARD ICE
CREAM, FRENCH ICE CREAM, FROZEN CUSTARD, AND OTHER SIMILAR
FROZEN PRODUCTS. AND ALL DAIRY PRODUCTS USED IN THE MANUFACTURE
THEREOF •
• ,CANNED MILK., MEANS CONDENSED, EVAPORATED OR CONCENTRATED
MILK IN HERMETICALLY SEALED CONTAINERS OR FOR MANUFACTURING
PURPOSES.
"CERTIFIED MILK" MEANS MILK FROM DAIRY FARMS OPERATED IN
ACCORDANCE WITH THE .,METHODS AND STANDARDS FOR THE PRODUCTION
AND DISTRIBUTION OF .CERTIFI~D MILK" LAST ADOPTED ~Y THE AMERICAN
ASSOCIATION OF MEDICAL MILK COMMISSIONS INCORPORATED", AND
THE PRODUCTION AND HANDLING OF WHICH SHALL BE CERTIFIED TO BY A
COMMISSION INSTITUTED IN COMPLIANCE THEREWITH.
"SECRETARY" MEANS THE SECRETARY OF AGRICULTURE OF THIS
COMMONWEALTH, OR HIS AUTHORIZED REPRESENTATIVE.
,.PERSON., INCLUDES SINGULAR AND PLURAL. MASCULINE AND
FEMININE. AND ANY INDIVIDUAL. FIRM. COPARTNERSHI~, lNSTITUTION,
ASSOCIATION. OR CORPORATION THEREOF.
Exhibit 2
Sample Input Text
Exhibit 3
Vocabulary
ACCESSIBLE
ACCOtH:lDATIONS
ACCOMPANIED
78.3.024,8.5
ACCORDANCE
9.3.010,4,52
102.3.033,14.35
ACCORDING
69.3.015,3.134
84
28.3.024,9.18
112.3.014,5.13
1.3.043,14.26
78.3.031,10.7
1.3.015,7.10
11.3.007,4.22
118.3.006,4.24
61.3.014,5.70
79.3.008,3.62
1.3.047,16.18
79.3.023,7.5
2.3.005,3.29
17.3.010,5.12
53.3.009,5.5
79.3.030,9.9
4.3.007,3.46
71.3.007,3.48
77.3.021,6.5
64.3.011,4.49
91.3.011,5.24
64.3.026,6.75
91.3.016,6.56
65.3.011,4.13
9.3.006,4.21
88.3.080,19.23
a separate document, and the documents were
assigned numerically increasing document
numbers. The documents average 200 words
in length.
Since we knew that the user might wish
to identify a word within a sentence within
a document, we assigned a numerical identification to each word. Exhibit 2 shows the text
of a typical document as stored on magnetic
tape. This document was entered as document
number 1. Exhibit 3 shows that the word,
ACCORDANCE, appears in document number 1, line 15, sentence 7, word 10, matching
the reference in Exhibit 2. A computer program
to prepare a vocabulary tape containing such
a concordance listing was written by the center. Thus, for each of approximately 20 states
there is a text tape and a vocabulary tape.
While the text tape is complete, common
words of no interest have been eliminated
from the vocabulary tape. These account for
about half of all the occurrences of words.
But let us return to the topic of primary
interest. How does an English teacher, lawyer,
or professor of education communicate with
the computer to use such text data? Suppose
he wants to find the documents dealing with
"food poisoning". Does he have to learn the
intricacies of computer coding?
The usual answer to this is that he has to
depend on some expert. He may hire a computer programmer, or he may even go further
and seek the help of a professional library
scientist skilled in all the devices of indexing.
In the spirit of a university atmosphere, we
felt there should be another answer for him.
He should be given aids so that he could
communicate with the computer directly himself. In this way he would be able to do his
own study and not be dependent on "experts".
For this reason, a User Oriented Search
Language was developed. One example will
serve to illustrate its use. Suppose I were
interested in the regulations concerning food
Exhibit 4
Search Request on Food Poisoning
REGULATION
FOOD
POISON
ILL
SPOILED
HOTEL
AA
AA
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
D
D
+
IFD+O+IO
D
IFD+O+I0
D
IFD+O+IO
0
REGULATIONS
PENALTY
PENALTIES
FINES
SUSPENSION
SUSPENSIONS
FOODS
FOODSTUFFS
FISH
MEAT
SEAFOOD
SEAFOODS
PorSONING
POISONOUS
ILLNESS
SICK
SICK
SICKNESS
SPOILING
SPOILATION
DECAYED
HOTELS
RESTAURANT
RESTAURANTS
EE
BB
MM
CC
JJ
GG
JJ
AA
BB
CC
EE
GG
MM
+6
CITE
+4
CITE
+2
PRINT
85
poisoning. Several concepts come to mind.
These can be identified by actual words that
appear in the texts.
AA The concept of food or fish or meat
or seafood as represented if one of
these words appears in a document.
BB The concept of poison or poisonous.
EE The concept of spoil or decayed.
GG The concept of a hotel or restaurant.
JJ
The concept of regulations or penalties or suspension (of a license).
MM The combination of the concept labeled AA, above, and the concept
labeled EE, above, in the same document would insure that we were
interested in spoilage of food. To
provide this, we wish to require that
the document to be retrieved would
contain at least one word in group
AA and at least one word in group
EE. We let MM serve as a label for
this group of documents. Now if there
were less than, say, 10 documents
dealing with the concept MM of
spoiled food, we would like to print
them out. If there were more than 10
such documents, we might wish to
insist that there be a reference to
concept CC (sickness) in the same
document and we would like a citation
to these.
Exhibit 4 shows how such a search request
is prepared by a user - namely, the lawyer,
himself.
+ is used for or
D is used for "and in the same document".
IF D+O+ 10 +6 means, "if the number
of documents in the list collected up to
this point is between 0 and 10, then
skip 6 lines to the line which has the
word PRINT and execute that command".
Cards are punched just as shown in Exhibit
4, one card per line. A header card identifying
the program to be used is also added, and this
small batch of 35 cards is fed into the system
shown in the flow chart in Exhibit 1. Preceding
this problem in the card input stack may ,be
an engineering problem, and following it may
8,6
be a business school inventory model study.
The output is .shown in bhibit Sa through Se.
A few comments will complete our report:
1. Notice the use of the IF statement provides a means for the user to control his
answers, conditionally dependent on
some intermediate results. This is a
necessary and really a key factor in communication between the user and the
computer. It avoids the necessity of his
returning many different times, dissatisfied with getting either too many irrelevant documents or having restricted his
search by the excessive use of the D or
"and" command and getting no output.
2. Because the search language is trivial
to learn, every user may phrase his own
search. His satisfaction is infinitely
higher. Personal participation means
much to him compared with dependency
on an "expert".
3. His learning curve on his own work goes
up very fast. He is quick to perceive his
own successes and failures. He comes to
understand his own use of language and
the language of his subject much quicker.
Many other computer programs for analysis
of natural text have been developed by our
center. Some of these are described in the
Communications of the Association for Computing Machinery, September, 1961, issue. But
the one idea I feel is an original contribution
to the area of information retrieval has been
in the introduction of a User Oriented Search
Language and the communication link it
provides a user with the computer.
The programs described in this paper were
written by Charles R. T. Bacon of the computing center staff, and have been made available through the GUIDE library by contacting
the IBM Program Librari&n, IBM, White
Plains, New York. Mr. Bacon's work has been
a significant contribution to the use of our
center in new areas beyond the usual scientific
applications.
The legal retrieval applications research
under the direction of Mr. John Horty is supported by the Council on Library Resources,
the National Institutes of Health, the U. S.
Office of Education, and the Ford Foundation.
Output of the Search Request shown in Exhibit 4
Exhibit Sa
87
Exhibit 5b
88
Exhibit 5c
89
Exhibit 5d
90
Exhibit 5e
91
By H. P. Luhn
Consultant
12. Automated Intelligence Systems
is engaged in
an effort toward a common goal, their success
depends to a significant degree on the facility
with which they are able to communicate.
Not only need information be exchanged
within the group but it is essential that access
be had to past and current external information that may have a bearing on the activity
of such a group. The quality of performance
of the group will furthermore depend on the
speed with which information is able to flow
through their communication network or can
be summoned from likely sources.
Presently the amount of information being
generated and accumulated is growing at an
alarming rate and this is bound to continue.
Specialization in all areas of human endeavor
has increased, while at the same time the
conventional boundaries of disciplines in
science and technology have vanished. Everybody needs to have access to knowledge relating to his interests, no matter where it might
exist. It appears doubtful whether sufficient
competent manpower will be available in the
future to perform the intellectual tasks required to organize and channel this information adequately and effectively. If one accepts
this prediction, the question arises whether
the principles of automation may be effectively applied to this area, as they have been
to the area of manufacturing.
In the following discussion, an answer to
this question will be sought by investigating
some of the problems that will be encountered,
by delineating a comprehensive intelligence
system, and by establishing some of the prerequisites for the solution of the intellectual
problems involved.
WHEREVER A GROUP OF PEOPLE
A COMPREHENSIVE
INTELLIGENCE SYSTEM
The information needed within an organization to carryon its business is manifold in
92
/1
character and substance. Certain units of information are generated to be utilized at other
points of an organization. At times a message
is directed to a very specific point while at
other times it is directed to substantial sections of an organization. Still other elements
of information enter an organization from
external sources and information generated
within an organization is directed to external
destinations. Information is stored within an
organization for future reference and access
is desired to this store at any time to answer
questions that come up in the course of doing
business. All these functions constitute a business intelligence system 2 which is the more
complex, the larger the organization. The
difficulty of adequate communication within
an organization arises because of incomplete
knowledge of one point of the organization as
to the information requirements of any other
point of the organization. This is particularly
true where communication is desired horizontally across a vertically oriented hierarchical
organization chart. An effective communication system should permit free flow of
information between any two points of an
organization. A model which will substantially perform this type of communication is
shown in diagrammatic form in bhibit 1 and
will be described in brief.
A first objective of the system is that of
selective dissemimition of new information. 3
In order to accomplish this, an interest profile
is created and stored for each action point of
a system. Such a profile consists of the description, in some appropriate language, of
the current sphere of interest or activity of
the action point. A document entering the
system, irrespective of its having been generated internally or externally, and despite the
fact that it might have been addressed specifically to an action point, is characterized with
respect to its information content in a lan-
guage similar to that used for the action point
profile. This document profile is then compared
with all of the action point profiles in store
and wherever a predetermined degree of
similarity between the profiles is detected, the
affected action point is made aware of the
information involved. If, upon receipt of this
notification, the action point desires to have
the complete information, the whole document
will be made available. Acceptance of this
information in one way or another will be
noted by the system and be used to update
the action point profile so that it may reflect
actual current interests.
The third function of the system is that of
interest or activity matching (skill matching).
If an action point wishes to find out which
action point concerns itself with a given subject, it may synthesize the action point profile
characteristic of the subject in question and
charge the system with comparing this synthetic profile with all of the action point
profiles and to announce the identity of those
of the action profiles which match the synthetic profile to a given degree. This function
will aid in discovering who is knowledgeable
or skilled in certain areas or where to send
inquiries of problems which cannot be resolved
by an action point contacted originally.
The fourth function of the system is that
of matching, as a matter of course, any new
profile with any part or all of the profiles in
store. This function serves to uncover duplications, to various degrees, between new
activities and current activities, whether
within the organization or between the
organization and the outside world.
It is easily recognized that the exhibit
model comprises most of the information and
communication activities conventionally performed in any organization today. The present
objective is to automate this system, that is,
to perform the various functions to the high-
The second function of the system is that
of information retrieval. By storing all document profiles and the documents characterized
by them, a library is built up for future
reference. If an action point wishes to obtain
information which might be contained in the
library, it composes a synthetic document
profile characterizing the information needed
in the language used for creating the document profile. By comparing this query profile
with all of the profiles of the library documents and by establishing the degree of
similarity wanted, the inquiring action point
may then be notified of those documents which
appear to be pertinent to the query.
Exhibit 1
IN
ORIGINAL
PHOTO-PRINnNG
TRANSCRIBING
PHOTO COPY
DOCUMENT
MICRD·COPIES
STORAGE
BUSINESS
INTELLIGENCE
SYSTEM
FEED BACK
MONITOR
ANNOUNCEMENT
SELECTION
"CCEPTANCE
QUERY
TRANSMmAl
93
est possible extent with the aid of machines.
Some of the problems that arise in replacing
intellectual and manual effort by automatic
procedures will be reviewed subsequently.
DIFFERENT TYPES OF SYSTEMS
So far the intelligence system has been described in general terms, covering its basic
functions. It is important to realize at this
point that in practice such systems will necessarily vary considerably in accordance with
the particular situation they are to serve.
Also, while the system has been treated as a
comprehensive one covering all phases of an
organization, there are many good reasons
why specific phases of an operation should be
treated individually in order to achieve utmost
efficiency. However, such specialization must
not prevent a special purpose system from
being an integral part of the comprehensive
system.
It is beyond the purpose of this paper to
enumerate the many variables that enter into
the design of a special system but, instead, to
create an awareness of the fact that these
variables range wide and far. Without such
realization it becomes difficult to interpret
generalities and determine the degree to which
they apply to a given situation.
In order to illustrate the range of diversity,
two tables are shown in Exhibit 2, based on four
arbitrarily chosen variables. One table relates
three degrees of access time to an information
store to three degrees of activity in terms of
inquiries per day. The other table relates three
degrees of complexity of indexing a document
to three sizes of document collection. Because
of the lack of any standard of measurement
at this time, an arbitrary index of difficulty
has been placed at the intersection of the
variables in question, such index being a relative value of rank. According to these tables
Exhibit 2
SOME CHARACTERISTICS OF IR SYSTEMS
Queries Per Day
Access
10
100
1000
1*
2
4
1 hour
3
5
7
real time
6
8
9
Time
1 day
Items in Store
Indexing Effort
10 thousand
100 thousand
1 million
low
1
2
4
medium
3
5
7
high
6
8
9
* Arbitrary index of difficulty
94
one may, for instance, characterize the information retrieval system of a medium size
chemical laboratory by index number 1 in the
inquiry table and 6 in the storage table, while
an information retrieval activity of an airline
reservations system might be characterized
by indexes 9 in the inquiry table and 2 in the
storage table ..
Typically all such systems involve the storage of items of information in such a manner
that certain ones of these may be recalled in
answer to a given inquiry. There is, however,
a great difference of format of the items
manipulated by various systems, bringing
with it varying degrees of difficulty in characterizing the items so that they may be
recognized. The process of characterizing
stored items is generally referred to as indexing. This process, as currently performed,
ranges from ordering factual data by unique
identifications such as names of persons or
things, to the categorization of descriptive
articles and books by the assignment of classifications or subject headings. We shall concern
ourselves here with only the more difficult
tasks of characterizing extensive texts as
would be typical of technical papers.
AUTOMATIC INDEXING
The problems that arise when contemplating the automation of intelligence systems lie
in two distinct areas, namely, the operations
concerning the meaning carried by documents
and the processing of all physical aspects of
documents. In making this distinction it is
realized that meaning per se is not negotiable
for purposes of communication, except by
means of arbitrary physical tokens such as the
spoken or the written word. While present day
technology provides many ways for processing or manipulating such physical manifestations, meaning comes into existence only at
the moment a human, familiar with the tokens
used, interprets them.
Since automation is the objective here and
since the intellectual effort in organizing and
processing information is the critical area,
machine methods must be found which will
characterize documents in a manner that will
produce results similar to those obtained by
intellectual effort. This reduces to the question
of whether some physical characteristics may
be discovered in the make-up of documents
which to a satisfactory degree are typical of
a given meaning content. The result hoped for
is that if two documents display similar physical characteristics, the probability is high that
upon human interpretation the two documents
will disclose similar ideas. There is no need
that the patterns by which such analogies
may be constructed and recorded by a
machine, be meaningful in themselves.
A first approach to the solution of this
problem is the performance of a statistical
analysis of the physical word pattern of a
document. This would result in a vocabulary,
including the frequency of occurrence for each
word which would constitute the profile of
that document. By comparing such vocabularies of various documents with each other,
an indication may be obtained as to the extent
that similarity of the vocabularies signifies
similarity of meaning contents. Preliminary
tests of this method have produced promising
results. 4 During these tests it was also found
that the vocabularies may be reduced substantially by disregarding certain classes of
words. Even when such vocabularies were
reduced to twenty to thirty top ranking words,
adequate results were obtained for discovering similarity of subject matter for the
purpose of selective dissemination of information. s In this latter case apparently high
ranking (significant) words were manually
extracted from the document, a procedure
which anticipates the eventual extraction of
keywords by machine.
Tl1e method just described is generally referred to as keyword indexing. While this
method has simplicity to its credit, considerable experience is required before the limits
of its effectiveness will have been established.
This method disregards any specific relationship that an author might have established
between the words of the vocabulary. The
assumption here is that within the framework
of a coherent presentation of a subject, there
are few ways in which the words of a vocabulary can be meaningfully associated or
related. 6
A refinement of keyword indexing consists
of indicating a first degree of physical relationship between words by denoting word
pairs. The improvement that is expected here
is that such words usually modify each other
and therefore become more specific. A simple
method is based on the notion that if two
words occur in the same sentence and are not
95
more than three or four words apart, there
is a high probability that the meanings of
these two words have been directly related.7
Again, no specific relationship is indicated,
relying upon probability as to meaningfulness
among all possible combinations.
Exhibit 3
Lattice of word. pair linkage••
1ST ORDU
5 WORDS AND OVU
2ND ORDER
2, 3 AND 4 WORDS
@> •
o.
96
FREQUENCY OF WORD PAl as
IS INDICATED BY NUMBER OF
INTERVEN ING LINKS
Even though this particular scheme has not
yet been evaluated on the basis of sufficiently
large samples, an impression of practicability
may be derived from Exhibit 3. This figure
represents a lattice of word-pair linkages
constructed from a list of word pairs automatically derived from a short report on
scientific research,· by the method just de·
scdbed. By tracing various paths interconnecting the less than 30 words, intelligent
statements may be synthesized which have a
high probability of matching those of the
author.
As long as such keyword indexing methods
have not been proven to be impractical, it
appears to be premature to consider methods
of higher complexity, such as would involve
the determination of parts of speech, for instance. There appears to be a point of diminishing returns in that the more specific an
indexing term becomes, the lower is the probability that it will match an equally specific
term in a searching procedure.
Assuming for a moment that auto-indexing
of the types just described will effectively
serve to characterize information contents of
documents for the purpose of comparison, it
is of interest to point out that this method
differs significantly from manually applied
indexing methods in current use. Such current
methods rely on the judgment of the indexer
based on his understanding, at least in general
terms, the message carried by a document and
that he is thereby enabled to assign to a
document as many class designations 01' subject headings as have been provided for by
a preestablished classification structure or
subject heading list. Depending on which of
several indexers perform this job and under
which conditions, such assignments will vary
appreciably. No matter how detailed the rules
might be to govern the indexer in doing his
job, the interpretation of such rules again will
vary. It is furthermore the job of the indexer
to predict the point of view that a future inquirer might assume in his search for past
knowledge. It is obvious that the decisions an
indexer is capable of making to categorize a
new document, can only be based on past experiences and that his predictions can at best
be subjective and incomplete. The philosophy
underlying this method assumes a low rate of
change of point' of view of the inquirers the
system is to serve. In the area under discus-
sion this is contrary to experience in recent
years and it is doubtful therefore whether
classification and subject heading systems can
survive here, even if their systematic updating
is accelerated.
In the case of auto-indexing, the characterization of the document is "derived" by
extracting certain physical elements from the
original text in accordance with some standard
rule. As long as the same rule is applied, the
product is the same, no matter by whom,
where and when the extraction is performed.
Since only the physical make-up of the text
is the basis of this operation, it can be performed by machine, provided the full text is
available in machine readable form. Except
for the effects of contraction, no bias has been
exerted on whatever meaning might have
been expressed by the tokens extracted. The
document remains unclassified in the conventional sense, although it might be said that
classification has been exercised on the primitive level of words. No intellectual effort needs
to be expended in this indexing operation.
At the instant of searching for information
relating to a given topic, through the intermediary of index terms, the respective procedures for the two systems differ as follows:
In the conventional system, classes or subject
headings have to be identified which most
appropriately delineate the topic of the query.
In the automatically indexed system an appropriate class or subject heading is created to
order for the occasion, so to speak, by synthesizing the vocabulary of keywords or word
pairs embracing the subject matter as if it
too had been stated in the form of a document.
In the first case the inquirer has to adjust
himself to the point of view that an indexer
might have taken in the past, while in the
second case he has complete freedom in reflecting his own viewpoint of the present.
As far as intellectual effort is concerned, it
will be seen that in the conventional system
it had to be expended on indexing each new
document entering the system, even though
only a fraction of documents of a collection
is usually ever referred to again. In the autoindexed system, on the other hand, intellectual
effort is spent at the instant of search. The
techniques of search will necessarily be different and in the case of auto-indexed documents, additional effort is required to
overcome differences of word usage and
problems of semantics, as discussed further
on. In those functions of an automatic system
where auto-indexed profiles are matched to
each other, no intellectual intervention is
required.
REORIENTATION OF
PROFESSIONAL SKILLS
The. procedures dictated by automation call
for a major readjustmep.t of professional
talent presently engaged in the information
services. For the present the idea of having
the average user of an information system
push buttons appears impractical since such
procedure presupposes complete familiarity
with all aspects of the system, attainable only
through extensive training .. Communication
with the system should instead be carried out
through the intermediary of an information
specialist.
Constant awareness of users' demands and
of the capabilities of the system will enable
the specialist to negotiate and amplify inquiries so as to produce optimum results. To
this end the prope:rties of mechanically transformed and condensed texts need to be well
understood. It is essential that statistical data
are collected regarding word usage in both
the indexing operation and retrieval operation. Cumulative dictionaries must therefore
be compiled and subjected to periodic analysis.
These latter operations would of course be
performeq mechanically as a by-product of
indexing, abstracting and retrieval processes.
The information thus obtained will assist the
--specialist in formulating inquiries more effectively and to adjust his procedures to the
peculiarities of the information stored. It will
furthermore enable him to develop thesauri
and check lists denoting synonyms or nearsynonyms and associated words which need to
be included in certain search patterns to home
in on pertinent information.
There are many other schemes which may
have to be evolved to cope with the dynamic
requirements of the users of information
systems. It is up to the experience and resourcefulness of the specialist as to the extent
to which automated information services .may
become useful and effective. Even when users
are enabled to communicate directly with
automated information systems, it will be because of machine routines developed by information specialists for guiding and supervising
97
an inquirer in his searching operations. There
is, therefore, a new challenge to the profession
of documentalists, information specialists and
librarians. Education and training directed at
these new requirements should be an early
concern of universities and library schools.
AUTOMATIC ABSTRACTING
\
98
Another intellectual operation which is
presently performed to facilitate communication is that of preparing abstracts. Abstracting is a highly developed art which
differentiates between many forms of abstracts depending on the purpose they are to
serve. The common objective is to characterize
the content of a document by deriving from it
such fractions as will convey the essentials
of the author's message.
A mechanical approximation to this process
would depend, as was the case in automatic
indexing, on the assumption that certain
physical patterns of texts coincide with desired patterns of meaning. Thus it may be
argued that the frequency with which certain
words have been used by an author is a
measure of significance and that the greater
the concentration of such significant words is
within a sentence, as compared with other
sentences, the more significant is such a sentence as a conveyor of information characteristic of the contents of the related document.
By selecting from a document an appropriate
number of these highest ranking sentences
and by reproducing them verbatim in the order
of occurrence, an extract is derived which may
approach the usefulness of conventional
abstracts. 9
While the mechanical solution to the procedure just described does not offer any
difficulties it has been hard so far to evaluate
the results. IO Fully automatic abstracting of
articles contained on punched tape created in
conjunction with conventional typesetting
methods was demonstrated for the first time
at the International Conference on Scientific
Information, Washington, D. C., October,
1958. II Subsequent experimental operations
have indicated the usefulness of such "AutoAbstracts" where great masses of documents
are involved and where prompt availability is
essential.
The perfection of automatic abstracting
methods is an important step toward the reali-
zation of automation. Further research and
development in this area is in progress.
AVAILABILITY OF
MACHINE READABLE TEXTS
In order that machines may perform the
various processes referred to above, written
material must be available in machine
readable form, such as punched cards, punched
or magnetic tape, or other storage media.
While the printed or written page may eventually join the ranks of such media, there will
always be an advantage in accommodating
the machine, where possible, by providing it
with input of the most suitable format.
This facility therefore becomes a major
requirement in automation.
Many miles of punched tape are produced
daily through the use of teletype and typesetting equipment. Such tape has so far been
regarded as a means to an end and has not as
yet found its way into the processing phase of
individual intelligence systems. The amount
of manual recopying of textual material and
the attendant proof reading that is being done
today is staggering. Yet there must be a lack
of awareness of this fact since improvements
of this situation are so slow in coming, even
though the development of remedies is well
advanced.
