CONTROLLING INTELLIGENCE INFORMATION

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP80B01139A000300110003-6
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 75
• Document Creation Date: December 12, 2016
• Document Release Date: June 29, 2001
• Sequence Number: 3
• Publication Date: September 1, 1966
• Content Type: PAPER

Approved For Release 2001/11/22 : CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/11/22 : CiA-RDP8QB01139A000300110003-6
SENDER WILL CHECK CLASSIFICATION TOP AND BOTTOM
UNCLASSIFIED CONFIDENTIAL SECRET
OFFICIAL ROUTING SLIP
TO
NAME AND ADDRESS
DATE
INITIALS
I
IHC Support Staff
2
Room E , Hea quarters
3
4
5
6
r
ACTION
DIRECT REPLY
PREPARE REPLY
APPROVAL
DISPATCH
RECOMMENDATION
COMMENT
FILE
RETURN
CONCURRENCE
INFORMA710N
SIGNATURE
Remarks
As requested a copy for your
retention.
FOLD HERE TO RETURN TO SENDER
FROM: NAME. ADDRESS AND PHONE NO. ~"~~' ~ ,DATE
19Au 69
IAL SECRET
UNCLA
FQIIM N0. ~] O 7 Use previous editions
1-67 LJ!
(40)
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Approved For Release 200~11/2~: CTA~DP80B01139A000300110003-6
Director of Intelligence Support
Directorate of Intelligence
Central Intelligence Agency
Approved For Release 2001,'I~,(2,,2~CJ(a.$DP80B01139A000300110003-6
 Approved For Release 200TTTT/22'''CI'A~DP80B01139A000300110003-6
Acknowledgment
A number of individuals have in one way or
another contributed to this paper, including -
(OCS) with particular reference to
ppenc x , and Project CHIVE personnel with
of the manusc:
particular reference to Appendix C.
(OCR) deserves special mention for mayor s~uppor
in the preparation of the report as a whole, and
my secretary, for the preparation
P.A.B.
Approved For Release 2001/11/22: CIA~-R+DP80B01139A000300110003-6
S~ C,[t ~i l
 Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
There are some things which cannot
be learned quickly, and time, which is all we have,
must be paid heavily for their acquiring.
ERNEST HEMINGWAY
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/1~ZZ:`GFk~7L'1~80B01139A000300110003-6
TABLE OF CONTENTS
I,
INTRODUCTION
TRENDS
1
2
II.
IMPACT
g
III.
CONTROLS
12
A.
Management Controls
12
B.
Selectivity
13
C.
D,
E.
Reduction
Division of Labor
Optimizing the System
15
lg
20
IV. AUTOMATION
V. CHALLENGE OF ASSUMPTIONS
APPENDICES: A. References
B. Selected Computer
Applications and
Techniques
C. CIA's Project CHIVE 44
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/1'f/Z2 "G1A~~80B01139A000300110003-6
Percentage Increases in Unique
Item Intelligence Documents 2a
Growth in Receipts of Individual
Title Open Literature Items 3
. Page Production of Translated
Foreign Intelligence Materials 4
Current Intelligence Production 6a
Schematic of Community On-line
Intelligence System (COINS) 21a
Schematic of Computer-based
Library of the Future 25a
Relation of Amounts of
Literature to Capacities
of Computer Memories
Illustrative Model of the
Intelligence Process:
- Collection 27a
- Processing, Analysis, Synthesis 27b
Project CHIVE
- Current System 47a
- CHIVE System 47b
- CHIVE Functional Overview 47c
- CHIVE Information Language 47d
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001~i~~+ r~~.~iDP80B01139A000300110003-6
Beginning with the first Methods Conference, we
have considered some aspect or other of the intelligence
information problem. This is not surprising inasmuch
as information is our business.
At eight years ago, provided
us with use u oughts on guiding collection._~ A
colleague at that time referred to consideration
of this particular aspect as contemp ating he tortured
progress of a complex organism in getting its food from
hand to mouth."
At that same conference, we also considered the
processing of information, and ventured some. judgments
about the nature and direction of prospective changes,
including the impact of automation.?t
At Washington, four years ago, brought
us up to date on the state of the ar a processing
information automatically.3~ And I injected a corrective
to the overly optimistic forecast made at the previous
conference concerning the appl?cation in the specialized
field of machine translation.4~
By and large these earlier reports and :findings
have held up well and remain valid. There would therefore
seem to be little profit in an attempt merely to update
what has previously been said. What would seem worthwhile
is to assess the impact on intelligence of the increasing
volume of information available. We should also strive
to formulate guidelines which might be useful to our
,superiors for coping with this so-called information
explosion problem.
Approved For Release 200~7~J~~~DP80B01139A000300110003-6
 Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
In the space of one hundred and seventy--six years
the Lower Mississippi has shortened itself two
.hundred and forty-two miles. That is an average
of a trifle over one mile and a third per year.
Therefore, any calm person, who is not blind or
idiotic, can see that in the...Silurian Period...
the. Lower Mississippi River was upwards of one
million .three hundred thousand miles long, and
stuck out over the Gulf of Mexico like a fishing-
rod. And by the same token any person can see
that seven hundred and forty-two years from now
the Lower Mississippi will be only a mile and
three quarters long, and Cairo and New Orleans
will- have joined their streets together....
.There is something fascinating about science.
One gets such wholesale returns of conjecture out
of such a trifling investment of fact....
,~ MARK TWAIN
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/1'f/T1 "CIA~DP80B01139A000300110003-6
I. TRENDS
Scholars have long cherished the right to complain
about their inability to cope with the amount of
information at their disposal. In 1613, it was Barnaby
Rich who referred to the multiplicity of books as one
of the diseases of his age. Books, he said, "doth so
overcharge the world that it is not able to digest the
abundance of idle matter that is every day hatched and
brought forth...."5~ In 1945, it was Vannevar Bush who
expressed concern over the startling growth in scientific
literature and described the investigator as "staggered
by the findings and conclusions of thousands of other
workers--conclusions which he cannot find time to grasp,
much less remember, as they appear."6~
Since then, whether in or out of Government, in
scholarly journals or during budget reviews, it is a
favorite theme of analysts to lament the unmanageability
of the in-basket, prompted, it is said, by the information
explosion. Most are quite ready to testify that this
situation is getting worse despite improvement in methods
for controlling information.
The reason for this is not hard to find. We
recently polled our collection, analysis, and information
handling staffs to determine, in-gross terms, what the
current trends in intelligence information appeared to
be and where the effects of change in data flows were
most severely felt. We indeed .found volume trends for
all types of information of interest to intelligence
invariably up. This is so for the large amo~~nt of open
source literature, that is, information published world-
wide in newspapers, periodicals, and books. It is so
for the amount of government field reports and the
finished analyses produced by the intelligence community
itself. It is very much the case as to techxiical
collection (SIGINT and overhead photography).
As can be seen, where we deal with receipts of hard
copy materials alone, the percentage increase since 1950
is very considerable. Only in the case of collateral
intelligence documents has the curve flattened since 1960
(see following chart).
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Approved For Release 200 ~~i~~~+~DP80B01139A000300110003-6
One of the most startling increases in information
receipts has been in the open literature of books,
serials, and newspapers. Volume figures rose sharply
in the years immediately following the death of Stalin
and reflect the relaxation of