Ideally there should be but one single
manual effort of creating an original record
or document. This effort should be applied in
such a form and with the aid of such tools as
will eliminate all further human intervention
in processing for purposes of printing, indexing, abstracting, storing, disseminating, retrieving or transmitting. A reasonable control
of the format of such original records or
documents is desirable and will contribute to
their negotiability wherever information is
exchanged in the public domain.
Since a substantial portion of the input to
automated intelligence systems would consist
of published information, the transcriptions
of such material into machine readable form
would entail an undue manual effort and expense for each individual system. It will therefore be another prerequisite that such material
becomes generally available in machine readable form. This may be accomplished through
cooperative arrangements among organizations of similar areas of interest. The text of
current material may be made available in
machine readable form by the respective
publishers on a subscription basis. In the
remaining cases professional societies and
commercial or governmental service agencies
may be expected to perform the task of transcribing needed texts into machine readable
form and storing them appropriately at
machine text centers.
The service of such centers would include
the furnishing of machine texts of current
literature relating to a prescribed area, on a
subscription basis, and the furnishing on demand of specific items, new or old. A further
service would consist of delivering this material in a specified format dictated by the
requirements of the user's system. This would
require that the centers are equipped to transform or reduce the machine texts to order,
relieving the user of this step and limiting
the transmittal of material to only that portion
or degree of information which is actually
needed.
A further requirement is that machine
texts are available promptly. Since machine
texts are already in the very (digital) form
needed for electrical transmission, existing
communication networks are capable of relaying such information from center to user
or exchanging it between centers or individual
systems. Eventually, by means of direct communication between electronic information
processing machines, the interrogation of
machine text centers or associated systems
may be carried out entirely automatically.
ACCESSIBILITY OF THE PICTORIAL
ASPECTS OF DOCUMENTS
As was mentioned earlier, the problems
connected with the automation of intelligence
systems consist of those concerning the manipulation of meaning and those concerning the
manipulation of the mere physical appearance
of documents. Automatic approaches to the
solution of the former have been dealt with in
some detail. It should be pointed out, however,
that a special case has so far been made for
texts and that the many other tokens of meaning, such as may be contained in pictorials of
all kinds, have not been gone into. The reason
for this is that the degree of difficulty in the
art of automatic recognition and characterization of pictorials appears to be of a much
higher order than that encountered with texts
and that proposed solutions are as yet in the
early stages of experimentation. However as
desirable and necessary as such refinement
may be in the future, it is felt that significant
progress can be achieved at this time without
this capability.
Dealing then with the purely physical
aspects of documents including pictorials, the
processing requirements which arise are usually referred to as document storage and
retrieval. Since the performance of these
functions does not involve intellectual processes per se but is instead the consequence of
such processes, this phase of the system will
not be elaborated on as being outside the topic
of this paper.
In the foregoing some of the problems in
mastering the intellectual aspects of intelligence systems have been presented and
possible mechanical solutions for them have
been reviewed. The approaches indicated in
detail are considered to be realizable in the
very near future. No attempt was made to
speculate on the feasibility of mechanizing
intellectual problems of higher order. The
position here taken is that it is entirely practical to start automatic systems today, provided that they permit step by step growth
and adjustment to changed and new conditions
as well as to technological advances for their
implementation. In fact, this adjustability
must be a primary design objective so as to
parallel the dynamics of modern life. The
availability of information processing equipment will have a particular bearing on this
aspect in that once a basic machine readable
record has been created, suitable transformations or extractions may be derived from it to
meet changing systems requirements as time
goes on.
I
REFERENCES
This article first appeared in The Clarification,
Unification, and Integration of Information Storage
and Retrieval and is reprinted with the permission
of the copyright owners, Management Dynamics,
P.O. Box 2864, Grand Central Station, New York
17, N.Y.
• H. P. Luhn, "A Business Intelligence System", IBM
Journal of Research and Development, New York,
N.Y., October, 1958.
3
H. P. Luhn, "Selective Dissemination of New Scientific Information With the Aid of Electronic Processing Equipment", American Documentation, April,
1961.
99
S. Stiassny, "Coefficients of Similarity Between
Documents", IBM Advanced Systems Development
Division, Yorktown Heights, N.Y., September, 1959.
5 A Selective Dissemination of Information System
utilizing profiles made up of 25 or less keywords
has been satisfactorily serving 450 professior.als and
technical people at the ASDD Laboratory of IBM
in Yorktown Heights, N.Y., since the Fall of 1959.
• An information retrieval system based on automatically derived keywords is operating experimentally at the Health Law Center, University of
Pittsburgh, Pa. So far over five million words of
machine readable texts have been processed and
stored.
7 H.
P. Luhn, "Auto-Encoding of Documents for
Information Retrieval Systems", IBM Monograph,
IBM Journal of Research and Development, New
York, N.Y., April, 1958.
"Plumb, R. K. "Experiments Suggest a New Ap ..
proach to the Treatment of Heart Attacks", New
York Times, September 22, 1957.
• H. P. Luhn, "The Automatic Creation of Literature
Abstracts", IBM Journal of Research and De1)elop?nent, New York, N.Y., April, 1958.
4
100
1OJ. Rath, A. Resnick, T. Savage, "Comparisons of
Four Types of Lexical Indicators of Content",
A merican Documentation, April, 1961.
"Papers for this conference were distributed to participants two months ahead for study. By arrangement with the Columbia University Press the Monotype tapes used in publishing these preprints were
malie available for experimentation. At the conference exhibit, IBM researchers demonstrated the
automatic transcription of these Monotype tapes to
magnetic tape via punched cards and thence the
automatic creation and print out of abstracts by
means of electronic d~l.ta processing equipment at
the Space Systems Center in Washington, D.C. All
this was done without any human intervention,
except for the handling of the input and output
records. Also, preprinted Auto-Abstracts of Papers
of Area 5 of the conference were made available
to participants at the beginning of the conference.
See: "An Experiment in Auto-Abstracting", Progress Report, IBM Research Center, Yorktown
Heights, N.Y., November, 1958.
By C. D. Gull
General Electric Company
13. Guidelines to Mechanizing Information Systems
As MY AUDIENCE, you may represent one of
many different operating information systems.
If you don't have such systems actively under
your own control, you have these systems
nearby in your organizations. We can assume
that most of these systems are manual, and
that you are aware that there is a trend away
from manual systems to a variety of mechanized systems. Some of you may be using
edge notched cards or the fully punched cards
which are used with electric accounting machines. A few of you may be using electronic
digital computers and devices which are peripheral to computers. A few information systems have been put through the gamut of
changes in the past ten years from manual
methods to edge notched or fully punched
cards and on to the use of computers.
You can check these trends for yourself by
studying the material about several companies
in the article in Fortune by Francis Bello,l in
the articles on information retrieval in the
Chemical and Engineering News of last JuIy,2
and in the latest publications in two of the
National Science Foundation's series, Current
Research and Development in Scientific Documentation,3 and Non-conventional Technical
Information Systems in Current Use. 4 The
trends are discernible in General Electric, Du
Pont, Eastman Kodak, Smith, Kline and
French, and in Charles Pfizer and Company,
for example. The examples in these firms include several different combinations and types
of equipment, but the trend is very definitely
evident.
You want to be able to answer these questions about your own information handling
needs:
Should we get on the bandwagon and join
the trend to mechanized information
systems?
Should we give up our manual system in
favor of a machine system?
Why should we undertake the transition
from a manual system to a machine
system?
How do we accomplish this transition?
Is a change to a machine system desirable
in itself?
Will a machine system produce the
required results?
Many detailed questions are given in the
following paragraphs for you to answer
before you can draw conclusions to answer
the above questions, particularly the last two.
The notable characteristic of this analysis
is that I am asking you a set of questions.
They are rhetorical only in the sense that you
will not be expected to answer them immediately. They are designed to enable you to
find out your own position on the questions
whiCh are asked. The reason for asking questions is that it is virtually impossible to give
guidelines that would fit all situations. Each
of you has a different problem, and the potential solution that may seem reasonable to you
is probably different in many specific details
from the solutions that your fellows will
adopt in answer to their problems. You will
have to derive the guidelines for yourself by
answering the questions.
It is desirable to emphasize at this point
that information systems process information
as their commodity. This attribute makes
them different from other control systems
where information is used to control the
processing of some other commodity. The
information handled in manufacturing processes is really secondary to the commodity.
manufactured, but in the business of handling
information, it is desirable to understand that
information is the primary commodity. Even
101
so, you may discover that the systems which
you have are hybrids because you are handling
physically discrete objects as well as information. You have to put these objects away,
serve them out, get them back, and serve them
out again to new customers. You are dealing
with commodities, these physical entities, and
you are also dealing with the contents, the
information therein. Another way of phrasing
this is the familiar question:
Are you indexing documents or are you
indexing information?
YOUR ORGANIZATION'S CAPABILITIES
FOR CHANGE
What are you own organization's capabilities for change?
Do you have sufficient competence in
management to contemplate a change
from a manual to a mechanized system?
Do you have adequate communication
with management?
Dr. James Hillier has given a very informative and penetrating analysis of how he
approached the managerial problem of what
to do about information services in a research
laboratory doing basic research work. s Essentially, he reached the conclusion that he should
not do anything at the present time, because
RCA maintains a conventional special library
and Dr. Hillier has assigned the information
handling duties for each research project to
at least one member of every project team in
his laboratory. He concluded, "That a centralized informatiQn service probably is
impractical and une20nomical and could even
have a detrimental effect on the creativity
level of the laboratory. That a rapid and efficient information retrieval system for specific
references is essential". In response to a question, Dr. Hillier said that he had not asked
his special librarians to participate in his
evaluation. He talked his evaluation over with
his creative people, who stated that the provision of information services is academic
because they don't read the literature. They
communicate only with their peers and productive colleagues. I believe that Dr. Hillier's
special librarians, if consulted, could have
demonstrated that creative research men cannot provide a rapid and efficient information
retrieval system for specific references, and
that it is not possible at this time to provide
102
such a system for a laboratory without centralizing the activities. It appears that Dr.
Hillier reached contradictory conclusions because of inadequate communication with the
personnel assigned full time to information
handling problems. He recognized this lack,
because he also concluded, "That there is an
urgent need for effective means of condensing
and consolidating technical information for
management's use". This conclusion needs to
be applied to information handling services
as much as to any other activities of an
organization.
The experience of the Battelle Memorial
Institute provides a contrasting example. A
number of years ago it was decided to institute a high level information system and
Battelle has followed that decision since. Battelle had a fine traditional special library and
put great emphasis on its research and development personnel to serve as outstanding
information specialists in their own projects,
but Battelle also added a high level manual
information activity to fill the gap which they
determined existed between the library'S service and reliance upon laboratory and technical
personnel. 6
Bernard K. Dennis of General Electric, has
described a different managerial environment,7 one in which he is permitted, as
manager of the technical information center,
to develop and carry out an aggressive program to provide information services. I believe
a similar situation has prevailed for Christopher G. Stevenson in managing the technical
information center at the Hanford Atomic
Products Operation, General Electric, Richland, Washington. Both men introduced
computers for information retrieval early in
the history of this application. Bernard Dennis has added another computer application,
the current dissemination of information, and
finds that it represents the greater load and
is more useful than retrospective retrieval by
computer. Stevenson has given up retrieval
by computer on finding that the number of
requests did not warrant the service, but he
has placed his serial and periodical order and
accounting processes onto a computer.
Another management situation is seen in
the Bell Telephone Laboratories, where W. O.
Baker, vice president in charge of research,
has actively supported the information activi-
ties. W. K. Lowry has put computers to regular
use there as part of the extensive information services provided to personnel of Bell
Laboratories.
These examples show that, when you begin
to look at mechanization of your information
system, you need to investigate very early the
various questions about the competence, openmindedness, and support of management.
A number of questions can be asked about
the new man and machine relationships:
Will you be able to prepare for the new
man and machine relationships?
Are your own personnel adequately flexible for this purpose?
Are they openminded; are they adaptable
to change; are they willing to vary their
routines to accomplish the objective of
making your information system more
productive?
Do you have within your own organization capabilities of undertaking a
preliminary analysis?
There is a useful bulletin from the Bureau
of the Budget in the Executive Office of the
President. Its title is Automatic Data Processing (ADP) Program of the Executive Branch:
Studies Preceding the Acquisition of ADP
Equipment. Just to give you the flavor of this
useful analysis, I quote from section IV,
Preliminary Analysis, substituting the phrase,
"Information System", instead of "ADP",
where "ADP" occurs in the text.
"An 'Information System' study can be a
lengthy, complex, and costly undertaking.
For these reasons, a preliminary analysis
or 'feasibility study' should be made to
determine the overall soundness of applying
the 'Information System' to the operation
of an organization.
"1. Informational Requirements. Establish
an inventory of 'Information System' processing by major functional areas ... Additionally, it is necessary to secure or develop
a distribution of man-hour effort and other
significant cost elements to be used in identifying costs by major operation, function
and/or organization" .
Bernard Dennis gave that kind of information
in his paper.'
"2. Analysis. The analysis of data developed should be undertaken in a logical
sequence. Several starting points are suggested, namely, unmanageable backlogs,
those operations which are the most costly,
take the most time, or are presently
mechanized" .
There are several more points listed. I am
skipping them to go on to "3. Determination of the Impact. The end
product of the preliminary analysis should
include determinations regarding the areas
susceptible to immediate or ultimate application, the size and cost range of equipment,
the extent of the conversion problem, the
sufficiency of available resources to support
the !Information System', the expected
economies and benefits, and the overall
effect of computer use on the personnel
procedures, and organization. The preliminary analysis should identify the organizational units that might be affected by the
possible utilization of the 'Information
System' and make adequate provision for
participation by staff representatives of
such units in the subsequent studies.
"4. Decision. If the preliminary analysis
shows that application of the 'Information
System' is generally justified, further and
more definitive plans are in order to proceed
with a full study". 8
These quotations give you some idea how
useful this bulletin is for your needs.
SYSTEMS DESIGN COMPETENCE
Can you detach a task force of your own
people for a preliminary analysis or an
actual design?
Where the manpower exists within your
information activity, can you detach it?
To cite an example, when we studied the
Undergraduate Division Library of the University of Illinois at Chicago, the Council on
Library Resources insisted that people from
the library staff be put full time on the study.
Two people were detached from the staff for
one year. Their task did not interfere with
the library'S daily operations, nor did they
need to feel any responsibility for current
operations of the library. Their only job was
to look forward.
COMPETENCE TO EVALUATE
STUDIES AND RECOMMENDATIONS
Will you be able to evaluate the study
103
which you prepare or which someone else
prepares for you?
This isn't an easy question to answer. There
are several publications to study in order to
prepare yourself to answer this difficult question. The book called System Engineering, by
Goode and Machol,' is the classic in system
engineering textbooks. The third volume of
the Handbook on Automation and Systems
Engineering, edited by Grabbe, Ramo and
Wooldridge, lOis just published and also useful
for this purpose.
ASTIA has published a number of reports
on their automation effort. The first was on
their preliminary design and what they intended to accomplish. I I They subsequently
issued 'reports on what they have done with
their computer and how it has affected the
services which they perform for the Government and for Government research and
development contractors. 12, 13 They are now
undertaking to improve the thesaurus of
descriptors which is used in their indexing
and retrieval operations. These reports provide a case history of the installation of a
machine system.
The Western Reserve University (WRU)
experience is an example of the development
of a mechanized information system for which
you cannot find guidelines. They had an intuitive feeling that machinery was needed and
that machinery required the use of a very
detailed analytical indexing method. Following their intuitive assumptions the WRU
personnel built a combination of certain
aspects of subject heading indexing and of
hierarchical classification into the telegraphic
abstracting .method. They built a paper tape
prototype of a searching selector. After trying
for a magnetic tape searching selector, WRU
began to use a general purpose computer to
accomplish retrieval and current dissemination for the metallurgical Documentation
Service of the American Society for Metals.
But note that the WRU personnel are
undertaking the tests after the information
system has been put into operation. By contrast, if you haven't already moved in the
direction of mechanization, you can try to get
some answers to the guideline questions which
are in this paper before you design your
system.
There is also another paper by Charles
104
Bourne and Douglas C. Engelbart who wrote
on the aspects of the technical information
problem. 14 Many of their questions and points
are related to the problem of a national centralized information system, but they are
useful and you can cut back from the national
aspect to your own applications.
THE PRESENT STATE OF YOUR
MANUAL INFORMATION SYSTEM
Eva Lou Fisher has prepared A Checklist
for the Organization, Operation, and Evaluation of a Company Library.15 The first fifteen
pages provide a number of questions which
will enable you to evaluate an existing manual
system which is oriented toward a special
library rather than toward an information
system. It gives a starting point in print from
which to begin to answer these questions:
Is our present system satisfactory?
What does our system fail to provide
now?
A careful examination of the present system must be m~de before you decide to explore
the desirability of a mechanized system.
Let us examine the current performance of
your information activity.
Has your activity reached the .point where
human decisions in handling information
need to be supplemented by machine
decisions?
Actually, this situation occurs when you begin
to employ machines. You are relying on machines to give you some of the answers,
perhaps all of them, when you ask a retrieval
question or when you disseminate information.
Do you even know what these human
decisions are?
From the standpoint of your people, could
you stretch their capabilities and their
present perfortnance to reach a satisfactory level in your own manual system
now?
This is a managerial problem. Its solution will
require ingenuity on your part. There is always the possibility that you can achieve this
better operational level without adopting
machines. Mr. Luhn's paper covered the point
of this next question very fully.16
Do you have a communication problem in
your information handling activity?
He answered it positively,
But have you looked at it as a communications problem?
Have you identified what this communication is - how it flows?
Do you require an optimal system?
Other than the form of the words,
Do you understand what an optimal
system implies?
There is an engineering meaning to this question, and it needs to be investigated.
One of the valuable points brought out at
the second of American University's Institutes on Information Retrieval was the
recognition by Dr. Robert M. Hayes, as a
mathematician and data handler viewing
libraries, that in many ways the present
manual systems are fairly close to being optimal. It has required perhaps a century to
develop the optimal state of a manual
operation of a library.
Is it realistic to ask for optimal performance within three months after the
installation of your machine system?
Whether you are going to be the operators
of the system or whether you are going to hire
machine oriented experts to be the operators,
Do you understand the operator's responsibilities in making the optimal system
work?
Do you already possess the quantitative
data and a qualitative evaluation of your
present performance?
Certainly attempts to answer this question
will uncover maj or deficiencies in many places.
Most of the systems with which I have come
into contact do not possess the type of information which Bernard Dennis has illustrated.'
The traditional library will know how many
books it cataloged last year, it will know how
much money was expended for books, and
perhaps how many cards were filed in the
catalogs. However, it will be unable to identify
the unit operations and the hours spent on
each operation, and consequently it will not
know the manpower costs on each of these
operations. These are facts you must obtain
before you can change from one system to
another.
In summary, here are some questions which
are useful to ask before proceeding to the next
phase of your investigation.
Can you actually define your problem?
Can you define the performance of your
system?
Can you now define what you want to
advance to?
Have you properly stated the basic question to yourself?
LOGICAL FLOW CHARTS
One of the contributions of the computer
industry to system engineering is a method
of analysis called logical flow charting. There
is a small and humorous example in Exhibit 1
to illustrate the product of flow charting. The
operation of making a logical flow chart need
not be a particularly difficult task It is·a form
of writing as well as a form of analysis and a
great deal depends upon the choice of words
within the boxes of this type of diagram.
Can you make a logical flow chart of
your existing system?
After the flow charts are made, the person
responsible for the administration of a certain
set of operations can often spot difficulties in
the system, such as unnecessary operations,
bad timing, and interruptions in the smooth
flow of the operations. Immediate action can
be taken to correct the situation administratively, and better performance can be provid~d
within a manual system.
After you have flow charted your existing
system, can you make a logical flow chart
to show how you wish to have the new
system operate?
Can you make a block diagram of the new
system?
Oftentimes the block diagrams are made first,
because they are more generalized and they
do not have to have the specific yes-no character of the flow diagram. If you implement a
system with a computer, the logical flow diagram will be required by the computer
programmers.
FUTURE REQUIREMENTS
AND EXPECTATIONS
We need to watch our tendency to wish for
too much too soon.
What are your expectations?
Have you identified the basic requirements which your system must fulfill and
105
"How to Get to Work in the Morning.
I,
Original source unknown.
Set alarm
Yes
?
Turn
off alarm
No
Kiss wife
Groan
(Repeat 3 times)
?
No
Shake wife
Kiss wife
Crawl out
Walk
out to car
Yes ._----"----......
Yes
Below 30
No
Yes
70 ?
No
Get back
in bed
Bathroom
chores
Turn on
heat
Go back
for it
Pos sible
dead end
Breakfast
Grumble
Yes
No
Give
her $10
106
Drive to plant
eliminated your desires and wishes about
the performance of the new system?
If you are going to require that all your
wishes and desires be fulfilled in these systems, certainly you will find them costly to
implement.
Perhaps you can separate requirements and
wishes by considering the information operation for which computers were first used: to
search for answers to questions.
Do you want a machine system for information retrieval?
Do you require speed in retrieval, or
exhaustiveness in retrieval?
It seems to be generally understood now
that the speed of retrieval for one person's
single question search is not improved by
using a computer, as compared to the time
that a person spends in consulting a card
catalog, for example. You must spend time to
code the question in a form that the computer
will accept and you have to take your turn
on a computer with the other people who use
it, with the result that it may take longer in
elapsed time to get an answer from the computer than it does to search manually in a
card catalog or in an abstracting and indexing
service.
As to exhaustiveness of retrieval, not only
is it costly to put information into a form that
permits retrieval, but also there will remain
a lot of material outside of the computer
facility which will have to be searched
manually.
Do you have such a large volume of
retrieval questions that you cannot handle
the searches with your present staff?
If yes, computerized retrieval offers the ad-
vantage of batching many questions for one
machine retrieval operation. This capability
offers a dramatic reduction in cost per question but offers no reduction in elapsed time
per question per person. This capability
has introduced a new problem, too, of overcapacity in retrieval services, or underconsumption, depending on your point of view.
Most of the installations which are performing computerized retrieval are operating
considerably under their capacity for answering batched questions. For example, Bernard
Dennis indicated he could batch 1,300 questions in one search; in actual practice, he
batched 250 questions for one search. 7 It is
more common to batch 5 to 40 questions per
search. There is a two-way communications
channel which his people are not loading to
capacity. He has difficulty in getting his 3,000
potential customers to realize that the system
has a capability which they ought to use.
Thus, you acquire an education problem when
you increase your retrieval capacity.
The retrieval situation is in better balance
at ASTIA, which has a large, daily question
load, derived from Government agencies and
some 1,500 contractors (3,200 contracts), to be
searched on a small computer with an uncomplicated program in contrast to Bernard
Dennis' situation of a small daily load to
be searched on a large computer with a
sophisticated program. 7
There are other activities you might consider:
Do you wish to accomplish the input
processing activities in your information
activity or can you obtain the desired
results elsewhere?
Do you want to undertake automatic indexing to replace human efforts?
H. P. Luhn indicated that he questions the
effectiveness of automatic indexing. 16 Don
Swanson has reported in Science magazine,
however, on the results of one experiment
which indicates that perhaps machine indexing is actually more effective than human
indexing. 17 ,
One of the basic problems is :
Do you want to sort, that is to reorder
or rearrange, the indexed information?
The material comes in chronologically (or
serially) and whatever indexing you accomplish acquires that same attribute. Yet you may
have to reorder the index so that you can use
it with a manual system.
Do you want to use electronic devices to
accomplish this rearrangement?
Computers can be programmed to sort text
into new arrangements more economically
than by having the same work done by people.
The rapid adoption of the new technique to
prepare key-word-in-context (KWIC) indexes
or permuted title indexes for publication is
evidence of this advantage. This technique is
being used to prepare all varieties of catalogs,
indexes and bibliographies and furnish them
107
108
with author, title, subject, and numerical
index sections as required. Publication of the
computer's work enables users to consult the
new arrangements without resorting to the
computer for search and retrieval.
Do you wish to add a new service, selective dissemination af current information,
to your informatiQn system?
This service is prohibitively expensive in a
manual system, but you can consider it with a
mechanized system. The computer matches
the results of current indexing against the
customer's questions collected and stored over
a period of weeks or months, and prints out
notices to individuals telling them that new
publications have been received which satisfy
their stored questions. The individuals return
marked cards to the information center,
requesting publications for themselves or to
be forwarded to others, or requesting modification of their stored questions. The center
tells local authors who asked for their works.
Altogether, the selective dissemination of current information is one of the most significant
advances in service in recent years.
Do you fully appreciate that selective
dissemination of current information is
the converse of retrospective retrieval
when you are using a mechanized system?
If you select equipment and procedures to
provide one of these services, you may have
the other service also. For example, in retrospective retrieval you match a small number
of questions against a massive index store;
in selective dissemination you match a small
index store to a large collection of stored
questions; and in each case you may print the
results in question order, index order, or
document-number order.
Do you want your binding records incorporated also?
Do you want to handle payroll and personnel records as well?
The large scale capabilities of the devices
we are talking about argue that you do want
to mechanize all of these operations, so that
you can load the machine adequately to achieve
low unit cost and maximum efficiency. Plan
to take as much of the workload off your staff
as possible and place it on the machine, leaving your personnel free to help your customers.
One question that calls for very serious
consideration is:
Do you want to put all of these operations
together and start out with a large jump;
that is, put every operation you can think
of onto the computer and its related
devices?
Or will you put the several operations
together one after another in a gradual
fashion?
Probably the organization which elects to
synthesize the system and start all the operations together will be the outstanding pioneer,
if it is successful in accomplishing the change
in one large effort, but there is still a real
challenge in changing one operation at a time.
Are you seeking improved performance
in your information activity; e.g., improved coverage, improved timeliness,
lowered cost?
Do you desire to free the staff to serve
the users?
These are questions that will aid you in paring
down the original question and separating
your desires from the practical accomplishments which are required.
ROUTINE OPERATIONS
Do you want to use these mechanized
devices for routine operations in your
library?
If you circulate materials, circulation records
may be necessary.
Do you want to mechanize circulation
records?
Do you want to keep your serial checking
records with the aid of a computer?
Do you want all your acquisitions work,
the financial and detailed record keeping
of your purchases, or gifts, or exchanges
to go through the computer?
CAPABILITIES OF MACHINE SYSTEMS
Some of the things that cannot be overlooked are:
Do you understand what can actually be
done by machines?
Have you an idea of the relative effectiveness you can attain from a manual
system, edge notched cards, or an electronic accounting machine complex?
What are the advantages of adding computers as a more sophisticated system
than the punched card and paper tape
systems?
The great amount of literature in this field
provides very little information and insight
into the answers to these questions.
SYSTEMS DESIGN FOR A MECHANIZED
INFORMATION SYSTEM
Let us assume that you have considered all
the preceding questions and that you wish to
undertake the transition from a manual system to some form of a mechanized system.
The transition involves system design.
Who will design this system for you?
There are various methods of determining
system design. For example, the government
sends out invitations to bid on its requirements. The resulting proposals are then carefully analyzed and a contractor is selected to
do the work. Private institutions and private
corporations can investigate the field of potential contractors by inquiry and select an
organization to do their systems design. It is
also possible to elect to do the systems design
within your own organization. However, very
few organizations have such capabilities in
their total structure. It is very unusual to find
an organization within the library and information business which can undertake the
design of its own system. It is usually necessary to call in specialized assistance from
outside the organization altogether.
SYSTEM DESIGN DIFFICULTIES
How difficult is your systems design
problem?
Do you have any good information on the
boundaries of the system?
What will your performance criteria be?
Will you have a queuing problem; that is,
too much of anyone operation on the
system at one time?
Is it possible for you to evaluate and
predict how successful your information
system design is going to be?
Is this something that you can undertake?
Are you going to use analytical or simulation techniques?
What kind of equipment difficulties are
you facing?
Are you aware of the various instructions
that can be given to a computer?
Are you aware of the problems of computer programming?
Do you have any concept of the time it
takes to do the programming?
Can you borrow computer programs
which have already been written and
have been proven satisfactory in similar
situations?
Can you use them without change; can
you modify them?
Can you use available sub-programs
which can be combined to produce full
programs?
The programming problem is certainly a
serious one; you may find its costs range
anywhere from $5,000 to $100,000 for certaip
types of information retrieval programs.
If you. are actually going from one existing
machine system into another one,
What problems of conversion are you
going to have?
Is there enough information in the
existing system already in a form compatible with a new system that conversion
will be worthwhile?
Are you going to have to undertake a major conversion of existing information?
Since universal print readers do not exist
yet, how much more key punching will
be required?
If you are in the area of the larger information systems,
Have your architectural and engineering
requirements been fully considered?
Also, there is the paramount problem of
costs:
Do you know what your costs were in
the past?
What are they now?
What do you expect they will be after you
have implemented a new system?
Who is going to do your implementation
work?
Is this going to be a different eroup than
the one who did your systems design, or
will you use the same group?
Who is going to retrain your personnel?
Are you going to do this internally, or are
you going to use specialized outside help
for retraining?
CHECKLIST OF RECOMMENDATIONS
As the closing paragraph, I shall quote
109
again from the Bureau of the Budget Bulletin
60-6, to provide a checklist of things you want
to be sure you are covering:
"Items to be included in the Information
Systems study recommendations. A carefully
documented Information Systems study would
normally include the following:
"a. Definitive indications of the benefits
that are expected to accrue from the
proposed system that cannot otherwise
be attained, and justification for these
benefits in terms of economic gains or
improved management effectiveness.
"b. A general description of the data processing needs of the activity in relation
to the basic mission of the activity.
"c. Workload and employment statistics
describing the magnitude of the
activity's problem.
"d. A chart depicting the organizational
effect of 'The Information System' on
the activity as well as showing the
location of the proposed data processing
function.
"e. Flow charts supplemented by clear and
precise narrative indications of the
nature and scope of each application.
There should be included for each application the following:
(1) A description of the contents and
purpose of proposed outputs.
(2) A description of the content and
source of all input data.
(3) Record lengths and number of
records.
(4) Processing frequencies.
"f. Narrative descriptions of those present
systems which correspond to the proposed application.
"g. Detailed comparative cost data for the
proposed system and the present
system, adjusted by improvements
developed in the course of the study.
"h. Indication of personnel implications in
terms of reduced or augmented staffing
needs.
"i. Indications of the availability of and
expected costs of site preparations.
"j. Projected schedule should be provided
covering plans for equipment selection,
detailed application development, personnel training and orientation site
1.10
preparation, installation, transition and
attainment of normal operations.
"k. Indications that consideration was
given to locating or obtaining equipment which could be shared with other
bureaus in your agency, or with other
agencies.
"1. All factors which might have a bearing
upon top management's decision, including problems or disadvantages."8
REFERENCES
'Francis Bello, "How to Cope with Information",
Fortune, Vol. 62, Sept., 1960, pp. 162-168.
2"Information Retrieval", Chemical and Engineering
News, Vol. 39, July 17, 1961, pp. 103-112; July 24,
1961, pp. 90-98.
'Current Research and Development in Scientific
Documentation, No.9, National Science Foundation,
Washington, D.C., Nov., 1961.
4Non-Conventional Technical Information Systems in
Current Use, No.2, National Science Foundation,
Washington, D.C., Sept., 1959, and its Supplement,
March, 1960.
sJames Hillier, "Management's Evaluation of Information Services", Information Retrieval Management, Data Processing Library Series, American
Data Processing, Inc., Detroit, Mich., 1962, Chap. 8.
6G. S. Simpson, Jr. and J. W. Murdock, "Qualitative
Approach to Scientific Information Problems", Battelle Technical Review, Nov., 1960, pp. 3-7.
7Bernard K. Dennis, "Financing a Technical Information Center", Information Retrieval Management, Data Processing Library Series, American
Data Processing, Inc., Detroit, Mich., 1962, Chap. 9.
"Automatic Data Processing (ADP) Program of the
Executive Branch: Studies Preceding the Acquisi~
tion of ADP Equipment, Bureau of the Budget,
Bulletin 60-6, March 18, 1960, pp. 2-3.
9Harry H. Goode and Robert E. Machol,. System
Engineering, McGraw-Hill Book Company, Inc., New
York, N.Y., 1957.
'0E. M. Grabbe, Simon Ramo, and Dean Wooldridge
(eds.), Handbook of Automation, Computation, and
Control, Volume 3, John Wiley & Sons, Inc., New
York, N.Y., 1961.
"Automation of ASTIA - A Preliminary Report,
Armed Services Technical Information Agency, Dec.,
1960, AD 227000.
12Automation of ASTIA-1960, Armed Services Technical Information Agency, Dec., 1960, AD 247000.
13Evolution of the ASTIA Automated Search and
Retrieval System, Armed Services Technical Information Agency, Jan., 1961, AD 252000.
'4Charles P. Bourne and Douglas C. Engelbart,
"Facets of the Technical Information Problem",
Stanford Research Institute Journal, vol. 2, no. 1,
1958, pp. 2-8. Also in Datamation, Sept.-Oct., 1958,
pp. 6"12.
15Eva Lou Fisher, A Checklist for the Organization,
Operation, and Evaluation of a Company Library,
New York Special Libraries Association, 1960.
16H. P. Luhn, "Automated Intelligence Systems," Information Retrieval Management, Data Processing
Library Series, American Data Processing, Inc.,
Detroit, Mich., 1962, Chap. 12.
17D. R. Swanson, "Searching Natural Language Text
by Computer", Science, 132, Oct. 21, 1960, pp. 10991104.
By Hattie T. Anderson
w. R. Grace & Co.
14. Compatibility of Information
and Data Systems within a Company
RECENTLY, I ASKED FOUR PERSONS knowledgeable in the field what "compatible" systems
meant to them. One replied that any systems
which work are compatible. The second stated
that compatibility doesn't matter; we will
simply have a separate program for each
system. The third thought of the capability
for hooking together various pieces of equipment. The fourth related compatibility to
universal language.
The literature today abounds with references to "compatible", "integrated", "convertible", "coordinated", "mergible", "company
wide", "universal", "centralized", "unified",
"flexible", and above all, "machine processable" information and data systems. Singly,
these loosely used and iII defined terms may
mean something different to each one of us,
depending upon our orientation within the
field and the problems we are trying to solve.
Cited together, however, they suggest a
definite, overall management approach for
handling information problems.
The purpose here is to examine such an
approach against a company background and
to ask ourselves whether, in spite of the
opinion of one of the above quoted experts,
compatibility between systems does matter.
Can management and should management try
to build compatible, integrated, company wide
information and data systems? If so, how?
To lend feasibility to our investigation, we
will first note some of the information problems which result from the characteristic
structure of a large, technology based company. Secondly, we will try to determine what
types of compatibility could be attempted
between scattered company information centers to alleviate the problems noted. Thirdly,
we will look at the more complex types of
compatibility and search for factors which are
basic to their achievement. Finally, we will
mention a few of the difficulties which would
be encountered in the development of a
grandiose, corporate information network.
A CORPORATE PICTURE JUSTIFICATION FOR COMPANY WIDE
SYSTEMS
A large, technology based company typically
consists of a number of decentralized divisions, affiliates, or subsidiaries, some having
many laboratories, plants, or offices of their
own. These are scattered throughout the
world. Intradivisional liaison is difficult
enough to maintain, not to speak of interdivisional liaison. Activities of the units that
require technical information range through
basic and applied research covering many
fields and scientific disciplines, patent work,
applications and product development, process
engineering, economic evaluation, commercial
intelligence, plant engineering and operations,
technical services, sales, management, etc. A
moment's thought tells us that there will be
measurable overlap in interests, activities, and
information needs between many parts of the
corporation, no matter whether these parts
are organized along product, functional, or
geographical lines. The overlap may vary in
type, extent, and frequency, but there can be
little doubt that it exists.
There are very likely those within our
company model who suspect that duplication
and waste are rampant, but some rationalize
it as a healthy thing. Some, through fear of
leak of proprietary information or through
divisional and professional jealousies, may
111
wish to keep it that way. Others think
the whole information problem is hopeless
anyway, so why worry.
There is a high probability that our company is thus characterized by excessive costs
resulting from repetition of work, frequent
lack of information nee!1ed for stimulation or
decision making, inadequate cross-fertiliz·ation
of ideas, loss of time advantage, and waste of
professional talents. Yet today, anything short
of maximum possible flow of information between the parts of a single competitive entity
has become intolerable in the face of keen
rivalry between companies and national economic systems. Indeed, we can theorize that
our company will be noncompetitive with
organizations that do succeed in breaking the
communication barriers.
Looking more deeply into the situation, we
find within the corporation numerous libraries
and information services operating with staffs
that are quite diverse in number and qualifications, ranging all the way from secretaries
through formally trained librarians, self
trained scientific librarians, literature scientists, information specialists, analysts, systems
engineers, programmers, consultants, and the
like. Services begin with simply ordering
books and shelving them by some homemade
scheme and extend to experimental study of
information storage and retrieval problems.
One unit may feel lucky to have bookcases and
a reading table, while another takes computers
and mechanization for granted. Collections
vary from a few text and reference books to
comprehensive and very costly subscriptions
to published or custom searching and current
awareness services.
It is important to note that a one to one
relationship between the need for information
and financial support for services does not
always exist. There are some laboratories that
are too small to pay for expensive collections
and information services. Others cannot convince management of their importance~ Conversely, Mooers' Law has taken over in still
others. Adequate means for equalizing the
dissemination and use of information based
on need are lacking.
A startling discovery might be the amount
of inefficiency and duplication that exists between the very libraries or information
centers which are the hope for reducing
112
much of the duplication elsewhere within the
company.
Examples of inefficiency may be seen in staff
situations. With trained personnel difficult to
find and keep, we often see novices independently struggling to rediscover even the most
elementary library procedures. With each
turnover in staff, especially in the small
library, there may be a total loss of familiarity
with sources of information and disastrous
lack of continuity and consistency in maintaining catalogs and retrieval systems. Crossfertilization of ideas and sharing of knowledge is just as lacking among mformation
personnel as among other professionals within
the company.
Inefficiency is further seen. in the failure of
company libraries to use their resources additively, even though each one differs in its
subject orientation and in the familiarity of
its staff with unusual or local sources of information. The latter is especially pertinent in
the case of overseas units, since it is often
difficult to identify, at home, foreign sources
of information.
Examples of duplication are found in acquisitions of very expensive reference materials
and in services provided, the costs of which
continue to rise at a rate which should be a
matter of serious concern and analysis by any
company. The same documents may be indexed and even translated several times.
Patent and literature searches may be repeated at great costs. Each unit spends many
man hours developing authority lists of terms,
some good, some poor - none compatible.
External to the corporation there are also
circumstances which bear on the feasibility
of compatible, integrated, company systems.
These, I think, are of three types.
One is the phenomenal development of the
technology for processing, communicating,
and disseminating -large volumes of information, even at long distances. Remote collections
can now be exploited where they could not
before. It may soon become negligence for a
large company to ignore new capabilities for
handling information.
The second is the origin of new sources of
scientific information and data, some of which
we cannot purchase and store on our shelves,
or if we can purchase something for searching within the organization, it may be tapes
or punched cards. I am speaking of the several hundred (427 is the latest figure I have
seen) 1 scientific information and data centers
which have cropped up everywhere with very
little apparent coordination. The small, illfunded, and ill-staffed company library cannot hope to keep up with these developments.
The larger information center may even find
difficulty justifying the use of some of the
expensive equipment, techniques, and search
materials, unless by so doing it can alleviate
some of the company problems already described and thereby reduce overall corporate
costs. Let me only cite the recent announcement by Documentation, Inc. of its forthcoming Mechanized Index to Chemical Patents
Literature to be provided on magnetic tape at
a cost of $12,500.
The third external development is the considerable thought and study now going into
the development of compatibility on an even
grander scale than we are considering - between large government, society, commercial,
and regional information services. Everyone
is aware of the need for compatibility or convertibility between media, machines, systems,
and terminology. Anyone interested in compatible company systems should certainly
watch these developments closely for what
they will teach and for ways of tying in with
them.
These are some of the situations which I
think justify a study of compatible, company
wide information and data systems. YoU: will
very likely wish to add other considerations
to those which I have sketched.
TYPES OF COMPATIBILITY
But can we now give the concept of compatibility of information and data systems
within a company more concrete meaning?
Within the framework of a corporate situation, such as that just presented, I would like
to suggest that it consists of equalizing, normalizing, maximizing, and optimizing the
dissemination of information among scattered, heterogeneous company parts; that
uncoordinated information systems and services are incompatible with company interests
if it can be shown that through some form
of cooperation, sharing, exchange, or joint
approach a greater amount of needed information can be more quickly or widely
disseminated or overall costs of acqUIrmg,
processing, storing, and disseminating this
information can be reduced.
This is not easy to demonstrate without
extensive study and, perhaps, actual test
efforts. However, we can gain some insight
into the potential rewards of such efforts
through a study of the activities tabled in
Exhibit 1. We should consider these activities
as they might be separately conducted in the
smallest, ill-equipped library to the largest,
best-equipped information center of our
imaginary company. We can try to identify
anv activities which we think
(1) entail wasteful duplication when con-
ducted without coordination, and
EXHIBIT 1
Identification of Types of Compatibility
Activities
Management
planning
personnel selection,
training, and use
forms, methods and procedures
systems engineering
experimental studies
other
Is there
Is there
problem of need for
duplication? exchanl!e?
Suggested
corporate
action
Acquisitions
selection
identification of sources
ordering
other
Processing (manual or machine)
bibl iograph ic descriptions
subject indexing or cataloging
data indexing
citation indexing
abstracting
translating
vocabulary development
coding
programming
storage
other
Reference
identification of sources
performance of searches
research liaison and
literature analysis
other
Dissemination
current awareness service
announcement of acquisitions
union lists
reproduction and display
circulation
bibliographic print out
reports, charts, tables
other
Interactions between above
113
(2) could be improved or reduced in cost
through some sort of exchange, sharing, or pooling of staff talents, information resources, and services.
In a private communication to the author,
Saul Herner noted that all systems are
compatible to some extent. To nurture compatibility, system ingredients, he said, should
be carefully analyzed and identified and common attributes and labors extended where
possible.
I believe that the concept of compatible
company systems can be applied to many
traditional, as well as newer, information
functions; that it embraces people, collections,
equipment, systems, and services. The process
of achieving compatibility starts with the
earliest efforts of one company library to seek
help from or give help to another company
library. It moves on through varying degrees
and types of coordination or integration of
information activities, such as the familiar
use of printed catalog cards. At the extreme
end of the spectrum of compatibility lies
complete standardization and centralization.
FACTORS BASIC TO COMPATIBLE
COMPANY RETRIEVAL SYSTEMS
Many of the problems which might come
up in our analysis of Exhibit 1 are relatively
easy of solution, once recognized, and for
reasons of time I would like to ignore them
in this part of our discussion.
The most challenging problems for company, as well as for society, government, or
commercial services, are those of compatibility between systems for indexing, retrieving, and disseminating information and
between the language of those systems. There
is yet, however, no meeting of the minds even
as to the meaning and purpose of such compatibility. For example, participants in a
recent ASTIA meeting on compatibility between thesauri placed major emphasis on
finding a means of "relating concepts from
one system or thesaurus to another, no matter
how the concept is presented verbally". There
are other strong proponents of the universal
language approach.
In a company situation, it seems to me that
we should consider the capability for merging
directly, or with nominal conversion, into a
single retrieval system all or parts of certain
114
indexes prepared at different locations. Indexing efforts could then be exchanged between
sites; repeated indexing of the same documents could be avoided; searches of local
indexes could be supplemented by central
searching of pooled indexes; questions would
need to be translated into only one system
language for broad searching of information
available throughout the company; innumerable man hours could be saved by joint, rather
than separate, design of systems and preparation of vocabulary; indexers using the same
system and vocabulary could easily be transferred from one indexing project to another
with changing workloads or for training purposes; no longer would the users have to guess
which of numerous indexes might contain the
information they need.
Whether all this can be achieved as a practical matter is uncertain, but we will proceed
on the questionable assumption that it can.
On whatever basis you might undertake to
develop compatible indexing and retrieval
systems, you will wish to build in maximum
flexibility and convertibility so that you can
study and manipulate the system in many
ways, adjust to changing requirements, and
take advantage of improved techniques and
equipment. Compatibility with the future
should not be sacrificed for compatibility with
all parts of the company. It is important to
remember that attractive but irreversible
decisions may beckon at every turn. A decision or action is totally irreversible if it
cannot be altered wihout completely redoing
past work. Such decisions should be made with
cautious deliberation where costly operations,
such as the analysis of documents by subject
specialists, are involved.
It is no easy matter to finally determine
which of your company indexes should be
made compatible (in the sense of mergibility
or convertibility, one to the other) and which
ones should be kept separate; which courses
of action will permit compatibility, and which
ones will not. I have listed in Exhibit 2 some
of the categorical investigations and a number
of specific questions which might shed light
on these matters. In considering these questions, we should look for
(1) factors which are basic to the achievement of compatible systems, and
(2) decisions which might be irreversible.
EXHIBIT 2
DEVELOPMENT OF COMPATIBLE STORAGE AND RETRIEVAL SYSTEMS
Discussion Questions
A.
B.
C.
What are the sources and special characteristics
of your records which will affect the compatibility
of their treatment within a single system and
between company units?
1. Do some deserve more costly treatment because
they are more valuable? What is the duration
of their value? Which documents have local
value only? Have some documents only intrinsic
value?
2. Must your identification be accession numbers,
Page numbers, line numbers, test numbers,
sample numbers, product numbers, run numbers, data sheet numbers, drawing numbers, or
a combination of such? Can different identifications be merged into one system?
3. If data is to go into the system, is it repetitive
(such as test or operating data), critical data,
or random numerical information?
4. Are the same records duplicated in several types
of documents?
5. Do you wish to treat interest profiles, questions
to the system, specialized internal and ex;;ernal
sources of information, or other nonconventional
items as documents to be fed into the system?
6. What are the physical characteristics of the
documents?
7. What are the sizes and expected growth rates
of various segments of the collection?
What are the corporate versus local objectives
or requirements?
1. Is the goal current dissemination, retrospective
retrieval, or both?
2. Do you wish to retrieve data, information, or
documents?
3. What free, auxiliary information should be
captured? Do you wish to retain complete bibliographic detail, citations, author abstracts?
How can this information be used now and in
the future?
4. Do you need to index chemical compounds? For
which collections and what purpose?
5. Must patents be indexed to permit retrieval
from a legal point of view?
6. What operations will be performed upon the
data or information in the system?
7. Will most questions be ones of recall of known
documents or retrieval of unknown documents?
8. Where must information and data flow within
the company? How will it be used? In what
form should it be disseminated? Is print out of
bibliographies and special reports required?
Will you reproduce or circulate documents?
9. ShoUld systems be user orientated? Is browsing
capability needed? Is on-the-spot searching
required?
10. What are the variations in time requirements
and volume of input and output?
11. What are the variations in depth and specificity
of search requirements?
12. What are the weeding requirements?
What system of indexing should be adopted for
corporate use?
1. Do objectives and characteristics of the documents require several systems?
2. Is there considerable vested interest in existing
systems? Should these bea model? If not,
should they be continued, converted, or terminated?
3. Should you use coordinate indexing, permuted
indexing, classification, or some other type of
indexing? Is there an application for more than
one type of indexing within one system? If so,
how will one complement the other?
4. Should you use interfixes? What kind? Should
systems which use interfixes be merged with
those which don't? Can the same system of
interfixes be applied to all types of documents
covering many disciplines?
5. Can data and information be merged in one
system?
6. Should you index chemical compounds by name,
empirical formula, and/or structure? Can you
adopt, modify, or combine some existing approach, such as that of Dyson, Wiswesser, the
Biological Coordination Center, Dow, Garfield,
Patent Office, etc? Should compounds be indexed
into a general or separate system?
7. Will professionals or non-professionals do the
indexing at each site?
8. How do you determine the optimum depth and
specificity with which to index for overall corporate needs? Should each center index to meet
its own or corporate requirements? Can documents which are shallowly or generically indexed be fed into the same system as deeply or
specifically indexed documents?
9. Will the system chosen lend itself to machine
indexing techniques?
10. Will the system be compatible with external
indexing services of greatest importance to you?
D.
How will you develop your vocabulary?
1. Should you choose Mooers' descriptor, Taube's
Uniterm, or Herner's bound term approach
or a combination or compromise of these?
2. How will your choice affect your system of
indexing?
3. How will you treat generics, logical relationships, synonyms, rarely used terms, etc.?
4. Can your vocabulary be integrated with, translated, related or converted to other vocabularies?
5. Can it accommodate new and changing terminology?
6. Can all subject disciplines and viewpoints be
reconciled in one vocabulary?
7. How should you handle numerical data?
E.
What format and method of storage should be
used for the collection and for the indexes?
1. Does it matter whether inverted or non-inverted
storage is used?
2. Should the index and document store be kept
separate or combined?
3. What are the pitfalls of coding and fixed fields
for print out, conversion, sorting, etc.
4. What consideration should be given to upper
and lower case, abbreviations, special characters, sub and super scripts?
5. How do you avoid duplication of input to the
system?
6. How important is space? Is duplication of document and index files justified at all sites?
7. Must you reproduce, display, transmit, or reorganize directly from the store?
(continued on next page)
115
8. Can the document store be organized in some
way to supplement, rather than duplicate, the
access points provided by the index?
F.
G.
How do you delete from the system?
1. On what basis will you delete from the system?
How often and how much? Will this vary between segments of the collection and between
company units?
2. Will weeding necessitate renumbering of items,
reorganization of files?
3. Should any data or information be retired into
special lists, compilations, tables, handbooks,
etc. ?
all types selected cOItlpatible as to input/output
media, speed of opeI\ation, principle of operation, function, etc.?
2. Are any locations committed to existing equipment? What are the limitations of this equipment?
3. What can you do to provide the capability of
conversion to newer and better equipment?
H.
What equipment should you choose to perform
local and corporate requirements?
1. Can any ol the activities be handled more
quickly, easily, or at less cost if machines are
used locally or at a central corporate center?
These questionR are not presented as a complete checklist for the study of compatible
company systems. They are intended only to
illustrate the types of questions that might be
asked.
DIFFICULTIES - SOME
GENERALIZATIONS
Innumerable factors would have to be considered before compatible, corporate systems
could be put into operation. It is difficult
enough to produce answers for a single use~
group. By the time a large corporate study
could be completed, some' of the early conclusions would already have been invalidated by
changing circumstances. The increased study
time required would invite criticisms from
those anxious to get started. No information
center ever seems to have enough staff to meet
even the daily demands placed upon it. All
those spending time on a grandiose, corporate
scheme, uncertain of accomplishment and
claiming· benefits not yet quantifiable, might
be considered impractical dreamers, ignoring the more immediate problems of their
customers.
The intricacies of the approach are apparent; We are normally inclined to start with the
simple and progress to the more complex only
as necessity demands; but in the development
of systems of maximum compatibility to meet
maximum needs, we might have to do the
opposite.
Variations in available funds, staff, and
equipment would also create p,roblems. For
example, if local capabilities were too limited
for participation in the corporate project,
should a central center step in? Or, if corpo-
116
If so, what types of machines are required? Are
What the the interactions within the overall
system?
1. Does any new decision invalidate past decisions?
2. Can work accomplished in one part of the system be'reused in another part?
3. How can you integrate all the infonnation and
data systems finally set up anywhere in the
company so that in searching none that are
pertinent will be overlooked?
I-ate requirements were greater and more
costly than the local requirements, who should
pay the extra costs?
Inconsistency of input to the system by
widely separated people having different
viewpoints would certainly be a danger.
What controls could be established to maintain equivalence in document analysis and
interpretation?
Those fearing leakage of proprietary information might have just cause for their fears.
Careful procedures would have to be worked
out to guard proprietary secrets.
Documentation is a young art. For the past
ten or so years, we have been preoccupied
with the development of new techniques, few,
if any, of which have yet proven themselves
to the exclusion of others. Now, with small
chance of consensus as to the best approach
for a single situation, talk of compatibility
emerges with standardization looming as its
ultimate manifestation. Premature standardization on a large scale to systems which are
not the best is entirely possible.
As clearly indicated by the earlier papers
of this Institute, criteria for the selection of
techniques have been the needs of isolated
user groups. If we are now to combine the
needs of many groups and service them from
merged systems, new criteria for development
of systems will have to be found.
The challenge is there. Some say it is insanity to think that each unit can continue
alone; others, that compatibility can be
achieved. So choose your form of insanity.
I
REFERENCES
Specialized Science Information Services in the
United States, National Science Foundation, 1961.
By Simon M. Newman
Consultant
15. Economic Justification -
F actors Establishing System Costs
THE ECONOMIC JUSTIFICATION for a proposed
or existing mechanized information system, or
even a partially mechanized one, is not simple,
nor is it easy. Figures from normal cost accounting procedures are inadequate, not
because the costs of any system cannot be
computed, but because, without the knowledge
of the contents of the output of the system,
such cost figures will not provide the proper
information for a valid justification. Why are
such cost figures not helpful? There are several intangible and ephemeral factors, other
than mere cost, which are more important in
any evaluation of a system.
The search to find a figure of merit on which
to base a decision to select a particular information retrieval system is of the same order
of complexity as trying to select the best new
model car for one's personal use. If one wishes
to purchase the best car, does he make his
standard of comparison, color, original cost,
operating cost, or some such other factor?
Concomitantly, how does he evaluate such
factors as easy riding, easy turning or parking, bucket seats, protection against crash
injuries, etc.? In the end, a rational selection
of a car or a retrieval system depends on
choosing a combination of cost and performance to meet existing requirements.
There appear to be four factors which
affect the rational choice of a system. I
have termed them: efficiency, utilization,
availability, and operating speed. The most
important and complex of these is efficiency.
EFFICIENCY
Efficiency is the ratio of the amount of pertinent information delivered by a system to
the total amount of pertinent information in
the original documents from which the system
files were made, provided that proper correction for any false delivery of information is
made. A system which is 100 percent efficient
will deliver all, and only all, the information
in the original documents in response to a
query requiring th'at information. Any system,
if economics are disregarded, can of course,
reach 100 percent" efficiency; one can search
every document for every question.
But how can one measure efficiency. The
yardstick has yet to be developed. Bourne,l
of Stanford Research, has recently completed
a six month preliminary study for the
National Science Foundation of how such
measurements might be made. Another study
by Arthur Andersen lwd Company,2 on the
same subject, is about to be published.
Measuring efficiency includes, at least, a
separate and detailed evaluation of how much
information in the document file is found in
the systems file, i.e., how much of the information in each document appears in the surrogates in the file which is searched. Of
course, there is no loss where the entire document is placed in the file, e.g., as one at the
University of Pittsburgh by KehJ,3 as used
in one of the programs for the General
Electric Search Comparator,4 and as prop@sed
by Newman 5 for research on U. S. patents.
But any system which uses surrogates has
some loss, and such a loss should not be
overlooked.
If surrogates are used, efficiency can be
shown to vary directly with the depth of
indexing, and with the sophistication of the
indexing system. Such sophistication can
provide, for example, the ability to search at
different levels of generality and specificity, or
117
from different aspects or viewpoints, or from
both. One major hidden cost in the use of
surrogates is the continuing expense of updating the search files, or the increasing loss
in efficiency from failure to update. The constant growth of our knowledge causes the
immediate deterioration of the retrieval value
of any surrogate. A document may convey
more information five years after publication
because of new knowledge developed during
the interim. Such additional information can
be harvested only by re-indexing, if surrogates
are used.
Efficiency is also lowered by noise, i.e., the
so-called "false drops" or system output that
is neither pertinent nor relevant to the question. This statement is made with the full
realization that both pertinence and relevance
are highly subjective terms, and that neither
of them has been quantitatively measured.
A theoretical study testing efficiencies of
retrieval systems has been done by Mooers.6
Cleverdon/ in England, has released a preliminary report on his research in measurements of the relative efficiencies of different
types of indexing.
Not every situation requires a high efficiency system. It is not necessary to use a
forging press to swat a fly. Small or inactive
files can often be serviced with modest means.
Simple edge notched card systems will often
perform satisfactorily for limited information
requirements.
UTILIZATION, AVAILABILITY AND
OPERATING SPEED
Utilization is the ratio of the actual use of a
system to its potential use. A perfect system
is not economic if it is not used. Mooers' Laws
is not a facetious formulation; some people do
not desire information. There are also people
who do not use the system because they do not
trust it, or because they do not know how to
use it. Utilization can be increased and costs
lowered by an educational campaign aimed
at the potential users of the system.
Availability is a measure of the ease of
accessibility to potential users. Physical arrangements for the system should be both
simple and attractive. If the location of the
user contact point is not close to his work,
speedy and detailed communication between
the user and the system must be provided.
118
The operating speed, i.e., the elapsed interval between the time a question is presented
to the system and the time it responds with
the requested information, should be min~mal.
A system will not be consulted if a user feels
that he can get the information faster by
doing the research himself. Mueller 9 found
that typically there is a 25 percent loss of
productivity by a worker awaiting a response
to a query.
GENERAL OBSERVATIONS
Mueller has published the only cost study
of an information retrieval system which I
have been able to find. However, he includes
in the input costs of his company's system,
the costs of internally generated reports.
Thus, these input costs include the cost of the
original research as well as the expenses of
editing and printing. Since these expenditures
would have been made even if a report were
not entered into the retrieval system, I question the value of this particular cost analysis.
Several theoretical studies 10 of cost analysis
exist, in which a mathematical model of the
retrieval system is created. Thi& is done by
formulating one or more equations which
attempt to represent the operation of the
system. Should the model parallel closely the
operation of the system, the figure of merit
yielded by the model will approximate the
costs of the system operation. These models,
although purporting to be of general application, seem to resemble the bookkeeping
inventory type of information systems. The
operation of such systems is so far removed
from the information retrieval systems which
we are discussing here, that the figures of
merit which are developed in these studies are
not helpful.
In order to justify a particular cost, one
must have some measure of the value of the
information which the system delivers. Quite
obviously, such value varies from search to
search, from question to question, and from
user to user. The value of information developed for retrospective searching, which is
often exhaustive, may be very different from
that developed for a current awareness search.
Engineers involved in technical development
may require different types of information
from scientists engaged in basic research, and
from administrators of research, development
and manufacturing departments. But even the
same information developed for different uses
may have different values for each of these
uses.
Some researchers have approached the
creation of systems, and of cost and efficiency
studies by studying the users of information.
Users' responses will often reflect the type or
kind of service and system with which they
are familiar. After some 33 years of experience, I am convinced that no user, no matter
how sophisticated he may be, asks the question to which he wants an answer; instead, he
asks the question he thinks the system can
answer. It may be misleading to depend upon
user reaction. Studying users and their habits
requires a high degree of understanding of
their problems.
Some help in pre-estimating operating costs
can be derived from time measurements which
most manufacturers list for operating sequences of their machines. We must first
collect data on actual costs of operation of
existing systems, such as that of Dennis,"
before figures of merit for evaluation of
information retrieval system costs can be
developed. It is from a correlation of such
sets of data from many sources, that systems
for devising figures of merit can be designed.
REFERENCES
C. P. Bourne, et al., Requirements, Criteria, and
Measures of Performance of Information Storage
and Retrieva,l Systems, Stanford Research Institute,
Menlo Park, Calif., 1961.
"Research Study of Criteria and Procedures for
Evaluating Scientific Information Systems, Arthur
Andersen & Co., New York, N.Y., March, 1962.
N.S.F. Contract C 218.
1
3William B. Kehl, "Communication Between Compuputer and User in Information Searching", Information Retrieval Management, Data Processing
Library Series, American Data Processing, Inc.,
Detroit, Mich., 1962, Chap. 11.
4G.E. Search Comparator (Advertising Circular),
General Electric Defense Systems Department, Information Systems Operation (DSD-820), Dec. 1961.
5S. M. Newman, Rowena W. Swanson, and Kenneth
Knowlton, A Notation System for Transliterating
Technical & Scientific Texts for Use in Data Processing Systems, U. S. Patent Ofike Research and
Development Report No. 15, U. S. Department of
Commerce, Washington, D.C., May, 1959.
·Calvin N. Mooers, The Intensive Sample Test for
the Objective Evaluation of the Performance of
Information Retrieval Systems, Zator Co., Cambridge, Mass., RADC-TN-59-160, 1959.
7Cyril W. Cleverdon, Interim Report on the Test
Program1ne of an Investigation into the Comparative Efficiency of Indexing Systems, College of Aeronautics, Cranfield, England, Nov., 1960.
·Calvin N. Mooers, "Mooers Law", American Docu·
mentation, XI, 3, July, 1960, p. 204.
"Max W. Mueller, "Time, Cost and Value Factors in
Information Retrieval", General Information Manual,' Information Retrieval Systems Conference,
Sept. 21-23, 1959, Poughkeepsie, International Business Machines, White Plains, N.Y., 1960.
10Mathematicai Models for Information Systems Design, RADC-TR-61-198, Magnavox Research Laboratories, Torrence, Calif., 1961.
11 B. K. Dennis, "Financing a Technical Information
Center", Information Retrieval Management, Data
Processing Library Series, American Data Processing, Inc., Detroit, Mich., 1962, Chap. 9.
ADDITIONAL REFERENCES
A. Melik-Shakhnazarov, Technical Information in
the U.S.S.R. (translated from Russian), M.LT.
Libraries, Cambridge, Mass., 1961.
Don R. Swanson, An Experiment in Automatic Tex,t
Searching, Ramo-Wooldridge, Canoga Park, April,
1960.
J. Verhoeff, W. Goffman and Jack Belzer, "Inefficiency of the Use of Boolean Functions for Information Retrieval Systems", Communications of the
ACM, 4, 12, Dec., 1961, pp. 557, 558, 594.
119
By Karl F. Heumann
National Research Council
16. International Activities in Documentation
IN TREATING THIS TOPIC, I don't like to look
back in history, because in a sense we are surrounded by history. In science information as
in other areas we can observe the role and
effects of nationalism and internationalism.
When we see groups of people fighting and
going through all the familiar steps to found
independent countries in Africa we're seeing
nationalism. When the United Nations, however, uses a variety of means to soften the
forces of nationalism, then we are seeing
internationalism.
Nationalism is extending the sites of international meetings into the far reaches of the
world. An example is the 1960 conference of
the International Federation of Documentation (FID) at Rio de Janeiro - it was the
first regular conference to be held outside
western Europe. According to a recent estimate, about 71;2 international scientific meetings are started every day and we can expect
further enlargement of the scientific world
and its conferences.
One of the ways internationalism affects us
is the need to gather documents from around
the world for such meetings. In 1956 the number of abstracts and chemical abstracts from
the USSR was nearly half the number from
our own country.
As a consequence of this geographic scatter, language problems are sure to arise. In
1957 Peter Bernais and 1 made- a count of the
languages of the original articles abstracted
in chemical abstracts. Russian language originals occurred about one-third as frequently as
English originals. A more recent study of
1960 material shows that in some sections of
chemical abstracts there were more Russian
language originals than English originals.
It is estimated that two new scientific periodicals are founded each day. The fact that
these journals start without regard for a total
120
plan is obviously making our international
information problems more severe. Founding
a journal in one country can sometimes be
defined as a nationalistic act.
It would be interesting to have statistics on
the countries of origin of articles in professional journals. Some years ago the editor of
a chemical journal made a study of this for
one year. After the United States, the country
from which the second largest number of
articles came was Canada; but the third one
may surprise you, it was Egypt. Since then an
Egyptian chemical journal has been founded.
This example rather suits both sides of my
case because it shows the international sources
of some of our leading U. S. journals. It shows
also that nationalism can act to alter the
pattern of these sources.
A trend that seems destined to continue is
the increase of scientists traveling to international meetings both in and out of the
United States. The 10th Pacific Science Congress held in Hawaii in the summer of 1961
had over one thousand scientists attending,
about half of them from the United States.
And there is the International Congress, to
be held in Moscow this year, where there will
be 125 Americans participating by giving
papers.
We can look forward to increasingly complex international relationships with other
countries having their effect on U. S. documentation, particularly having an effect on the
raw materials for science information.
INTERNATIONAL ORGANIZATIONS
There are several international organizations concerned with documentation. Although
most are professional associations, UNESCO
is an intergovernmental organization which is
giving limited but increasing attention to
documentation through both its library and
science offices. In my OpInIOn the role of
UNESCO should be studied further. It has
the opportunity to draw attention to problem
areas which may then be attacked by other
organizations at the UNESCO level. For
a variety of reasons, including political,
UNESCO has not seen fit to operate vigorously in problems of documentation, with the
exception of one good bibliography.
Among other intergovernmental organizations with active interests in documentation
are EURATOM and the International Organization of Patent Offices. The latter is developing an intensive program of information
exchange in Latin America.
Nongovernmental international organizations are often weak and have very little
support. They also have the problem of infrequency of meetings as intergovernmental
organizations do. The most important information activity internationally is the program
of the International Federation of Documentation which is frequently referred to by its
French initials, FID. The FID now has 43
national members, and has a secretariat in
The Hague.
Until a few years ago FID was primarily
concerned with the Universal Decimal Classification. Specialists in the United States have
not been greatly interested in problems of
classification, but in almost all other parts of
the world classification is the normal way of
dealing with scientific information. FID has
had some 50 years of experience in promoting
and refining a world-wide classification system. Recently the FID has broadened the
scope of its staff and conference attention to
the full scope of documentation problems.
The International Council of Scientific
Unions is another important nongovernmental
international organization. The primary activity of ICSU is abstracting. Its function is
often misunderstood; it is not to provide the
individual with abstracts, it is to help the
editors of abstract periodicals in their own
work. The abstracting program was started
with physics only. An attempt was made to
get the editors to contribute to abstracting
periodicals regularly, to exchange pages by
airmail, and to settle some standards in terminology and dialectics. Progress might seem
slow, but I think it has been quite effective
in the given field. In addition ICSU has now
expanded into chemistry, and is on the verge
of including the biological sciences.
The International Federation of Library
Associations has an interest in international
documentation. Another new organization
with a related interest is the International
Federation of Information Processing Societies (IFIPS) which holds its triennial
Congress in Munich in 1962. IFIPS is especially concerned with electronic computers
and associated equipment and the theory and
technique for their effective use in processing
data and information in other forms.
There are several United States organizations which have some relation to international
documentation. A recent article in the FID
Journal by Burton Adkinson is a comprehensive and well balanced statement: "Science
Information in the Federal Government".
The Office of Science Information Service
in the National Science Foundation is well
known for the publication of a directory of
nonconventional systems for dealing with
scientific information. Within the Office of
Science Information Service there is a foreign
science information program concerned with
the development and support of more effective international science communication. Dr.
Adkinson, Head of the OSIS, has become the
president-elect of FID. I think this is a reflection of the fact that the United States is becoming much more involved in the international exchange of information.
There is also in the National Science Foundation an Office of International Programs.
It has been established recently, and its relationship to international programs is not yet
crystallized. The National Science Foundation
maintains an office in Tokyo, which is a very
interesting experiment in attempting to get
scientific information across international
boundaries. It is probable that this program
will be expanded.
The Department of State has a science
advisor and several science attaches, who
also deal with scientific information. The National Institutes of Health has recently established an Office of International Research, and
intends to put a number of staff officers in
various countries, the first of which is now
in Paris, primarily to make sure that information about research supported by NIH in
other countries comes back to the United
121
States. It is important to assure that the
results of research support funds which we
provide to other countries are reported back.
Other agencies of the Federal government
which have foreign science information programs are Atomic Energy Commission, the
Department of Agriculture, the Library of
Congress and the Department of Defense. It
is important to call attention to the dangers
of a great many uncoordinated agencies dealing independently in information. A real
problem is that scientists in other countries
find themselves being talked to over and over
and over again by a variety of people from
the United States. It is possible that the science
attaches of the Department of State can play
a constructive central information and informal coordinative role for representatives of
various agencies active in a foreign country
or area.
I am associated with the NatIOnal Research
Council of the National Academy of Sciences.
The Academy and Council are nongovernmental although they perform many services
to government. There is in the Academy an
Office of International Relations which is concerned with documentation as well as other
topics. The Academy has a committee interested in the work of ICSU, which provides a
channel from the United States into the International Council of Scientific Unions, largely
through the Academy staff. The Office of
International Relations also has a Committee
on UNESCO and it is through this committee
that many of the recommendations about
scientific affairs, and in fact international
affairs, are made to our U. S. National Commission for UNESCO. The Academy also in
its Office of International Relations has an
active Latin American program which includes
some aspects of information work.
122
The United States recently has changed its
adherence to FID, so that the Academy could
become a U. S. national member for the
Federation. In common with other adherences
of the Academy to the international organizations, we have formed the United States
National Committee for FID. This committee
has as its purpose broad representation inside
the United States, so that there is a better
channel between the people in the United
States interested in this topic and the International Federation of Documentation. The
committee is made up of thirteen representatives of professional societies, eight governmental organizations, several members at
large, the Secretary of the Academy, and all
the people in the United States who are
elected officers of FID.
I would like to close by citing four topics
which should be the subject of some concern.
First, there is not enough continuing study of
international trends in scientific information.
There are individual stUdies, but they tend to
be fully individual and not to have long term
substantive value. Second, we have information about information centers in the United
States, but we do not have adequate information about centers in other countries that
might well provide service to us.
Third, I think there is a lack in the use of
international patent information and that the
International Organization of Patent Offices
could well be prevailed upon to make this kind
of information more readily available. And,
finally, as a personal matter, I am interested
in some day having an International Union of
Documentation rather than the tremendous
scattering of activities found in our present
situation.
By Saul Gorn
University of Pennsylvania
17. Computers, Communications,
and Science - Extending Man's Intellect
IT IS AMAZING TO ME how the subject of this
paper crops up again and again and how
relevant the material is to different types of
audiences. I have used this material in addressing management people on the impending
shift in manpower, and in addressing high
school teachers on the impending changes in
educational ideals. It is material which appears in sections of introductory courses that
might be called "computer appreciation". It is
concerned with specialization in science and
how it is related to the development of more
and more interdisciplinary areas. In the way
that is most relevant to the area of computers
we have introduced it into the introductory
course in the curriculum toward Master
and Ph.D. degrees in the Computer and
Information Sciences at the University of
Pennsylvania's Moore School of Electrical
Engineering. I call it "a philosophy of
communication" .
THE DEVELOPMENT OF MACHINES
AND THEIR CHARACTERISTICS
The general attitude in this philosophy of
communication comes up very naturally in an
introductory course. Visualize it as given to
a group of students who are going to be professionals in the field of computer and information sciences. I first go through a rapid
review of the history of computation. This
field appears to have developed at a pretty
slow pace over the ages until the 20th century.
At the point of the middle of the 20th century
I talk about the influence of the war in accelerating the need for computations. I
indicate that production itself was outstripping information about what was being
produced at that time. A specific instance is
the case during the war where guns and ammunition were being produced at a faster rate
than the tables needed to fire them accurately.
This prompted the Defense Department to
ask Moore School of Electrical Engineering
to give a general training program on computation, so that more people could help out in
this situation. Right next door to this training
group was a group of people working on
pulsed digital techniques for radar. Somebody, probably Eckert and Mauchly, got the
idea that since radar involved very rapid
counting of pulses, perhaps these techniques
could help the people with the bothersome
computation problem.
The turning point of putting electronics
into computers was conceived in that way,
resulting in ENIAC, and then a long series
of other computers whose names ended in
AG. After the AG's came all the numerical
series - the 700s, the 1100s, the 1400s, etc.
By the end of the first hour the blackboard in
the classroom shows an enormous growing
tree of different types of computers, all digital. Then I say: "Now wait a minute! How
many firing and bombing tables do we have
to compute? How can we explain this sort
of an explosion ?"
The explanation has two sides. Somehow
these machines had certain properties which
were just what was needed at the moment.
This would explain the development from
nothing to more than a billion dollar industry
in only ftfteen years. Conversely, what was
there about society and its structure that made
this happen and that made this just the thing
that was needed at the moment?
In talking about the machines, we point out
that it was unfortunate that they were called
123
computers. The French simply invented a new
word "ordinateur" for the purpose. We point
out that these machines are really general
purpose symbol manipulators or information
transformers. They allow one to do incredibly
complicated syntactic processing of groups of
symbols, where these symbols may be interpreted in any way one pleases. We rapidly get
into the idea of loop control, which makes the
machines general purpose. A key feature of
this loop control is the common storage of
instructions and data, which makes it possible
for the machine to shift the interpretations of
what is in the storage so that at one instant
a word might be an instruction where, at another, it will be data. The machine is capable
of changing its own instructions as it proceeds. The result of this is the ability to
simulate the self-referencing subtleties of
speech and thought.
SOCIETY AND THE
INFORMATION PROBLEM
Now what about the picture of society?
Why did it need just that kind of instrument
at that time?
Let us consider a cybernetic picture of
society. The definition of cybernetics as Norbert Weiner gave it (just about fifteen years
ago) is the "theory of communication and
control in animals and machines". That title
itself was considered very dramatic at the
time. Actually it was rather conservative,
because one can look upon any growing
organism in a cybernetic fashion: human
society, a culture, a company, a language, a
technology, language systems, domains of
knowledge, etc.
In a fam'ous essay by J. B. S. Haldane,
called On Being the Right Size, the author
begins by remarking that it is fortunate for
humans that insects do not grow to the size
of elephants. He then remarks that, considering the almost mechanical breathing system
an insect possesses, the internal communica~
tion system for its oxygen, if you will, it is
physically impossible for such a system to
develop enough pumping energy to control a
musculature (the control system) beyond a
certain size.
COMMUNICATION AND CONTROL
An important feature of organisms that
can be studied in this cybernetic fashion is
then the relationships between their internal
communications and their internal controls;
every growing system maintains willy-nilly
a balance between communications and' control, even if that balance means the death of
the system. Those cultures, those civilizations,
that became unbalanced on the control side
developed a kind of stupidity, in which information necessary to intelligent behavior
arrived too late. Like the dinosaurs, they died.
Those civilizations that were overbalanced on
the communications side could understand
everything perfectly but could never do anything in a suitably coordinated way, like the
Greek civilization or some of the past French
Republics. They also died. If we look at the
development of human cultures as far back
as we can, we find that there are certain
critical times when there were vital changes
in the structure, in the whole organization of
society. These vital changes or revolutions
seemed to come in pairs. I submit that the first
of each pair was a control revolution, and
that rapidly following after it, within a
few hundred years, there was a resulting
communications revolution.
Far back in history there was the agricultural revolution which was a control revolution. Mankind became able to control his
physical environment considerably more than
he had previously. This was followed by an
urban revolution within less than a thousand
years. This was a communications revolution.
The same thing has happened more recently.
About 150 years ago there was a control revolution called the industrial revolution, when
men became capable of a much greater control
of the physical and social environment because the machines took over large scale
routine labor. The rapid expansion of the
population which could now be supported
created an imbalance on the control side. A
crisis has developed which has created a need
for a reorganization. of the communications
system. People concerned with scientific
information are generally acutely aware of
this problem. They refer to it as the information crisis. Others in our society express it by
saying: there is too much red tape, too much
paper work, there is too much traffic, there is
too much specialization.
The effect of the first industrial revolution
was that the machines took over large scale
routine labor. In the second revolution the
machines are now taking over large scale
routine thinking. These symbol manipulators,
these information transformers, have made it
surprisingly, sometimes painfully, evident
that the biggest part of what we have considered to be thinking was in fact routine and
therefore mechanizable.
MANPOWER AND EDUCATION
The result of each of these violent changes
in social structure was a change in manpower
distribution and types of manpower needed.
There was a corresponding change in educational ideals because a different distribution
of manpower needed a different kind of educational system. Thus in the first industrial
revolution the large mass of unskilled labor
that the large societies had come to depend
upon seemingly became at the time unnecessary. Actually this mass was freed for even
more work that did result later. The time lag
was painful, but it was only a time lag. While
the unskilled laboring classes became by and
large unnecessary, there was an enormous
class of white collar workers (that had not
been needed before) which were now needed.
The corresponding shift in educational ideals
was from an emphasis on teaching techniques
in some kind of guild system to teaching
everybody reading, writing, and arithmetic.
Just at that time it became economically important and economically feasible to provide
universal education. This also accelerated the
communications crisis that resulted.
Since the machine is taking over large scale
routine thinking, there is now another corresponding shift in educational ideals. Teaching
reading, writing, and arithmetic universally
isn't enough. Machines can do such things. In
many cases they can do them better. We now
need a different kind of personality, a person
who is going to decide what a machine ought
to do, or to ask a machine to do tremendous
jobs for him. When programming a machine
for a problem, a person must consider all the
ramifications of what, say, 15,000 clerks may
have handled separately, for the machine is
the equivalent of a staff of 15,000 clerks.
Therefore, his type of thinking must be not
the mechanical clerical type any more, he must
use executive ability and creative thinking.
Suddenly we are faced with the fact that we
must educate most of our population to be
executives and creative thinkers.
Further, the trend toward specialization
must erase itself too. The reason for specialization in the first place was that there was
too much detailed information about the
world to be crammed into anyone head.
Therefore, this knowledge had to be split into
domains, with different people cramming in
different kinds of information. However, the
storage and retrieval of that information and
most of the work involved in it was a mechanical job. If the machines can take over
that mechanical portion then the reason for
the specialization disappears. But then you
have to teach everybody the broad principles
of every domain of knowledge, and the
arrangement of the domains of knowledge in
a taxonomic tree. They would have to go on
from there to retrieve any details they needed;
but they would program a machine to handle
that retrieval for them. In other words, we
must now learn how to educate a whole generation of Leonardo da Vinci's. Do we know
how to produce even one?
Even proving theorems in geometry is by
and large a mechanical job. When I talk to
mathematics teachers I point out that the way
geometry is being taught presents a worm's
eye view of the subject. They have been making the students act like machines in step by
step proofs without ever understanding how
that which they proved was conceived of in
the first place. One has to balance off this
worm!s eye view of mechanical proof with
almost instinctive flashes of insight obtained
by a bird's eye point of view of what a subject
is like and what is likely to be true in the
field.
STRUCTURAL CHANGES IN
DOMAINS OF KNOWLEDGE
Let us now consider the particular causes
of these periodic violent changes in the structure of a growing organism. The simplest
example is the development of the sciences;
how they grew, why they grew, the order in
which they developed and the way that each
one individually developed. The biological
principle "ontogeny recapitulates phylogeny"
seems to be applicable here. We remember
that this principle points out that the development of an individual animal from conception
on takes it roughly through the same stages
that evolution took among the species until
that of the animal was reached. Similarly,
125
the history of the development of each individual science seems to have followed the
same pattern as the history of the development of the sciences as species. The phases
and changes in each science, to my mind, are
due to an information retrieval problem. In
talking to management people, I would characterize the problem as the growth of a file
and what has to happen to it.
What happens is that there must be a balance, as the file grows, between the retrievability of the information and the needed communication flow of the information; the control of the information due to the structure
of its arrangement must balance the method
by which it is communicated. What has happened,· then, is that the continuum of a domain of knowledge when a revolution is due,
either splits into distinct fields or changes
phase radically by a change in the structure
of its arrangem~nt. That is, once the mass of
information got beyond the critical mass,
usually either fission or fusion occurred.
Fusion is the change in the structure of the
individual science itself to make it a much
more compact carrier of information; for
example,. general laws are highly compact
bouillon cubes of information. Fission is the
breaking up into various areas of specialization, which we have already mentioned. The
fission or fusion occurred in each case because
the information got beyond our capacity, and
its former control began to require too much
time to retrieve. In order to be able to retrieve
the information, you either have to have separate specialized lumps or have a violent
change in structure.
A CLASSIFICATION OF SCIENCES
Let us, then, look at the ontogeny and phylogeny of the development of the sciences.
This particular point of view was that of the
positivistic school of philosophers and scientists led by Auguste Comte over a hundred
years ago. Since they were developing a new
field called sociology and they considered themselves philosophers, they put the different
sciences in a scale with philosophy and social
science at the top. Mundane sciences like
physics and mathematics were at the bottom
of the scale. However, they did agree that
mathematics had a more crucial value than
that indicated by being at the bottom of the
scale. In my scaling of the sciences I consider
126
that philosophy has a special role (as you
might guess from the tone of this talk), and,
further, that mathematics has a special role
also (as you might have guessed if you know
my background).
Originally the wise men of the tribe in a
small cultural unit had to carryall the basic
information, the natural philosophy that the
tribe needed at critical moments. This was all
one domain. Knowledge was considered, as it
should again be considered, as one continuous
domain. However, because of the ever repeating information retrieval problem, this
knowledge was fragmented into different
domains. History indicates a natural
chronological sequence.
Mathematics seems to have been the first
field to have split off and to have developed a
life of its own; then physics developed. The
more powerful the mathematics available, the
faster the physics developed. Then there was
chemistry, and the more powerful the understanding of physics, the faster chemistry
developed. Next in order follow the biological
sciences, the psychological sciences, and the
sociological sciences. It seems reasonable that
these sciences should have developed in this
order when we examine the unit of discourse
in each of these sciences. The unit of discourse
in physics at one time was the atom, a rather
small unit. However, that has been fragmented
recently, fission and fusion again. The unit of
discourse in chemistry is a molecule composed
of atoms. You have to have sophisticated
knowledge from physics in order to begin to
get certain very basic knowledge in chemistry.
The unit of discourse in biology is the living
ceIl, an incredibly complicated group of molecules. Consequently you have to have very
sophisticated information in chemistry before
you can begin to get the really basic information in biology. The unit in psychology is the
individual, and a human being is an incredibly complicated concatenation of biological
cells. The unit in sociology is a structured
group of individuals with some kind of communication and control structure among them.
Again we have to have rather sophisticated
information in biology before you can get
very basic information in psychology. In fact,
the pressure of the needs of information in
psychology are such that it had to develop in
two directions at once. The psychologist
looking from the point of view of a sociologist
developed into one kind of school. The psychoanalysts are an example of this. The
psychologists' thinking of the individual as a
biological system developed the other point
of view; deterministic psychologists, for
example, developed here. They both have to
work at the same time because of the pressure
on both sides.
MATHEMATICS AS AN EXAMPLE
OF ADVANCED DEVELOPMENT
Mathematics happens to be the oldest of
the sciences, and since I believe that the
ontogeny-phylogeny principle in biology is
really a principle of the development of growing organisms of any kinds, we should see
that principle at work in mathematics. If we
look at the phases in the development of
mathematics, we can see what is happening
and can begin to predict the sort of thing that
will be happening among the other sciences.
Originally, even within the memory of man,
mathematics was talking about the real world.
Originally, of course, there was the question
of control of communication about the envir~nment. Since one type of new society was
then agricultural, there had to be information
about continuous quantities. That is, they were
concerned more about measurements than
about counting. When the Nile overflowed
and the boundary lines were wiped out, they
had to be able to get the boundary lines back
again after the flood subsided. Or they had
to construct a pyramid with an exact rigqt
angle, with one side pointing due east and
west and another due north and south, etc.
If the society happened to be not primarily
sitting in one place to control the environment
but moving around and communicating, then
the emphasis was on counting discrete items.
For instance, nomadic groups had to be able
to keep control of numbers of goats or the
number of members of a tribe. Thus they
stressed counting; the other society stressed
measuring. In the urban revolution both had
to deve10p together for the purposes of
commerce.
COLLECTION OF DIREOT FACTS PHASE
Thus the rules that were used in both
counting and measuring were directly connected with physical entities. They had to file
away the facts that two sheep plus two sheep
was four sheep, even this is already an
abstraction, and two dogs plus two dogs was
four dogs, and two beans plus two beans was
four beans in the information system. This
filing system soon became unwieldy. They
must therefore contract the structure by
creating the abstract concept, 2, and in the
filing system they have just the one statement
2 plus 2 equals 4, and then also they have a
large list of digitalized phenomena to which
counting applies: dogs, sheep, cattle, people,
and so on. Now they have in a very compact space, many sentences replaced by one
sentence, and a list of appropriate replacements in such sentences. That long list
together with the one sentence took up much
less information space or retrieval time than
the original did. This achievement of compactness indicates the general pattern.
At this first level of abstraction, these concepts, numerical concepts for example, were
still considered rather concrete. As mathematics goes these days they are close to a
concrete level, mathematics being only rarely
fascinated by particular numbers. In those
days there were different schools of thought
for the rules to be used with these first level
abstractions. For example, in constructing a
right angle, there was a school of thought that
claimed that if you take a rope and tie two
knots in it and form a triangle with the three
portions three units long, four units long, and
five units long, respectively, then the angle at
the knot that is opposite the side that is five
units in length will be a right angle. Further,
anyone in his right mind would know that
that is the way to arrange things if he wants
to get his pyramid squared away. There might
esaily have been another school of thought
which went in for sides which were five,
twelve and thirteen units in length. This
school might have believed that with these
bigger lengths they wer~ bound to get a more
refined, right angle than with the 3, 4, and. 5
system. It is conceivable that they got into
arguments and occasionally killed each other
in the course of these arguments; they took
these things seriously in those days. It is a
recorded fact that the Pythagoreans killed
a member who revealed to the outside world
that, contrary to their doctrine that all nature
harmonized by having all quantities ratios of
natural numbers, the square root of two was
irrational.
Thus this fact was added in the mathe-
127
mati cal filing system and the amount of
information about right triangles began to
pile up: 3, 4, 5; 5, 12, 13, and so on. The filing
system began to get to the point where it
developed a critical mass and either fission
or fusion had to occur. In this case it was
fusion, the Pythagorean theorem. The result
was not a list of what can be substituted for
the lengths of the three sides of the triangle
but rather a statement about the relationship
the numbers in such a triple must have. Thus
we have an infinite number of potential facts
in that one sentence - a terrific condensation
of information. Now when this one sentence,
the Pythagorean theorem, was put into the file,
the structure of that filing system condensed
remarkably. The average retrieval time was
extremely fast, although in any particular
case in addition to the looking up of the theorem you would have to pay the price of
computation time to verify the numbers
considered.
Recall that the first phase was just the collection of facts from experience. This was a
file with the limited structure of a single
sequence, wherein you are lucky to find something. You had to hunt through this linear
file from beginning to end, the way a person
of the Ancient Mariner type gives you. advice.
You ask him for advice, and he'll give you an
autobiography; and when the relevant information that you really want passes by, you
stop him.
After this (linear file) collection of facts
phase, we evolved to the selection of the
appropriate abstract concepts, in other words,
the selection of the right grouping of things
to be observed. We had to learn what it was
that we ought to be perceiving. At this stage
of the game it was decided that numbers were
important, that measurements were important and that the concept "animal" was
important, whether talking about sheep, dogs
and so on. Just as a child has to learn that
the concept dog is important when he sees
Towser in one place and Fido in another place,
etc. In the classification phase one is inventing,
not discovering, the combinations and groupings that will permit a tighter and more
compact arrangement of information. In
botany it happened about 200 years ago with
Linnaeus. A new language, a new jargon, was
developed for that information. This language
permits classification of millions of things in
128
an appropriate fashion for an appropriate
purpose; we call this the taxonomic phase.
GENERAL LAWS
This same thing was happening in mathematics when it was decided that the concept
of an angle or a triangle was important. Thus
there was developed a filing system in which
a portion was concerned with rightangles,
with triangles, circles, etc. For example, a
right triangle would be obtained if the three
sides of the triangle had lengths of 3 units,
4 units, 5 units; 5 units, 12 units, and 13 units,
and so on. Again· the file began to get too big,
fission or fusion had to occur. In this case it
was fusion again so that we derived general
laws which like the Pythagorean theorem
cover a fantastic number of facts, actual or
potential, in one statement.
The more pragmatic mistakes one makes in
selecting the collections to be given abstract
names in the classification phase, the more
complicated the general laws become. The better your choice of taxonomic groups, the
neater the general laws become. Very often
after a certain amount of generalization has
been done, it would be decided that certain
previous concepts in the classification system
should not have been considered to be the
important ones at all. That is why I stressed
the fact that classifications were invented
rather than discovered. A grouping would be
rejected if by using it the description of a
situation or how to extend beyond it becomes
much more complicated than when compared
to what would happen if another grouping
were used. Thus facts themselves change
pragmatically, since the things that are to be
perceived have to change; consequently there
is a feedback between descriptions of experience about things perceived and the decisions
on what should be perceived. Neither ideas
in Plato's sense, nor objects in the sense of
materialism, are merely out there waiting to
be passively perceived. Perception and knowledge are together a dynamic interlocking
process; asking which is prior, mind or matter, is as absurd as asking which came first,
the chicken or the egg.
As classifications change properly, the
general laws become neater. This is important,
because it indicates why neatness and esthetics
are the criteria used in the development of
the world of communications, whose key
creators are our artists and scientists. The
function of the arts and sciences is to sharpen
the tools and substance of communication.
These creative communicators with their
esthetic criteria are considerably closer together in aims than the world or they themselves realize. They are almost diametric
extremes to the control type represented by
executives, army generals, etc., whose working
value system must be almost exclusively
ethical rather than esthetic. It is the professionals such as doctors, lawyers, teachers,
engineers, etc., who have the difficult task of
transforming information into action.
DEDUCTIVE SYSTEMS
Now these general laws begin to pile up.
We have the Pythagorean theorem, we have a
theorem about base angles of isosceles triangles, and a host of other theorems. An
enormous file beings to develop. Again it gets
out of hand, and either fission or fusion must
occur. In this case again fusion occurred.
Euclid discovered that if one began with a
few general laws and certain logical rules
ab~ut how to derive new ones from the old,
one could derive all the general laws. This was
the development of deductive systems.
Portions of physics have gone through the
deductive phase since Newton's time, the area
called rational mechanics is an example. Most
of physics has gone into just that stage now.
Chemistry is only beginning to enter this
phase. So is biology if one looks at the work of
Rachevsky.
The deductive systems began to pile up
during the 19th century. The big stimulus was
the attempt to prove Euclid's parallel postulate; the best minds in the world were not
able to do it. Then it was discovered there was
a good reason why it couldn't be proved. This
axiom was independent of the others. To
show that it was independent, perfectly selfconsistent geometries were discovered that
were not Euclidean. Thus there were developed many different deductive systems in
geometry alone. The latter half of the 19th
century and the beginning of the 20th century
resulted in a tremendous proliferation of
deductive systems. Again we faced the information retrieval problem which is always
with us.
FORMAL SYSTEMS
It became apparent that either fission or
fusion had to occur. For a while it was fission
into various geometries, types of algebra, and
types of analysis; but then it was discovered
that many of these deductive systems could be
made to fit into exactly the same forms. A
divorce of some of the key mathematical concepts from the real world came in when one
deliberately made some of the wor Id used
meaningless in order to obtain such formal
identifications. Thus by putting different
meanings into those words, one got different
deductive systems from one system. This
introduced formal systems, complete formalization. Pure science was completely formalized; applied science put some of the
semantic content and meaning back into such
syntactic formalized systems by assigning
them to these undefined terms and checking
whether the axioms became true. This was
the new meaning of the word application.
For example, let us look at developments in
projective geometry in which points and lines
are considered meaningless (deliberately) .
The only permissible interpretations are those
which can be obtained from the axioms alone.
Proofs are derived from logical rules applied
to these few sentences (the axioms). You are
supposed to think of nothing else at the risk
of bringing in an irrelevancy and not having
a good proof. Yet one application to the real
world of this formal system would be to have
"point" meaning a position in space, and
"line" meaning the path of a light ray. Another application would have "point" meaning
a member of the board of directors of a company, and "line" meaning a subcommittee of
the board of directors. If you take some of
the first axioms of projective geometry with
such an interpretation, 10 and behold, you are
reading the charter for the board of directors;
for example, "Any two people of the board of
directors must be on one and only one committee together". This is the same as saying
"two points determine aline". And if you
take the word "point" to mean a certain field
in the country and "line" to be a group of
fields to be dusted in one day, the axioms of
the geometry now read like a crop dusting
contract. There are many other applications
of the same formal system.
Similarly, in algebra there are highly
abstract systems for groups, fields, rings, and
algebras. And for such systems there are
theories on the structure of the set of subfields
of fields, subgroups of groups, and how to
129
extend them. It is this type of theory which
can be important in solving the problem of
how to make a taxonomic language grow
under control in an information retrieval
system. If you cannot expand such a system,
it will rapidly lose its value.
Such then are the trends that we see in the
development of mathematics; but we can also
expect to see them appearing in all the
sciences. Each will have to go through those
various phases, if the principle of ontogeny
recapitulating phylogeny applies to the various sciences. My claim is that over and over
again it was the information retrieval problem in a given domain of knowledge which
brought about these changes in phase. When
the domain of knowledge got too big to be
controlled by the old system, the whole communication system had to change, whether by
fission or by fusion.
The picture of the sciences which I have
just given, you can recognize, is a naive one
because it seems to imply a linear arrangement. Further, you now see why I considered
mathematics separately. As anf science gets
to a completely formalized state the completely formalized part is mathematics and is
considered as such. And what is philosophy
doing outside of this arrangement of the
sciences? It is examining the basic concepts
and axioms in all of the sciences, comparing
theHl for consistency, sitting on the sideline
being critical and sneering; a very important
function. For sitting off at the sidelines as it
does it can see if the basic concepts are not
jibing properly, look clumsy, etc. It has had
a tremendous effect, but nothing is harder to
read than last year's philosophy. For if philosophy has done its job properly, we can no
longer see the problems that it is discussing,
having already changed our methods of perception in accordance with its criticisms.
Philosophy has still a controlling influence on
the structure of the domains of knowledge
and how they fit together.
INTERDISCIPLINARY SCIENCES
But in this day and age we see the development of many interdisciplinary sciences. Thus
there isn't a linear alignment of science as
described above. We should actually be thinking of the domains of knowledge as originally
thought of; a unified and continuous host of
information. This information we quantize
according to our needs, marking out those
domains that seem to be relevant at any given
time. Those domains also change with the
times. Moreover, as scientific information becomes much denser, the lines of demarcation
separating erstwhile domains are no longer
as clear as, say, the separation of galaxies
in a thinly popula,ted universe.
Possibly more important now than the traditional separation of scientific knowledge
into the classic sciences is the distinction between the pure formalized science (mathematics), the applied science, and pragmatics.
The formalized is concerned with pure syntax,
the relationships among symbols independent
of their meaning. The applied is concerned
with the semantics of the words used, that is,
the relationship between the symbols and their
meanings. We are also concerned, however,
with the relations between the symbols and
their users or interpreters. This is pragmatics.
Computer and information sciences can be
distinguished from the mathematical portions
of science precisely in its concern with pragmatics. The emphasis is on how interpretation
is controlled, and how interpretations shift.
The fact that I couched this general picture
in information retrieval terms is due to the
fact that I have been considering scientific
information pragmatically. I believe you have,
too. In this new domain of computer and
information sciences we are studying a growing mechanical language. However, in order
for us 'to know that we are studying pragmatically important concepts we need information from you. It is possible to carve out a
concept which would not be pragmatically
important and spend our days doing research
on it. Yet, practitioners of information retrieval systems are the ones that have to meet
pragmatic problems day to day. There are
various studies of mechanical languages which
are interesting and some which are useful.
However, as matters stand, in studying pure
theory, we are in the position of a man who
was stuck with an elliptical billiard table two
feet high, with a leopard skin top. He made
one because someone had ordered it, but that
perSon didn't want it any more and cancelled
the order. The pure scientist is in the position
of having a load of these tables. If anyone
happens to want one, we have them.
/
130
By E. M. McCormick
National Science Foundation
18. The Management Process
and Science Information Systems
THE USUAL MANAGEMENT CONTROL CYCLE is
familiar to scientists since it is essentially the
scientific method itself. The first part is the
determination of goals and objectives for the
system being considered. The second part of
the cycle is concerned with evaluation. Here
the existing system is compared with the goals
and objectives of the system and a decision is
made on action to be taken. This results in
.implementation of the revised system. As
operation of this system continues, it is important to collect and organize informatior
which can then be used for further evaluation
by comparison with the goals and objectives
to determine the value of the system. This
may result in different goals, decisions, implementations, operations, etc., which cause the
cycle to be repeated as often as required to
set up the desired management situation and
to maintain it under dynamic conditions.
The management control cycle can be considered in two ways with regard to science
information centers. The first is as a part of
the overall process involved in scientific research and development activities. Thus, the
science information center is used for collecting and organizing the information for
management decision. Further, the management situation must be one in which scientific
and technical information are significant parts
of the management process. As such, it represents the viewpoint of the boss or the manager
of the science information center.
The second way of considering the management control cycle is in terms of the science
information center itself. Thus, the goals and
ojbectives are those of the center, not those of
the overall activity. The operations are those
of the center itself, etc. This is primarily the
viewpoint of the manager of the science information center who is concerned with
developing an optimum system to provide
service to the overall organization.
As has been indicated, these viewpoints
have much in common. Thus, much of the
material presented here applies to both. However, it is helpful sometimes to consider
whether any specific management situation
pertains to the overall system or to the science
information center itself.
Before the principal phases of the management control cycle are considered individually,
it is necessary to review certain aspects of
the management situation in which science
information centers exist.
THE MANAGEMENT ENVIRONMENT
Just what is the situation which makes the
management of science information centers of
importance to management? Traditionally,
the technical information function is' handled
by providing library services to the staff. In
general, this is a relatively small part of
management responsibilities. However, in
many cases the handling of scientific information has become a much more significant
problem in recent years. The reasons for this
are discussed in some of the papers in this
book. Briefly, business, government and industry have found that their products and
services depend more on the information
produced and used by scientists and engineers.
Management itself finds that more of its
decisions are based on technical decisions.
Further, the cost of this information is increasing and its management becoming more
difficult.
131
In these papers the emphasis is on the
management problems which differ from the
general management problems of scientific
research and development activities. Thus,
much of the material must be considered in
terms of the existing management environment of scientific research and development
activities. Further, the point of departure or
reference standard in considering technical
information centers is the conventional library. Most of the papers are concerned with
how these centers differ from or compare to
these libraries.
SCIENCE INFORMATION CENTER
DEFINITION
The essential elements of a science information center are identified by several of the
authors. The common factor in these definitions is the aggressive attitude toward
information collection and distribution. This
is contrasted with the comparatively passive
operations of the usual library. The technical
information center volunteers its expertise
to the scientist. It offers a wide variety of
services to its customers and is generally
aggressively striving to render more and better service. In many ways, of course, it is a
special library in that its collection is narrow
in breadth, comprehensive in depth, and
directed to sophisticated clientele. However,
the technical information center is more likely
to use nonconventional methods of handling
information. This is probably due to the fact
that the items in the collection and the demands made on the system are not always
amenable to conventional techniques. In addition, these technical information activities
frequently are originated to meet technical
needs by technical people who are familiar
with newer techniques for data processing
and information handling.
GOALS AND OBJECTIVES
It is apparent that one needs first to
determine the reasons for having a science
information center. What is its purpose? How
will its function contribute to the overall
objectives and goals of the organization?
Then: what must be the goals and objectives
of the science information center itself?
Although obviously important, it is generally difficult to set meaningful goals and
objectives for science information centers.
132
Many just grew from library operations. Yet
the reasons for this growth need to be investigated to determine what it means in terms of
the goals and objectives of the company. Perhaps they have also changed. This is illustrated
by situations in which management has
changed the fields of interests of a company
without informing the library or science
information center so that it would be
prepared to serve these new needs.
INFORMATION NEEDS AND USES
The first requirement in setting up a system
is to determine the need for the service. Yet
this is quite difficult to do, and many systems
are in operation with certain assumed needs
for the service but where there is no good
measure of the need. Some user studies have
been made but these are not conclusive with
regard to the need for information. It is
pointed out that the way scientists use an
existing system is not necessarily the way
they would use a better system. Scientists are
conditioned to respond to a system in anticipation of the system response. This results in
their not asking the question they need an
answer to but rather asking another question
which they feel can be answered and from
which the desired answer can be derived. This
results in some doubt that the actual information needs of a system can be determined by
the scientist himself. It would appear that
market survey techniques could be useful in
this situation.
In considering this aspect of objectives and
goals it will be noted that three papers referred to Mooer's Law. This states that, "An
information retrieval system will tend not to
be used whenever it is more painful and
troublesome for a customer to have information than for him not to have it". It is
suggested that this law is not entirely facetious and merits consideration when planning
the services a science information center is to
provide.
EVALUATION AND DECISION
The next step3 in the management process
are concerned with evaluation of the system
and the decision on a better system. Evaluation of an information system is fraught with
difficulties. Yet these difficulties are not so
different from those with which management
is already familiar. Although objective evalua-
tions are only rarely possible, subjective
evaluations are frequently made for management purposes.
Evaluations are made in terms of goals
and objectives. Thus, since science information
centers are established to directly benefit the
scientist in his work, evaluation is also concerned with the user. All systems provide for
feedback from the user; quite often this feedback is informal. However, more sophisticated
techniques are now being applied to this problem. Many system studies go beyond the "how
did you like it?" type of survey. There must be
dependence on opinion, but management has
generally recognized the importance of intuition over logic in many complicated situations.
Management's interests are served somewhat indirectly, i.e., through the scientist.
Thus, management must not only consider the
user's evaluation but also whether the assumed user need fs valid. Again it is quite
likely that the informal information system
may be much more significant in comparison
with the formal than is sometimes presumed.
ECONOMIC JUSTIFICATION
The economic justification for information
activities, especially in an industrial environment, is considered by several of the authors.
In addition to the general factors which must
be considered, the specific problem of justifying a mechanized information system is
presented in some detail.
A part of this economic justification involved determination of what can and should
be done within the company and what should
be done outside. Many of the existing information systems are primarily concerned with
internal information where general disclosure
would be prejudicial to the company. Yet an
increasing number of commercial activities
are being established to handle nonproprietary
information where there is an economic
advantage in handling this common information. Since these endeavors are competitive
and must be self-sustaining, their pricing
of various information functions is helpful
in management evaluation of the cost of
information systems.
MECHANIZATION
Many of the problems in evaluating an
existing information system and in making a
decision on a new system are encountered
when mechanization is considered. There is
much management interest in the question.
To what extent is mechanization the answer
to many of the problems in processing technical information? This topic is considered
thoroughly in one of the papers and referred
to in others. The approach is the same as for
automatic data processing systems for commercial applications. Of course, there are
technical differences between these two fields,
but primary management considerations are
similar. The experience in data processing
should be helpful for information handling.
The successful application of mechanization
in information must first meet the requirements for data processing. Beyond this there
a~e additional requirements for information
handling. The emphasis in most of the papers
is on these special requirements. Again the
presentations are in terms of the usual
management situation in science and technology which presumes some knowledge and
experience in automatic data processing.
IMPLEMENTATION AND OPERATION
After the existing system has been evaluated and a decision made on a new and
presumably better system, there are many
considerations involved in implementing this
decision and in operating the system.
Organizational location
The position of the technical information
system in the management hierarchy is considered in some of the papers. The importance
of having it as far up in the organization as
possible is stressed. The factors which result
in the responsibility for this service being
pushed upwards in the organization are also
considered.
It is suggested that technical information
centers do not automatically get the support
of top management but rather earn it by
demonstrating their value to the organization.
On the other hand, these centers are more
likely to be recognized and use made of them
when it is known that top management solidly
supports them.
Of course, there are situations where the
center has developed to meet a special need
comparatively low in the formal organization
chart and have stayed at this level. However,
there are also examples where the need has
133
been recognized first by management and the
resulting center was established to meet these
needs.
In addition to position in the formal organization, the position in the informal
organization is also important. This is one
reason for having technical peers in the information center - it permits better access to
the important informal organization thus
getting around organizational lines which
sometimes hinder the communication process.
Personnel administration
Personnel problems were considered by
several of the authors. It was felt that there
were special requirements in the technical
information situation which merit special
consideration.
There was general agreement that there is
a need for technical personnel in the technical
information center. These individuals should
first have competence in the fields of science
and technology represented in the center, and
second should have skill in using the tools of
the center. The principle need for this competence seems to be in handling the contacts
between the center and its customers. It is
deemed necessary for the customer to be able
to deal with someone in the center who can
talk his language and understand his problem.
It appears that much of the benefit of this
situation is not necessarily the actual technical
competence of the representative of the information center, but that he is or represents a
group that the technical customer would
accept as a peer. This peer concept seems to
be quite important in handling the basic communication problem between the center and
the customer. Again the contrast to the usual
library environment can be noted.
The peer concept indicates that the center
representative is interested in the user's problem as such rather than in the integrity of the
sources of information. The trend here is
analogous to that in ADP - one obtains personnel who are primarily skilled in the technicalities of the business itself and then one
trains them to use the tools of the service
activity (ADP systems or technical information systems) which are of use to the business.
This has obvious advantages in the integration
of the service activity with the other activities
of the organization.
134
There is a question, however, as to whether
this technically competent individual should
be identified as a subject matter specialist or
as some form of information specialist. Some
consider that there is or should be a profession
of information specialists. Although they may
have technical backgrounds, their main occupation is handling information competently.
Thus, in a sense they would most resemble
librarians although possibly trained in more
of the new technology of information processing. It is also proposed that the technical
information field needs personnel whose basic
skill is in researching, analyzing, and documenting the significant aspects of various
fields of science and technology. These writers
would produce state of the art reports which
would serve many of the basic information
needs of scientists, especially with regard to
sifting out the important from unimportant
material.
The other end of the gamut is the situation
where the information function is handled by
a researcher who does not lose his basic
identity as a researcher. Perhaps one member
of a team would spend more time than the
others in the information gathering and disseminating capacity. However, he would do
this as a secondary activity and would probably not have any special training in information processes. Although this may be inefficient as compared to a technical information
center, it is defended (at least in part) by
the fact that the scientist retains his first class
status by continuing to be identified as a
scientist. Otherwise, as an information specialist, as such, he would become a secondclass citizen. Obviously, the situation is
different in various types of scientific and
technical enterprises.
This peer concept, in its various forms,
emphasizes the need for prestige and status
for personnel in' the technical information
center. Of course, management recognizing
this need can meet it by these as well as
various other techniques.
Financ'ial support
The relatively high cost of some information
centers as compared to library services aggravates the problem of financial support. Most
centers are supported as a part of the overhead account as service activities. However, in
one of the papers there is a report on the prob-
lem of support of an information center and
of obtaining some measure of the value of its
services by making charges for these services.
Only partial support of the center is attempted; the rest is a direct subsidy as is the
usual practice in the operation of information
centers and libraries. This technique has disadvantages as well as advantages; it will be
of interest to see how well it works out
over a period of time and whether other
information centers adopt the idea.
Public relations
Libraries have a traditional image of service
but do not normally engage in an aggressive
campaign to sell their services. Yet science
information centers seem to be characterized
by an aggressive attitude toward the collection of information and toward the dissemination of information to the users. This results
in selling, promotion and marketing activities
which are not those usually associated with
a library.
COLLECTION AND ORGANIZATION
OF INFORMATION
In order to close the management control
loop, it is necessary to collect and organize
information about the operating system that
will permit further evaluation, modification
of goals, and revision of the system if necessary. Again the problem is familiar to
management. Further, the need for obtaining
valid cost information is obvious.
Cost information
The precedent for technical information
service is the traditional library function.
This was considered to be a necessary part of
the overhead operation. As such its support
was a relatively small percent of the overall
cost. Yet as the technical information function
has grown, the cost of the service has in some
cases increased to several percent of the
operating budget. Yet, there are no good
measures of the value of the information
which could be compared to the cost of information. Actually, in the cost accounting sense
there is very little information on detailed
costs of library operations. For that reason
one of the papers gives information on costs
in a successful mechanized and integrated
technical information center. The presentation
here is costing in the sense that an operating
profit-making activity must be conscious of
cost. Conventional information activities normally do not sell their product, hence, are not
in a position to be as cost conscious as they
might otherwise be.
In addition, often many of the costs of
technical information are hidden. When there
is a well-defined information activity, it is
possible to determine costs with reasonable
accuracy. However, in many situations the
information function and the research and
development function are combined making it
difficult to separate these costs. The policy of
considering information to be an essential
part of the research process tends towards
this. However, efforts are being made to
identify the costs of science information
activities. At the national level some government agencies are identifying them as line
items in their budgets and efforts are being
made to have other agencies also do this.
Information consolidation
In collecting and organizing information
about systems, it is suggested that the precedents from data processing experience may
be misleading in science information activities. Data processing is generally successful
in meeting the needs of management because
it is possible to consolidate data in the system
to correspond to the various levels of management. This enables management to judge data
about the system used for data processing.
However, in science information centers
the basic commodity is information which, in
general, is not capable of being consolidated
as in a profit and loss statement. It is generally
only available to and useful to the scientists
and engineers themselves, not to the higher
levels of management. Thus, their information
about an information center is not as certain
as with data processing. This presents special challenges in gathering the information
needed for the evaluation of information
systems.
135
Bibliography
(The bibliography is selected from "Information Handling and Science
Information - A Selected Bibliography", American Institute of Biological Sciences, Washington, D.C., B.S.C.P., 1962, and reproduced with
permission of the Biological Sciences Communications Project of the
American Institute of Biological Sciences and the Center for Technology
and Administration of The American University).
Adkinson, B. W., THE FEDERAL GOVERNMENT
AND U. S. SCIENTIFIC INFORMATION. J. Chem.
Doc. 2(1 ):48-50. Jan. 1962.
In addition to presenting a general review of the
scope of the Federal Government's role in research
and development, and a review of the character
of the Federal Government's scientific information
activities, the article presents a review of the sixpoint program which has been developed by the
National Science Foundation for improving the
dissemination of scientific information, particularly among U. S. scientists and engineers. The
author concludes that the dissemination of scientific information must be recognized as an integral
part of research and development.
Adkinson, B. W., UNITED STATES SCIENTIFIC AND
TECHNICAL INFORMATION SERVICES. Spec. Libr.
49(9):407 -414. Nov. 1958.
This discussion of United States scientific and
technical information services is divided into the
following parts: the present situation in the
United States; some of the reasons behind the
current problems; the. principal scientific and
technical information activities of the Federal
Government; and a proposed program to remedy
this country's scientific and technical information
ills.
Agard Evans, A. B.,and Farradane, J., TRAINING
THE SCIENTIFIC INFORMATION OFFICER. Internatl.
Conf. Sci. Inform. Proc. 1958(2):1489-1494. 1959.
Points out the difference between the function of a
scientific information officer and a librarian. A
syllabus is presented for a post-graduate course
of training for a student who is already a subject
specialist.
American Management Association, Administrative Services Division, ADVANCES IN EDP AND
INFORMATION SYSTEMS. New York, 1961. 1 87 p.
(Report No. 62).
Papers in this volume are "based on material
originally presented at AMA's Seventh Annual
136
Data Processing Conference, held in March 1961".
Partial contents: Advances in information retrieval and data acquisition: I. Progress in the
design of information retrieval systems, by M.
Taube, p. 51-63; II. Improved information storage
and retrieval systems, by B. E. Holm, p. 64-74.
Asmonas, Y., SYSTEMS OF SCIENTIFIC AND
TECH~
NICAL INFORMATION SERVICES: LONG RANGE
PLANNING. Rev. Doc. 27(2):81-85. May 1960.
A proposal for the establishment of an international technical information system which would
be based on three main international centers (one
for the Americas; one for Europe, the Near East
and Africa; one for Asia and Australia). In addition, there would be a national center for each
country, and a center for each scientific or technical discipline. The paper also includes discussion of
problems which would be encountered in the development of such a system.
Atherton, P., and Clark, Y., A SUGGESTED CLASSIFICATION FOR THE LITERATURE OF DOCUMENTATION. Amer. Doc. 12(1 ):38-48. Jan. 1961.
The classification scheme presented here had its
origin in a class project undertaken by students at
the University of Chicago, Graduate Library
School during the spring quarter 1960. A description of the principal features of the classification
with reference to .David J. Haykin's six basic
building rules for special classification is followed
by two appendices: I. Classified Guide to American
Documentation 1950-1960, vol. I-XI; II. Suggested
classification for the literature of documentation.
Bach, H., SCIENTIFIC LITERATURE USE: A SURVEY.
Spec. Libr. 48(10):466. Dec. 1957.
The author draws from reports by Bernal, Herner,
and Urquhart on the ways in which scientific personnel gather information, some observations as to
certain general practices and attitudes among
scientists which he believes should be kept in mind
by those who manage science libraries. ©ADI
Balke, S., WIRTSCHAFT UND DOKUMENTATION
[Documentation and economy]. Nachr. Dok. 12
(1): 1 -5. Mar. 1961.
limitations of machines in this field, and a presentation of three methods for accomplishing machine
indexing.
Fundamental problems of documentation are discussed from the professional and organizational
point of view. It is stated emphatically that the
expert must be provided with the material for his
work whatever the financial cost. Development in
documentation in West Germany must take the
form of cooperation between institutions staffed
by highly qualified personnel. Complete centralization as in the Soviet Union is neither possible nor
desirable. Documentary activity will be coordinated at national level by the proposed Institut fur
Dokumentationswesen. ©ADI
Booth, A. D., THE EFFICIENCY OF CERTAIN
METHODS OF INFORMATION RETRIEVAL. Inform.
& Control 1 (2):159-164. May 1958.
Becker, J., INFORMATION PROCESSING. Washington, D.C., Amer. Inst. BioI. Sci., BioI. Sci. Commun.
Proj., 1960. 13 p.
Describes the information framework in all special
libraries, public libraries, and information centers
as containing the following elements: the information received which is usually in printed form, the
analysis of this information for indexing and
cataloging, the index to the information, and the
stored information or file of information. Discusses various efforts to mechanize each of the
elements in the framework.
Bello, F., HOW TO COPE WITH INFORMATION.
Fortune 62(3):162-167, 180, 182, 187, 189, 192.
Sept. 1960.
... Information retrieval is reported to be in 1960
just about where electronic data processing was
10 years ago. A history is given of "modern"
information retrieval, beginning with Vannevar
Bush, head of the Office of Scientific Research and
Development, with his article, "As we may think",
in the July 1945 Atlantic Monthly. Since then, information retrieval has attracted two general
types of enthusiasts, those interested in building
machines and those interested primarily in theory,
with the shift in the last few years to debates on
indexing and retrieval theory . . . The article
chronicles worldwide activities in retrieval from
the Rapid Selector of Ralph R. Shaw to the Itek
Corp. system of the Minicard type for the Air
Force, the computer-produced index Chemical
Titles of Chemical Abstracts Service, and the
"biggest single abstracting service in the world",
VINITI, the All-Union Institute of Scientific and
Technical Information of the U.S.S.R. ©ADI
Bohnert, l. M., NEW ROLE OF MACHINES IN
DOCUMENT RETRIEVAL: DEFINITIONS AND SCOPE.
Inst. Inform. Storage & Retrieval, 3rd, 1961.
Papers presented, p. 8-21. 1962.
A discussion of the role of machines in indexing
and document retrieval, and a review of the recent
research in this area. The article includes definitions for the terms "document retrieval" and
"machine indexing", as well as a discussion of the
Methods are discussed for the retrieval of items
of data which are stored as entries in a table in
the store of a computer. The average number of
look-up operations required to find an entry is
computed for several methods. It is shown that if
advantage is taken of the relative frequencies
with which the different entries are looked up, the
average Dumber of look-ups may be substantially
reduced. The results are applied to the problem of
using a computer as a mechanical dictionary.
Author.
Bourne, C. P., THE HISTORICAL DEVELOPMENT
AND PRESENT STATE-OF-THE-ART OF MECHANIZED INFORMATION RETRIEVAL SYSTEMS. Amer.
Doc. 12(2): 108- 110. Apr. 1961.
The last ten years of equipment development and
the application of mechanization techniques are
reviewed for each of several functionally separate
approaches, such as punched card systems, computer systems, and magnetic media systems. Comments are made on the progress to date, and the
degree of activity to be expected during the next
few years for each of these approaches. Author.
Bourne, C. P., and others. FINAL REPORT: REQUIREMENTS, CRITERIA, AND MEASURES OF
PERFORMANCE OF INFORMATION STORAGE AND
RETRIEVAL SYSTEMS. Prepared for Office of Science
Information Service, National Science Foundation.
Menlo Park; Stanford Res. Inst., 1961. 132 p.
A preliminary study was made of the requirements, criteria, and measures of performance of
information storage and retrieval systems. Using
an interview guide and a methodology developed
during this study, a total of 92 applied electronics
researchers and 11 metallurgists were interviewed
in an attempt to measure and rank several different requirements for information. It was found
that some requirements could definitely be measured, and that there was general disagreement
among the users about the relative importance of
various information requirements. The methodology and the interview guide could be extended,
with minor modifications, to other technical subject fields. In addition to the study of information
requirements, three separate and complementary
tools were developed for the analysis and evaluation of information retrieval systems: (1) a coarse
screening procedure; (2) two different performance evaluation procedures; and (3) two cost
analysis procedures that used computer programs
to simulate the operation of candidate systems to
determine their operating costs over wide ranges
in operating conditions. A general functional
137
model of a storage and retrieval system was developed for use by these cost analysis programs.
A number of specific research tasks were also suggested to further develop the techniques for the
determination of user requirements and the measurement of the performance of information storage
and retrieval systems. Author.
Boutry, G. A., THE ICSU ABSTRACTING BOARD:
THE STORY OF A VENTURE IN INTERNATIONAL
COOPERATION. Internat!' Conf. Sci. Inform. Proc.
1958(2): 1503-1515. 1959.
A brief history and description of the International Council of Scientific Unions Abstracting
Board. Included in the paper are: a detailed discussion of the International Journal of Physics
Abstracting and mention of future activities and
proposals of the ICSU Abstracting Board.
Brossman, S. W., THE CASE FOR CENTRALIZED
INTERNAL COMMUNICATIONS. Res./Devlpmt. 11
(7):17, 19,21,23. July 1960.
Discussion of the need for, and organization,
operation and control of a communications office
with authority not only over printing and distributing, group filing, report writing, and library
cataloging, but also over manual writing, revising,
and cutting, as well as communications systems
and procedures.
Brownson, H. l., THE DOCUMENTATION RESEARCH
PROGRAM OF THE NATIONAL SCIENCE FOUNDATION. AUTOMATIC DOCUMENTATION IN
ACTION. Proc. 1 959: 1 -9. 1961.
The article discusses the background of the N ational Science Foundation and names other government agencies engaged in or supporting
research and development problems in documentation; the work of private foundations such as
the Council on Library Resources; the publication
of the Current Research and Development in
Scientific Documentation pamphlet; the Research
Information Center and Advisory Service on Information Processing at the National Bureau of
Standards; several projects being supported in
information storage and retrieval and five projects
in the field of mechanical translation.
Bush, V;, PROFESSIONAL COLLABORATION.
Science 125(3237):49-54. Jan. 11, 1957.
In this paper on the responsibility and opportunities of the professions, particularly those of
medical men and scientists, Dr. Bush foresees that
the time is coming when the practice of medicine
will rest upon a systematic understanding of the
life-processes in all their complexity. The full
integration of uncoordinated, empirical data may
not come in our lifetime, however. There may be
required new methods of thought, novel ways of
recording and transmitting the accumulated experience of the race, ways as yet unconceived of
bringing to bear on complex problems the interrelated efforts of diverse minds. We may witness
138
new devices as powerful, versatile, and rapid as
digital computers in the realms of computation
and analysis, but capable of interrelating and
ordering masses of primary and inexact observations into meaningful arrays. There may be means
for communicating the knowledge of a group which
will render obsolete the cumbersome writing of
papers and the chaotic task of storing and consulting them. ©ADI
Cahn, J. N., "B.I.T.S." - BUREAU OF INFORMATION FOR TECHNOLOGY AND SCIENCE PROPOSED AS "KEY STATION" OF FEDERAL "NETWORK" OF INFORMATION SERVICES TO U. S.
FREE ENTERPRISE SYSTEM. New York, 1962. 3 p.
Excerpts of address delivered at Engineers Joint
Council, New York, January 17, 1962.
Presents a brief background on government concern with the rapid growth of scientific information. The author proposes a two-fold plan for
Federal action in this area: 1) the establishment
of a Federal network of libraries and information
centers consisting of voluntary collaboration
among four main elements the three great
National library systems, the major inter-agency
information systems, the major intra-agency systems, and the Federally supported information
centers; and 2) the formation of a "Bureau of
Information for Technology and Science" as the
principal point for gathering and distributing
tax-supported, non-classified information to all
individuals requesting it.
Case Institute of Technology, AN OPERATIONS
RESEARCH STUDY OF THE DISSEMINATION AND
USE OF RECORDED SCIENTIFIC INFORMATION.
Cleveland, 0., 1960. 63 p.
A study of three aspects of the problem of dissemination and use of recorded scientific information. Part One is a study of the reading behavior
of chemists and physicists using the direct observation method to determine what is read, why it
is read, and the differences in reading behavior
between groups. Part Two is an economic analysis
of journal publications to determine the average
cost of publishing journals of various types and
to determine the cost per reader of the most frequently read journals. Part Three is a study to
determine how condensation affects the comprehension of scientific articles.
Cattaneo, A., L'INFORMATION TECHNIQUE DANS
UNE GRANDE INDUSTRIE [Technical information
in a large industry]. Rev. Doc. 24(2):88-94. May
1957.
Presents the requirements of the information
service in a large industrial firm and its organization at FIAT in Milan, as well as considerations
for the improvement of technical information.
Clapp, V. W., THE COMPUTER IN THE LIBRARY.
Computer App!. Symp., 1960. Proc. 7:35-45. 1961.
The author concludes that "computer applications
will progressively take place in library work - in
some areas more rapidly than others". He points
out that such applications are challenging to those
engaged in computer work "because they involve
some of the most subtle and difficult factors affecting human activity and accomplishment - namely
information and meaning".
CLARIFICATION, UNIFICATION
OF INFORMATION STORAGE
proceedings of Feb. 23, 1961
York City. New York, 1961. 94
& INTEGRATION
AND RETRIEVAL,
symposium, New
p. Ref.
The symposium was sponsored by Management
Dynamics, Lincoln Square Chapter (N. Y. City)
Systems and Procedures Association, and ScienceTechnology Group, N. Y. Chapter, Special
Libraries Association. The speakers endeavored to
limit the time given to history and to the blue-sky
future. The effect, rather, is intended to emphasize the present state of affairs - concentrating on areas of opportunity for the user, the
researcher. and the manufacturer.
Cloud, G. H., KEY TO PROGRESS - SPEEDY USE
OF PUBLISHED INFORMATION. Monitor 2(3):4, 53.
Ref. May 1957.
Contains brief description of the Houston Research
Library, Humble Oil and Refining Co. Coordinate
indexing is used for report literature. Some 1,600
reports are "deep" indexed by more than 26,000
entries, with 4,500 Uniterm cards in the subject
index. LC classification is used for books. ©ADI
Coblans, H., SOME NOTES ON THE ORGANIZATION OF SPECIAL LIBRARIES. Unesco Bul. libr. 12
(11/12):261-266. Nov./Dec. 1958.
Points out the main differences between general
and special libraries; and discusses matters of
documentary reproduction, acquisition policy,
translations, abstracts and indexes, periodicals and
reprints as they relate to special libraries.
Cohan, L., and Craven, K, SCIENCE INFORMATION PERSONNEL; THE NEW PROFESSION OF
INFORMATION COMBINING SCIENCE, L1BRARIANSHIP AND FOREIGN lANGUAGE. New York, Mod.
Language Assoc. Amer., 1961. 74 p. Ref., p. 42-43.
This study, conducted for the U. S. Office of Education with support from the National Science
Foundation, describes "science information" as a
new profession and the science information specialist as its practitioner; and includes description
of work elements and job titles in the field. The
staff requirements for science information personnel both current and anticipated in government
agencies are tabulated, and a graduate school program leading to an M.S. degree in information is
outlined.
Colmen, K. S., INFORMATION IN ACTION. Res.
Devlpmt. 12(2):6- 18. Feb. 1961.
One of the most important means of improving
communications is by making our everyday procedures more effective as communication media.
Communication of company objectives, R&D
objectives, and philosophic approaches for effective communication within an organization are
discussed.
Columbia University, Bureau of Applied Social
Research, REVIEW OF STUDIES IN THE FLOW OF
INFORMATION AMONG SCIENTISTS. New York,
1960. 2 v. Ref.
Prepared for the National Science Foundation. H.
Menzel, study director; L. Lieberman and J. Dulchin, research assistants. The first volume of this
two-part report presents a review and comparison
of the various types of "User studies" (studies on
the flow of scientific information) which have
been conducted. The studies are arranged by topical organization of content, and refer especially
to the behavior, habits, usages, experiences and
needs of research scientists in obtaining scientific
information. Volume II is a collection of tables
which complements the information presented in
Volume 1.
COMPUTER APPLICATIONS
SERVICE.
Detroit,
American Data Processing, Inc., 1962. 2 v.
Cordonnier, G., OPTIMATION DE l'ORGANISATION DOCUMENTAIRE [Optimization of documentation]. In Kent, A., ed. Information retrieval and
machine translation, pt. 2, p. 1205-1237. New
York, Interscience, 1961. Also in Rev. Doc. 27(1):
12-31. Feb. 1960.
This is a state of the art review of the various
techniques in use for machine translation and
information retrieval. Particular attention is given
to systems which involve photostorage and the
"Selecto" system.
Costello, J. c., SOME SOLUTIONS TO OPERATIONAL PROBLEMS IN CONCEPT COORDINATION.
Amer. Doc. 12(3):191-197. July 1961.
Discusses the operational problems attendant to
the functioning of coordinate indexes with emphasis on the intellectual and human considerations, those which determine system efficiency and
output quality.
Coughlin, R. E., DOCUMENTATION FOR SCIENCE
AND ENGINEERING. Burlington, Conn., Burlington
Pub. Co., 1960. 177 p.
A review prompted by what the author regards as
a major shortcoming of many of today's technical
communications programs, namely the seemingly
unchecked trend toward specialization at the expense of integration. It is contended that implementation of documentation programs for science
and technology as a whole will offset the detrimental effect of over-diversification of personnel
and facilities. Toward this end, details are given
on practical matters of establishing a documenta-
139
tion program, including organization, administration, standardization, and education. ©ADI
DATA PROCESSING ANNUAL. 3. PUNCHED CARD
AND COMPUTER APPLICATIONS AND REFERENCE
GUIDE., 1961. Detroit, Mich., American Data Processing, Inc., 1961. 320 p.
Reviews of books in the data processing field, p.
210-220; Books in the data processing field, p.
221-226; Bibliographical index to articles in
periodicals, p. 227-274.
DATA PROCESSING EQUIPMENT ENCYCLOPEDIA.
Detroit, American Data Processing, Inc., 1961. 2 v.
Dyson, G. M.,and Farradane, J., THE AIMS OF THE
INSTITUTE OF INFORMATION SCIENTISTS, LTD.
n. p., 1961. 6 p. Prepared for the General Papers
Session of the Division of Chemical literature,
American Chemical Society, Sept. 5, 1961.
Vol. 1, Electromechanical devices: punched card,
punched tape, related systems. Vol. 2, Electronic
devices.
The authors recite the background which leads up
to the formation of the Institute and then explain
its purposes and plans for the future.
DATA PROCESSING LIBRARY SERIES. Detroit, Mich.,
American Data Processing, Inc., 1962. 5 v.
Edmundson, H. P., and Wyllys, R. E., AUTOMATIC
ABSTRACTING AND INDEXING - SURVEY AND
RECOMMENDATIONS. Commun. ACM 4(5):226234. May 1961.
DATA PROCESSING YEARBOOK. Detroit, Mich.,
American Data Processing, Inc., 1962.
Dennis, B. K., HIGH SPEED LITERATURE SEARCHING ON AN IBM 704. 25 p. In Information Retrieval Systems Conference, Poughkeepsie, 1959.
General information manual [papers]. New York,
Internat!' Business Mach. Corp., 1959.
A detailed description of General Electric's Flight
Propulsion Division's Technical Information Center's high speed literature searching system which
uses the IBM 704 computer. Also describes the
Uniterm coordinate indexing system which was
previously used and explains why the Technical
Center decided to mechanize its searching system
and why the 704 was selected for use in the new
system.
Dubeste,r, H. J., THE LIBRARY OF CONGRESS
LOOKS AT MECHANIZATION OF INFORMATION
RETRIEVAL. In Clarification, Unification & Integration of Information Storage & Retrieval; proceedings, Feb. 23rd, 1961 symposium, p. 37-45. New
York,1961.
Mechanization is used in several operations at the
Library of Congress, such as those of the Order
Division and the Loan Division. However, the
management of the Library of Congress has many
questions which they feel must be answered before
all or parts of their information system are
automated. The first of these questions is: "Is
automation needed?"
Duyvis, F. D., DOCUMENT REPRODUCTION SERVICES; THEIR EFFICIENT ORGANIZATION AND
MANAGEMENT. Unesco Bul. libr. 14 (6): 241-259.
Nov. 1960.
Twenty-five principles are enunciated and explained. They cover such matters as scheduling of
work, laboratory maintenance, transport of docu-
140
ments to the camera, communication and control
of orders, pricing, accounting, and standardization.
The author recommends a series of case reports
to provide comparative financial data, and the
establishment of an international organiZation
for reproduction services possibly under the auspices of IFLA and FID. A bibliography of more
than 30 references is appended. ©ADI
This report presents a new concept in the automatic analysis of documents: the relativefrequency approach to measuring the significance
of words, word groups, and sentences, which employs as the primary criterion of a word's significance the contrast between the word's relative
frequency of use with the document and its relative frequency in general use. This approach is
discussed in detail, as are its applications to problems of automatic abstracting, indexing, and
translation. The report includes a summary and
comparison of the pioneering automatic analysis
studies of P. B. Baxendale, H. P. Luhn, and V. A.
Oswald, Jr. ©ADI
Engelbart, D. c., SPECIAL CONSIDERATIONS OF
THE INDIVIDUAL AS A USER, GENERATOR, AND
RETRIEVER OF INFORMATION. Amer. Doc. 12 (2):
121-125. Apr. 1961.
Urges the development of automatic systems for
individuals which would be a way to store, retrieve, and manipulate the information within
each individual's private domain, with informationpacket sizes that match his actual needs ( i.e.,
separate concepts, facts, considerations, etc.). This
private automatic system could go far toward
increasing the effec.tiveness of the individual's
capabilities to the level needed for the extended
and complex problems that are the pressing ones
of our day. The author also describes such a
system which he developed and named "microdocumentation" .
European Productivity Agency, TECHNICAL INFORMATION AND THE SMALLER FIRM; FACTS AND
FIGURES ON PRACTICES IN EUROPEAN AND
AMERICAN INDUSTRY. Paris, 1958. 69 p. (Project
296/2).
This is the final report of the survey of the
methods used by small and medium sized manufacturing firms to obtain technical information.
European Productivity Agency, Division for Tech·
nical Action and Productivity in Agriculture and
Food, AGRICULTURAL DOCUMENTATION: THE
RESPONSIBILITIES OF LIBRARIES AND INFORMATION SERVICES. Paris, 1961 . 77 p.
Report of a seminar, Apr. 20-23, 1960, arranged
by the European Productivity Agency, and organized by the Ministry of Food, Agriculture and
Forestry and the Research Council of Food, Agriculture and Forestry of Germany and the International Association of Agricultural Librarians
and Documentalists.
Fairbairn, R. E., INFORMATION METHODS:
ADAPTATION TO GROWTH. Aslib Proc. 9 (10):
301·314. Oct. 1957.
On the premise that information services cannot
be planned ahead, but must be developed in close
connection with the experience of services requested and rendered, the writer discusses the
stages of development in information and related
services; the differences between commercial and
economic, and technical information and intelligence work; the management of an information
service, and future expectations and realities. In
the last, he cautions against the enthusiasm generated by the advent of microcopying techniques
and computer systems as cure-ails for information
problems. Also outlined is a suggested role for
ASLIB in furthering information services. ©ADI
Farradane, J., THE FUTURE OF INFORMATION
WORK. Aslib Proc. 12(5): 191·199. May 1960.
The information officer, who is deputizing for the
researcher in his reading, must necessarily be an
information scientist. Few of his essential tasks
have affinities with librarianship and newer developments in information work are still less related.
Although the work may include in addition to the
more usual activities of abstracting, indexing,
keeping research workers informed on literature
. . . and research in documentation, the author
believes that the general standards are deplorably
low, and that this is largely because managements
and governments fail to appreciate their importance . . . Many problems can be solved only
through collaboration and international cooperation. The hope for the future lies in raising standards, improving techniques, and thus attracting
higher grade personnel, and in stimulating government and industry to greater awareness of the
potentialities of information services. ©ADI
Foskett, D. J., INFORMATION SERVICE IN LIBRARIES. London, C. Lockwood, 1958. 142 p. Ref.
(Crosby Lockwood's new librarianship series, 4).
Contains chapters treating origins of information
service in libraries, role of the information officer,
selection and acquisition of stock, arrangement
and indexing, dissemination of information, refer-
ence service, educational role of an information
service, reports and correspondence registries,
training and qualification for information work,
and further developments in information service
including translation, documentary reproduction
and microfilm. ©ADI
Francillon, M., INFORMATION RETRIEVAL: A VIEW
FROM THE REFERENCE DESK. J. Doc. 15(4): 187198. Ref. Dec. 1959.
A view of the problems in information retrieval
from the standpoint of the librarian as liaison
between the researcher and information. The
author believes that attempts to improve information retrieval must take into account this relationship. He emphasizes the idea by describing and
presenting results of a study conducted from
records of reference service in the IBM Research
Library at San Jose, California.
Frank, 0., CO-OPERATION BETWEEN FIRMS IN
THE FIELD OF INFORMATION. Rev. Doc. 28 (3):
89-92. Aug. 1961.
Report on cooperation between firms in the field
of information based on the author's practical
experiences in the coordination of documentation
and information work. The following points are
discussed: Ways of disseminating information;
necessity of good planning; inquiry into the
amount, availability and presentation of information in a particular field; organization questions;
compilation of the information; distribution of
the information; cooperation with other information centers; savings through cooperation; and
experience with cooperation in information.
Funk, C. E., THE INFORMATION CENTER OF
AMERICAN CYANAMID'S STAMFORD LABORATORIES. J. Chem. Educ. 34 (10):507-509. Oct. 1957.
Describes the organization and operations of the
Technical Information Section at Cyanamid's research laboratories, including the molecular formula index for company technical reports and
screening programs. ©ADI
Garrett, P., A CLASSIFICATION SYSTEM FOR ANY
DATA BANKING (INFORMATION STORAGE AND
RETRIEVAL) PROCESS. Santa Monica, Calif., Benson-Lehner Corp., 1959, 11 p. Research report
59-6 on Contract Nonr-266600 with the Office of
Naval Research.
Study was started to discover and state explicitly
the fundamentals of data banking (more commonly called information storage and retrieval).
A clear exposition of these fundamentals will be
useful in directing the development of equipment
to meet particular application requirements. A
logical framework or hierarchical tree is displayed
that includes all possible data banking processes
and shows their similarities and differences. The
basis of this framework is the organization of the
141
store and the method of search, not the physical
form of either the items of information or the
search equipment. Author.
Griffin, M., THE LIBRARY OF TOMORROW. Soc.
Tech. Writers & Pub. Proc. Annu. Conv. 8:66-71.
Ref. 1961.
Glass, H. B.,and Norwood, S. H., HOW SCIENTISTS
ACTUALLY LEARN OF WORK IMPORTANT TO
THEM. Internat!' Conf. Sci. Inform. Proc. 1958(1):
195·197. 1959.
A projection of what the library of the future will
be. To meet the increasing volume of scientific and
technical information, the libraries will use machines for cataloging, indexing, abstracting, crossreferencing, and retrieving of information. The
library will be part of 'a global network of communication centers. Tht technical writer will have
the added task of adopting input material for
machine scanning. Concludes with a plea that
librarians must not be unprepared to load the
machines or to retrieve from them when they are
available for the efficient operation of the libraries
of the future.
Report of a survey in which 50 biological scientists
were interviewed to determine how and where
they had learned of sources of information pertinent to their fields of interest.
Glimn, A. F.,and Greenway, R. D., INFORMATION
STORAGE AND RETRIEVAL. Part II, Equipment
applications. Bethesda, Md., Defense Systems
Dept., Gen. Elect. Co., 1960. lOp. Ref.
Machines alone can never solve the fundamental
problems of information storage and retrieval.
They can only provide the means for implementing
solutions which would not have been practical
using manually operated systems. Thus, the solution to the problem of evaluating and specifying
equipment for information storage and retrieval
systems must be evolved by conducting a complete
and exacting operations analysis study to determine the system parameters. This paper discusses
the factors involved in evaluating the specifying
equipment and equipment configurations for
information storage and retrieval systems.
Goldman, A. J., and others, MATHEMATICAL
RESEARCH RELATED TO INFORMATION SELECTION
SYSTEMS. Washington, U. S. Nat!. Bur. Standards,
1961. 58 p. (NBS Report No. 6883).
Mathematical research relevant to information
selection systems is described. Three areas of
mathematical problems related to mechanization
of the information retrieval process are consid~
ered: (1) efficient formulation of search questions; (2) rapid classification of incoming documents on a given one-dimensional scale, and (3)
categorizing and manipulating documents and
their relationships.
Greenway, R. D., and Russell, M. V., LET'S DESIGN
AN INFORMATION SYSTEM. Washington, D. C.
Gen. Elect. Co., Defense Systems Dept., Inform.
Systems Sect., 1960. 11 p. Ref. Presented as a
contributed paper to the 18th National meeting of
the Operations Research Society of America,
Detroit, Mich., Oct. 10-12, 1960.
Discusses the complex inter-relationships which
exist between techniques for indexing and devices
for machine search and retrieval of indexed information. A generalized approach to indexing is used
to show the equipment requirements as a function
of system requirements. The future of mechanized
information systems is discussed. It is imperative that every information system provide for
collection, reduction, storage and retrieval,
manipulation and presentation of information.
142
Hanson, C. W., ORGANIZATION OF MATERIAL
AND SERVICES IN A SMALL INFORMATION DEPARTMENT. Rev. Doc. 28(3):105-109. Aug. 1961.
A discussion, from actual practice, of precisely
what can be done by a small industrial firm to
provide itself with an effective information service; how it can be done; and how much it will cost.
The basic aims of this information service are:
(1) to keep the staff informed of scientific and
technical developments relevant to their work, and
(2) to help them find solutions to specific problems
as they arise.
Hattery, L. H., INFORMATION AND COMMUNICATION IN BIOLOGICAL SCIENCE. Washington,
D.C., The Amer. Univ., School of Govt. & Pub.
Admin., Cent. Techno!. & Admin., 1961. 99 p.
Report of the Biological Sciences Communication
Seminar, sponsored by the Biological Sciences
Communication Project, AIBS, and held at The
American University, June 19 to July 1, 1961. The
purpose was to identify and review problems associated with the flow of biological science information from scientist producer to scientist consumer.
Developments in the various areas of study are
given in this report, along with the seminar program and procedure, a roster of participants, and
a list of committees with their purposes and
reports.
Heald, J. H., TRANSITION FROM A MANUAL TO
A MACHINE INDEXING SYSTEM. Inst. Inform.
Storage & Retrieval, 3rd, 1961. Papers presented,
p. 170-190. 1962.
An account of the problems and requirements involved in the conversion of ASTIA's major information system from one of manual operation to
one of automation. Included is a detailed description of the Thesaurus arrangement developed
especially for this transition.
Herner, S., THE RELATIONSHIP OF INFORMATION
- USE STUDIES AND THE DESIGN OF INFORMATION STORAGE AND RETRIEVAL SYSTEMS. Washington, D.C., Herner & Co., 1958. 30 p. Technical
Note No.1 on Contract AF 30(602)1857.
A review is presented of the results of past studies
of the information-gathering methods of workers
in various fields, and the general applicability of
these results to the design and improvement of
information programs and systems is shown. Two
cases from the literature are used to illustrate
storage and retrieval systems that do and do not
meet the requirements of their users, and to show
the contrasting need of the pure scientist for mere
references to information and of the applied
scientist for direct access to actual information.
Author.
Heumann, K. F., THE BIG BLACK BOX AT YOUR
BECK AND CALL. Spec. Libr. 51 (9):483-484. Nov.
1960.
Suggestions are given as to how a librarian should
approach and utilize the capabilities of data processing systems. Manufacturers' representatives
and programmers for the systems are recommended as people who can be of much assistance.
The experience of a reference librarian, or of one
skilled in indexing, classification, or subject
headings, is said to be of much value in utilizing machine techniques for searching technical
literature. ©ADI
Hillier, J., MEASURING THE VALUE OF INFORMATION SERVICES. J. Chem. Doc. 2(1):31-34. Jan.
1962.
A discussion of the value of an information service
from the point of view of management. The author
contends that: (1) Management's considerations
must go beyond the provision of extra brains to
perform functions which have grown beyond the
capability of the individual, (2) the additional
brain-power which is serving the creative individual must, in some way, operate in conjunction
with the individual's experience-based optimization of his information input, (3) there must be
a balance between a centralized information service and a dispersed system, where the service is
performed by individuals who are an integral part
of the research teams, and (4) we must undertake
further studies to develop a more complete understanding of the relationship between information
flow and creativity.
Howerton, P. W., DOCUMENT SYSTEMS: TO AUTOMATE OR NOT TO AUTOMATE. Washington, D.C.,
1961. 12 p. Paper presented before the Interagency Records Administration Conference, National Archives and Records Service, Mar. 17,
1961.
Documentation problems started with the first
recording of knowledge. Some of the solutions to
these problems have been less than satisfactory.
A satisfactory documentation system need not
depend on complex devices to be useful, but it must
satisfy the need of the customers who will use its
services. In such a system, machines are powerful
tools which, under the guidance of human judgment, can do routine operations that free the
creative talents of man.
International Business Machines Corporation,
REFERENCE MANUAL: INDEX ORGANIZATION
FOR INFORMATION RETRIEVAL. White Plains,
N.Y., 1961. 63 p.
A simplified presentation of the methods and patterns of organizing indexes. It is geared mainly to
systems personnel in an effort to help them understand the problems to be encountered in organizing information for retrieval. A glossary of terms
and a bibliography on information indexing are
included.
INTERNATIONAL CONFERENCE ON SCIENTIFIC
INFORMATION. Washington, D. c., Nov. 16-21,
1958. Proceedings. Washington, D.C., Nat!. Acad.
Sci. - Nat!. Res. Counc., 1959. 2 v.
This conference was sponsored by the National
Science Foundation, the National Academy of
Sciences - National Research Council, and the
American Documentation Institute. The discussions covered seven different areas: 1. Literature
an::! reference needs of scientists: knowledge now
available and methods of ascertaining requirements; 2. The function and effectiveness of
abstracting and indexing services; 3. Effectiveness of monographs, compendia, and specialized
centers: present trends and new and proposed
techniques and types of services; 4. Organization
of information for storage and search: comparative characteristics of existing systems; 5. Organization of information for storage and retrospective search: intellectual problems and
equipment considerations in the design of new
systems; 6 .. Organization of information for storage and retrospective search: possibility for a
general theory; 7. Responsibilities of a government, professional societies, universities, and industry for improved information services and
research.
International
Federation for
Documentation,
MODERN DOCUMENTATION AND INFORMATION
PRACTICES; A BASIC MANUAL edited by Dr. O.
Frank. The Hague, 1961. 225 p. Ref. (FlO Publication 334).
This manual, published with the assistance of
Unesco, "deals with the fundamentals of documentation and information work, and especially
with the organization of documentation and information centres". Contents: Chapter 1. Introduction, by O. Frank; 2. Organizational forms and
purposes, by O. Frank; 3. The need to organize,
by J. W. Holmstrom; 4. Intake and outflow, by
G. S. Stekhoven; 5. References and storage of
documents, by H. Elsner; 6. Ready made references, by J. E. Holmstrom; 7. Information retrieval, by J. E. Holmstrom; 8. Mechanization using
manual apparatus, by H. A. Elsner; 9. Mechaniza-
143
tion using electrical apparatus; by J. E. Holmstrom; 10. Reproduction, by O. Frank; 11. Internal
organization, by J. E. Holmstrom; 12. Overcoming
language barriers, by J. E. Holmstrom; 13. Human
factors, by J. Koblitz; 14. Co-operation and coordination, by J. Koblitz; Appendix 1. Recommended books and articles; Appendix 2. Organizations concerned in the promotion and improvement
of documentation and related practices.
Jeanmaire, R., ROLE D'UN SERVICE DE DOCUMENTATION DANS UNE GRANDE ENTREPRISE INTERNATIONALE [Role of a documentation service in
a large organization]. Nachr. Nouvelles 35(5):
153-158. Sept.jOct. 1959.
A documentation department can make an important contribution to the creative, and advisory
activities of an organization: the documentalist
can make of his department a laboratory of ideas,
suggest methods of analysis or reveal descriptions
of processes in furthering experimental work,
assist commercial development by surveys of possible markets, the setting of high standards in
publication, contribute to good public relations,
etc. ©ADI
Kent, A., EXPLOITATION OF RECORDEDINFORMATION. I. DEVELOPMENT OF AN OPERATIONAL
MACHINE SEARCHING SERVICE FOR THE LITERATURE OF METALLURGY AND ALLIED SUBJECTS.
Amer. Doc. 11 (2): 173-188. Apr. 1960.
Describes the background, plans, initial stages.
and problems of the operational mechanized
searching service in metallurgy and allied subject
fields. The article is based on the experiences of
the Center for Documentation and Communications Research at Western Reserve University
which is supported by the American Society for
Metals and the National Science Foundation.
Kessler, M. M., AN EXPERIMENTAL COMMUNICATION CENTER FOR SCIENTIFIC AND TECHNICAL
INFORMATION. lexington, Mass., Mass. Inst.
Tech., lincoln lab., 1960. 19 p. (Report No. 4G0002). Contract AF 19(604)5200.
A discussion of the problems of scientific communication and some of the thinking and research
recently performed on this subject at Lincoln
Laboratory. The report stresses that the evaluation
of new ideas and components must be made in a
system environment and not in terms of parameters unique to each component. A distinction is
made between scientific message units and their
mode of propagation. The message units (scientific
talks and papers) are considered adequate for their
functions, but they are encountering increasing
losses and delays in propagation. The analysis
indicates that the most critical item in need of
research and invention is an indexing or addressing scheme that will direct papers to readers.
Valid directional indexing should be sought in the
operational history of the author and the intended
144
reader. A communication system for experimental
purposes is briefly outlined. The three principal
channels of the systems are (1) Non-interrogated
broadcasting, (2) Interrogated point to point
transmissions, and (3) Non-interrogated pOInt to
point transmission. It is stressed that the critical
need at this point is for a deeper understanding
of the logic of scientific information and its flow
habits. It is recommended that a pilot system of
scientific communication be established to serve
as a test bed for components development and
systems research. Author.
loosjes, T. P., ORGANISATION UND DIENSTE DER
HOLLANDDISCHEN DOKUMENTATIONSZENTRALE
FUR DIE LANDWIRTSCHAFT [Organization and
services of the Netherlands Documentation Centre
for Agriculture]. Nachr. Dok. 10(2):59-65. July
1959.
A description of the Netherlands Documentation
Centre for Agriculture, as an example of centralization of documentation effort in a subject field.
Activities and services of the Centre, and some
details of operational costs are included in the
article.
lowry, W. K., AUTOMATIC EQUIPMENT FOR INFORMATION HANDLING. Automatic Documentation in Action. Proc. 1959:51-57. 1961.
A description of the application of 9 pieces of
electronic equipment to information handling
functions at the Bell Telephone Laboratories. The
philosophy behind these applications is that the
greatest advantages will be realized from using
machines for clerical operations in the information service. The author concludes that although
they have barely scratched the service in applying
machine techniques, they have already realized a
great reduction in space requirements and are able
to offer much better service on engineering
drawings.
lowry, W. K., and Albrecht, J. c., A PROPOSED
INFORMATION HANDLING SYSTEM FOR A LARGE
RESEARCH ORGANIZATION. Internat!' Conf. Sci.
Inform. Proc. 1958(2):1181-1202. 1959.
The authors have designed this information center
with a large research organization in mind, but
most of the approaches and procedures may also
be applied to smaller-groups. It is advocated that
this (or any other) system be centered around the
information requirements of the group involved,
in order that collection and maintenance of superfluous materials be avoided.
luhn, H. P., SELECTIVE DISSEMINATION OF NEW
SCIENTIFIC INFORMATION WITH THE AID OF
ELECTRONIC PROCESSING EQUIPMENT. Amer.
Doc. 12(2):131-138. Ref. Apr. 1961.
Describes a service system of dissemination of
new scientific information in which machines are
used.
MACHINE INDEXING: PROGRESS AND PROBLEMS.
Papers presented at the Third Institute on Information Storage and Retrieval, February 13-17,
1961. Washington, D.C., The Amer. Univ., School
of Govt. & Pub. Admin., Center Technol. & Admin.,
1962. 354 p.
The purpose of the Institute was to explore one
attack on the problem of communication of scientific and technical information - that of machine
indexing. The twenty-one papers presented are
intended to provide a composite report of the
state of the art in 1961.
McCormick, E. M., DIGITAL COMPUTER PRIMER.
New York, McGraw-Hili, 1959.
McCormick, E. M., THE RESEARCHER AND INFORMATION STORAGE AND RETRIEVAL. In Clarification, Unification & Integration of Information
Storage ond Retrieval; proceedings Feb. 23rd,
1961 symposium, p. 79-83. New York, 1961.
Discusses three general points pertaining to the
subject which" can be summarized as a consideration for more systems analysis in information
storage and retrieval". Points are: (1) existing
studies for systems evaluation, (2) association of
ideas, and (3) "the amount and type of information interchange there is or will be between
information storage and retrieval systems".
McCormick, E. M., WHY COMPUTERS? Inst. Inform.
Storage & Retrieval, 3rd, 1961. Papers presented,
p. 220-232. 1962.
A discussion of computers based on the assumption that they are, could, or should be considered
the ultimate in information processing devices.
The author cites reasons for their use, and
discusses their advantages, limitations, and
application.
McGuire, H. G., and Stoddard, T. L., INFORMATION PRESENTATION SYSTEM. Bethesda, Md.,
Gen. Elect. Co., Defense Systems Dept., 1960. 14 p.
Machine-controlled information presentation subsystems used for displaying information at electronic speeds are the subject of this paper. The
over-all effectiveness of the display subsystem
depends on the net effectiveness of its three complimentary elements: Men, Machines, and Methods.
These elements are interdependent, and their
arrangement in a proper relationship involves
balancing the capabilities and limitations of each
element in order to achieve the objectives of the
system.
Maizell, R. E., INFORMATION GATHERING PATTERNS AND CREATIVITY; A STUDY OF RESEARCH
CHEMISTS IN AN INDUSTRIAL RESEARCH LABORATORY. Amer. Doc. 11 (1):9-17. Ref. Jan. 1960.
A comparison of creative and "noncreative" research chemists with respect to the ways in which
they use their professional and technical literature.
The creative chemists differ from the "noncreative" in that the former read more technical
literature on the job, are less reluctant to use
literature of greater reading difficulty, are less
influenced in their independence of thought, read
more extensively and consult more frequently the
older material, are more inquisitive and have
broader cultural interests. The findings of the
study are believed to be helpful in planning library
and information services, in refining future inquiries into the ways in which scientists use
recorded information, and in improving tests
for the identification of creative ability among
chemists. Author.
Melik-Shakhnazarov, A. S., TECHNICAL INFORMATION IN THE U.S.S.R. Translated from the
Russian by B. I. Gorokhoff. Mass. Inst. Tech. Libr.
Libr. Monog. 3, 122 p. Ref. 1961.
Under a grant from the National Science Foundation, the Libraries of the Massachusetts Institute
of Technology are sponsoring a series of studies
by Boris 1. Gorokhoff of the dissemination of
scientific and technical information in the Soviet
Union. This project is part of a foreign science
information program of the Foundation's Office
of Science Information Services, which include
critical and descriptive studies of the organizations and activities of information services
abroad . . . It has therefore been decided to incorporate in the Library Monograph Series of
M.LT. this translation of A. S. Melik-Shakhnazarov's N auchnotekhnicheskaia informatsiia i propaganda v mashinostroenii (Technological information in machine building), published in Moscow
in June, 1960. Notwithstanding its title, the book
is actually the most comprehensive review of the
Soviet techn,ical information system to appear in
the U.S.S.R. in recent years. Foreword, page v.
Mooers, C. N., THE "TAPE TYPEWRITER PLAN", A
METHOD FOR COOPERATION IN DOCUMENTATION. Cambridge, Mass., Zator Co., 1960. 22 p.
(AFOSR-TN-60-532). Contract AF 49(638}-376. Also
in Aslib Proc. 12(8}:277-291. Aug. 1960.
The "Tape Typewriter Plan" makes possible widespread cooperation in documentation between
libraries, while at the same time facilitating the
internal clerical and cataloging operations at
individual libraries.
Mueller, M. W., TIME, COST AND VALUE FACTORS IN INFORMATION RETRIEVAL. 12 p. In Information Retrieval Systems Conference, Poughkeepsie, 1959. General information manual
[papers]. New York, Internatl. Business Mach.
Corp., 1959.
This paper deals with a study conducted by the
Operations Research Division of the Lockheed
Aircraft Corporation (California Division) with
the specific intent of unearthing some of the eco-
145
nomic factors behind the library information
handling system in a typical engineering organization. While many of the findings are still of a
preliminary nature, they show the need for further development of the commodity concept and
provide some interesting design criteria for improved information systems. Introduction.
Murdoch, J. W., and Simpson, G. B., PRACTICAL
INFORMATION HANDLING. Indus. Res., Apr./May
1961: 41-43.
An optimum information system must provide for
the following: the user must guide the selection
of inputs; information processing time must be
held to a minimum; information, and not references to information, must be immediately available; the user will not be able to state his
information requirements exactly; and the user
must be able to proceed easily from one segment
of the system to another.
Myatt, DeW., and Upham, T. E., A QUANTITATIVE
TECHNIQUE FOR DESIGNING THE TECHNICAL
INFORMATION CENTER. J. Chem. Doc. 1(3}:18-24.
Nov. 1961.
Based on their definition of a technical information center, and their assertions as to what the
mission of an information center should be and
the relation between the resources which the center must collect and the people served by the
center, the authors present 4 guide-lines for designing and operating an information center. In
conclusion, they urge that anyone designing an
information center begin by looking at the
situation with the greatest perceptivity.
Opler, A., and Baird, N., EXPERIENCE IN DEVELOPING INFORMATION RETRIEVAL SYSTEMS ON
LARGE ELECTRONIC COMPUTERS. Internat!' Conf.
Sci. Inform. Proc. 1958(1):699-710. 1959.
Report of research into the use of large electronic
computers for information retrieval systems. The
research was conducted by programming an IBM
702 Electronic Data Processing Machine for the
search of two chemical files. Results are reported
from the viewpoint of all those concerned with
retrieval systems - the user, the documentalist,
and the administrator.
Samain, J., CLASSIFICATION AND INFORMATION
RETRIEVAL. Automatic Documentation in Action.
Proc. 1959:273-282. 1961.
First, this is a discussion of the problem that the
information stored in the system is very often
expressed in very different terms from those used
in the question asked of the system. Second, the
consequences of this problem which confronts
every documentation system are drawn. Third,
this describes the organizational scheme based on
this problem and which has been applied in the
National Center of Scientific Research CC.N.R.S.)
of Paris.
146
Schultz, C. K., RESEARCH IN INFORMATION RETRIEVAL - SOME RECENT CONTRIBUTIONS. In
Clarification, Unification & Integration of Information Storage and Retrieval; proceedings, Feb.
23rd, 1961 symposium, p. 47-55. Ref. New York,
1961.
An informal survey of research in information
retrieval as found mainly in published reports.
Sewell, R., TECHNICAL INFORMATION SERVICES
IN GERMANY, LUXEMBOURG, BELGIUM, AND
HOLLAND. Aslib Proc. 11 (1 0}:240-265. Oct. 1959.
A report of a three weeks' trip to 22 technical information centers on the continent of Europe. The
three aims of the trip were: 1) to examine methods
of operating information services, 2) to study the
attitude of industrial management and research
toward technical information, and 3) to contact
some of the leading people in the techQical information field in Western Europe. Each center
visited is fully described, and its available
information and library services are given.
Shaw, R. R., A MEDICAL INTELLIGENCE PROGRAM
FOR THE NATIONAL INSTITUTES OF HEALTH. n.p.,
1961. 117 p.
Report of a survey of library and bibliographic
services at the National Institutes of Health with
recommendations for optimum services ... Chapter X, Bibliographic Intelligence Services, is of
particular interest. In it is outlined a section
offering services tailored to the needs of particular
research projects together with a screening service
and special indexes for techniques, instrumentation, and news of medical research. ©ADI
Singer, T. E. R., ed., INFORMATION AND COMMUNICATION PRACTICE IN INDUSTRY. New York,
Reinhold, 1958. 304 p. Ref.
A compilation of chapters presented as guides to
subject fields such as operations research and the
technical information program, chemical research
file departments as information services, patent
searching, abstracting, indexing, training the
literature scientist, etc.
Spitzer, E. F.,and McKenna, F. E., THE INDUSTRIAL
INFORMATION DEPARTMENT. In Singer, T. E. R.,
ed., Information and communication practice in
industry. p. 1-53. Ref. New York, Reinhold, 1958.
A general description of the organization and
operation of industrial information departments.
The description is rather short, but the reader is
guided to more detailed information on the subject
through the bibliography provided with the
article.
U. S. Armed Services Technical Information
Agency, CONTROLLING LITERATURE BY AUTOMATION. Presented at the IV. Annual Military
Librarian's Workshop sponsored by Armed Serv-
ices Technical Information Agency . . . 5-7 October 1960. Washington, 1960. 130 p.
Contents: Preparing for automation, by Lt. Col.
W. Hammond; Selection, training and relocation
of personnel, by L. R. Barnes; File conversion,
data cleanup and inventory control, by F. A. Keller; Human aspects of ADPS, by E. S. Pope;
Building an information retrieval system, by H.
Rehbock; Impact of automation on the organizational structure of ASTIA, by H. Miles; The road
ahead _. integrated data processing, by W. A.
Barden; The tape typewriter plan, by C. Mooers;
and, A generalized computer method for information retrieval, by C. Schulz.
U. S. Congress, Senate, Committee on Government Operations, CO-ORDINATION OF INFORMATION ON CURRENT FEDERAL RESEARCH AND
DEVELOPMENT PROJECTS IN THE FIELD OF ELECTRONICS. An analysis of agency systems for
storage and retrieval of data on ongoing work
and of views of private companies on indexing
and communication problems. Washington, Govt.
Print. Off., 1961. 292 p. Ref. (87th Congress. 1 st
Session. Committee Print).
This publication is a companion print to Senate
Report No. 263, 87th Congress, 1st Session, issued
May 18, 1961. The present print narrows the focus
of the preceding publication. It explores in the one
field of electronics the type of problem which the
earlier print surveyed for all scientific fields namely, how to manage information on research
and development work in its prepublication stage.
- from Senator Humphrey's Letter of Transmittal. In addition to the main body of the report
this publication contains the statement and findings of Senator Humphrey, chairman of the Sub-
committee on Reorganization and International
Organizations, and the Summary and Recommendations of the Consultant, J. Stern.
U. S. National Science Foundation, Office of
Science Information Service, CURRENT RESEARCH
AND DEVELOPMENT IN SCIENTIFIC DOCUMENTATION, No.9. Washington, Govt. Print. Off.,
1961. 270 p. Ref. (NSF-61-76).
This is the 9th in a series of semi-annual descriptive reports on current research and development
in scientific documentation.
U. S. National Science Foundation, Office of
Science Information Service, NONCONVENTIONAl TECHNICAL INFORMATION SYSTEMS IN
CURRENT USE, No.2. Washington, Govt. Print.
Off., 1959. 66 p. (NSF-59-49). The information in
this report supersedes, for the most part, the information given in publication No.1 of the same
title. Supplement to No. 2 (NSF-60-14) issued in
1960, brings the information up to date.
Describes technical information systems currently
in operation which embody new principles for the
organization of subject matter or employ automatic equipment for storage and search.
U. S. National Science Foundation, Office of
Science Information Service, SPECIALIZED
SCIENCE INFORMATION SERVICES IN THE UNITED
STATES; A DIRECTORY. Washington, 196]. 528 p.
(NSF-61-68).
A directory to 427 physical and biological sciences
information services in the United States as compiled from information collected by questionnaires.
Designed as a reference aid for working scientists,
engineers, librarians, and document lists.
147
IN DE X
Abstracting/ automatic, 98
Center/information/definition, 132
Abstracting/by the author, 28
Center /information/designing, 45-53
Abstracting/objectives, 16-17
Center /information/ external aspects, 32-33
Activities of library of Congress/ technical
information cenfer, 35
Center/information/internal aspects, 31-32
Adkinson, Burton, 121
Centers/inputs, 44
Aerospace Materials Information
Project (AMI), 62-64
American Documentation Institute, 77
American Institute of Biological Sciences, 14, 81
American Petroleum Institute, 43
Centers/technical information, 31-33, 34-36
Change/organizational capability, 102-103
Characteristics/information retrieval systems, 94
Charts/logical flow, 105
American Psychological Association, 81
Checklist for the Organization, Operation and
Evaluation of a Company library, A , 104
American University, 14, 105
Classification of sciences, 126
Anderson, Hattie T., Ch. 14, 111-116
Cleverdon, Cyril W., 118
Annual reviews/literature, 18
Columbia University, 79
ARIES project/case history, 50-52
Arthur Andersen and Company, 117
Communication Between Computer and User in
Information Searching, 83-91
Articles in international journals, 120
Communication/control, 124
Asbury, W.
c.,
Ch. 6, 37-44
ASTIA, 28, 104, 107, 114
Author abstracting and indexing, 28
AutoCom/ a dissemination tool of Automatic
Information Retrieval System (AIRS), 66-67
Automated Intelligence Systems, 92-100
Automatic abstracting, 98
Automatic indexing, 95-97
Automatic Information Retrieval System (AIRS) at
General Electric Company: Flight Propulsion
Division, 61 ff.
Automation expense trend/FPD Technical
Information Center, 72-73
Bacon, Charles R. T., 86
Baker, W.O., 102
Battelle Memorial Institute, 25
Bell Telephone laboratories, 102
Bernais, Peter, 120
Bernal, J. D., 80
Biological Sciences Communication Project, 14
Bourne, Charles, 104, 117
Brownson, Helen l., Ch. 10, 76-82
Bryant, Margaret, 81
Business intelligence system, 93
Cohn, Julius N., Ch. 3, 21-30
Case history /upper atmosphere/ARIES, 50-52
Case Institute of Technology, 79
Center/defining its function, 34
Center/financing, 61-75
Center/humanistic features, 48-49
148
Center / mission, 46-47
Communications of the Association for
Computing Machinery, 86
Compatibility of Information and Data Systems
within a Company, 111-116
Compatibility/types, 113
Computers, Communications, and Science Extending Man's Intellect, 123-130
Control/communication, 124
Control cycle/management, 131
Cooperative activities between centers, 43
Coordination/information services, 23-24
Cost/systems, 135
Costs/factors establishing system, 117-119
Costs/operating/Flight Propulsion Division,
Technical Information Center, 71-72
Costs/science information, 13-14
Council on library Resources, 86, 103
Creativity/information, 56, 59
Crisis in documentation, 15-20
Current awareness, 58
Current awareness/GE Co. Flight
Propulsion Center, 62-63
Deductive systems, 129
Definition/science information center, 132
Definition/technical information center, 46-48
Dennis, Bernard K., Ch. 9,61-75, 102, 103,
105,119
Depositories/regional, 19
Derivation/optimal service design, 50-51
Design/information project, 50-51
Design/systems, 103
Designing an Information Center to Meet
a Real-System Requirement, 45-53
Handbook on Automation and Systems
Engineering, 104
Determining documentation needs, 81-82
Hattery, Lowell H., Ch. 1, 10-14
Hayes, Robert M., 105
Dissemination versus consolidation of
information systems, 111-112
Health Law Center, University of Pittsburgh, 84
Herner, Saul, 114
Documentation crisis, 15-20
Heumann, Karl F., Ch. 16, 120-122
Documentation, Inc., 113
Hillier, James, Ch. 8, 54-60, 102
Horty, John, 84, 86
Documentation/international activities, 120-122
Documentation needs/determining, 81-82
Howerton, Paul W., Ch. 4, 31-33
Docu mentation Needs of Scientists, 76-82
Humanistic features/technical information
center, 48-49
Humphrey, Hubert H., 24
Identification/types of compatibility, 113
Documentation purposes, 15
Documents/pictorial aspects, 99
Economic Justification - Factors Establishing
System Costs, 117-119, 133
Efficiency/systems, 117
Engelbart, Douglas
c., 104
Environment/management, 131-132
Esso Research and Engineering Company, 37-44
EURATOM, 121
Evaluation/information services, 54-60
Evaluation/studies and recommendations, 103-104
Evaluation/system, 132
Evaluation/technical information center, 42-43
Expenditures/Federal/research and
development, 22-23
Factors establishing system costs, 117-119
Federal expenditures, 22-23
Federal role/technical information center, 34-35
Federation of American Societies for
Experimental Biology, 18
Financial support/system, 134
Financing/Flight Propulsion Division,
Technical Information Center, 68-72
Financing/Technical Information Center, 61-75
Indexing/automatic, 95-97
Indexing/by the author, 28
Indexing/cost factor, 117
Indexing/objectives, 16-17
Industrial organization/technical information
services, 37-44
Information center/designing, 45-53
Information center/financing, 61-75
Information center /fu nctions, 34-36
Information center/government role, 34
Information center/humanistic features, 48-49
Information center/inputs to, 44
Information center/status, 31-33
Information/ network, 21
Information resources inventory, 25
Information retrieval/GE Co. Flight Propulsion
Center, 61 ff.
Information/scientific/responsibility for, 14
Information service/evaluation, 54-60
Information service/industrial
organization, 37-44
Information service/mechanization, 23
Information/society, 124
Information systems/compatibility, 111-116
FIND-X/Manual version of AIRS, 66
Fisher, Eva Lou, 104
Flow charts, 105
Ford Foundation, 86
Formal systems, 129
Information systems/management process,
131-135
Information systems/mechanizing, 101-110
Functions of a Technical Information Center, 34-36
Funding information programs, 27
General Electric Co., 61-75
Information systems/overall system, 21-30
Information/use and practices, 78-80
Inputs/technical information center, 44
General Electric Search Comparator, 117
Gibson, R. E., 48
Intelligence systems/automated, 92-100
Inter-agency information compatibility, 27
Inter-agency information modernization, 27
Gorn, Saul, Ch. 17, 123-130
Guidelines to Mechanizing Information
Systems, 101-110
Gull, C. D., Ch. 13, 101-110
Haldane, J. B. S., 124
Interdisciplinary sciences, 130
Inter-library plan, 26
Intermediary information repositories, 13
International Activities in Documentation, 120-122
149
International aspects/information, 29
Media/communication/diagram, 12
International Conference on Scientific Information,
Washington, 1958, 77, 98
Media/transmitting information, 11
MEDLARS project, 17
International Council of Scientific Unions, 121
Menzel, Herbert, 79-80
International Federation of Documentation
(FlO), Rio de Janeiro, 1960, 120, 121
Moise, J. E., Ch. 6, 37-44
International Federation of Information
Processing Societies (IFIPS), 121
Mooers' law, 80, 112, 118, 132
International Federation of library
Associations, 121
Myatt, De Witt 0., Ch. 7, 45-53
Mueller, Max W., 118
International organization/documentation, 120-121
National Academy of Sciences - National
Research Council, 77, 122
International Organization of Patent Offices,
121,122
National referral service, 24
Inventory/information resources, 25
Network/information, 21
Newman, Simon M., Ch. 15, 117-119
Johns Hopkins University Applied Physics
laboratory, 48
Objectives/abstracting, 16-17
Kehl, William B., Ch. 11, 83-91, 117
Operating costs/Flight Propulsion Division,
Technical Information Center, 71-72
Kent, Allen, 23
Key-word-in-context (KWIC), 107
Organization/Automatic Information Retrieval
System (AIRS), 67
Knox, W. T., 29
Organization/capability for change, 102-103
language problems/documentation, 120
Organizational plan/Esso Research and
Engineering Co., 39-40
language/terminology, 10
lattice/word-pair linkages, 96
leake, Chauncey D., Ch. 2, 15-20
lee, Milton 0., 18
library scientists as reviewers, 19, 20
linkages/word-pair, 96
localized systems of communication, 12-13
logical flow charts, 105
lowry, W. K., 103
luhn, H. P., Ch. 12, 92-100, 104, 107
lyght, Charles E., 24
Machine readable texts, 98-99
Machine systems/capabilities, 108
Management control cycle, 131
Management environment, 131-132
Management Process and Science Information
Systems, 131 -135
Management/technical information, 60
Management's Evaluation of Information
Services, 54-60
Manpower/education, 125
Materials/ aerospace information, 62-64
McCormick, Edward M., Ch. 18, 131-135
Mechanization/information service, 23
Mechanization/information systems, 101-110
Mechanized Index to Chemical Patents
literature, 113
150
Mooers, Calvin, 80, 112, 118
Perry, James W., 23
Personnel/administration/system, 134
Personnel/Automatic Information Retrieval
System (AIRS), 67
Personnel/technical information center, 41
PERT (Program Evaluation and Review), 29
Physics Abstracts, 79
Pictorial Aspects of Documents, 99
Pitt Executive System for Tapes (PEST), 83
Planning and policy/Department of Defense, 27-28
Pre-design surveys/technical information
center, 49
Price, Derek, 19, 81-82
Principles/technical information center, 41-42
Processing input/information system, 13
Producer/information, 11
Publication rate/articles, 37
Public relations/centers, 135
Purposes/documentation, 15
Radio Corporation of America, 56
Regional repositories, 19
Reorientation of professional skills, 97
Repositories/regional, 19
Resources inventory/information, 25
Responsibility for science information, 14
Reviews/critical/literature, 18
Royal Society's Scientific Information Conference,
london, 1948, 76-77
Saltzberg, P. lo, 38
Technical information/management, 60
Science and Technology Division of Library
of Congress, 35
Technical Information Services in an
Industrial Organization, 37-44
Science information/responsibility for, 14
Tenth Pacific Science Congress, Hawaii, 1961, 120
Terminology and language, 10
Science information systems/management
process, 131-135
Texts/machine readable, 98
Sciences/classification, 126
Thesaurus/compatibility, 114
Sciences, interdisciplinary, 130
Scientists' documentation needs, 76-82
Training/interdisciplinary, 28
Trudeau, Arthur, 23
Search request/prepared by user, 86-91
UNESCO, 120, 121, 122
Selectivity /information, 57
United Kingdom Atomic Energy Authority, 79
Services/information project, 51
United States Air Force, 25
Shera, Jesse H., 20, 23
United States Atomic Energy Commission, 122
United States Bureau of the Budget, 103
United States Central Intelligence Agency, 32
United States Department of Agriculture, 122
United States Department of Commerce, 35
United States Department of Defense, 27, 122
United States Department of State, 121, 122
United States Library of Congress, 35-36, 122
United States National Aeronautics and Space
Administration, 22
United States National Agricultural Library, 35, 79
United States National Institutes of Health, 86, 121
United States National Library of Medicine, 17, 35
United States National Science Foundation,
23, 28, 32, 43, 79, 121
United States Office of Education, 86
United States Office of Naval Research, 49
United States Office of Technical Services (OTS), 36
United States Senate Committee on Government
Operations, 20, 26
University of Illinois, 103
University of Pennsylvania, 123
University of Pittsburgh, 83-91, 117
User/information, 11, 132
User/information/relation to centers, 31
User Oriented Search Language, 85, 86
User/prepares search request, 86
User/technical information, 54-56, 57
Users/documentation, 76-,82
Users' needs/conferences on, 76-78
Users' needs/importance of determining, 81-82
Users/responses, 119
Weinberg, Alvin, 2.2
Weiner, Norbert, 124
Western Reserve University, 104
Wiesner, Jerome, 22
Word-pair linkages, 96
Yockey, H. P., 20
Sherrod, John, Ch. 5, 34-36
Skills/ reorientation, 97
Society/information, 124
Solid Propellant Information Agency, 48
Staffing/technical information center, 41
Stanford Research Institute, 81, 117
Status of Technical Information Centers, 31-33
Studies/information use and practices, 78-80
Surveys/pre-design/technical information
center, 49
Swanson, Don, 78, 81
System/business intelligence, 93
System/ concept/information, 12
System/costs/factors, 117-119, 135
System/ design/ difficulties, 109
System Engineering, 104
System/evaluation, 132
System/financial support, 134
System/Information Systems, 21-30
System/location/organizationally, 133
System/personnel, 134
Systems/automated intelligence, 92-100
Systems/ capabilities of machine, 108
Systems, compatibility/information
and data, 111-116
Systems/ concepti documentation, 10-14
Systems/deductive, 129
Systems/ design/ competence, 103
Systems/design/mechanized, 109
Systems/efficiency, 117
Systems/formal, 129
Systems/information/management process,
131-135
Systems/mechanizing/information, 101-110
Systems/retrieval/ characteristics, 94
Systems/types of, 94-95
151
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.3 Linearized : No XMP Toolkit : Adobe XMP Core 4.2.1-c041 52.342996, 2008/05/07-21:37:19 Create Date : 2015:08:25 12:08:10-08:00 Modify Date : 2015:08:25 11:52:22-07:00 Metadata Date : 2015:08:25 11:52:22-07:00 Producer : Adobe Acrobat 9.0 Paper Capture Plug-in Format : application/pdf Document ID : uuid:933c0008-9442-7c47-a488-83f38fd7224e Instance ID : uuid:c6e4dfd9-6028-8245-a696-69c2b8bda8a0 Page Layout : SinglePage Page Mode : UseNone Page Count : 153EXIF Metadata provided by EXIF.tools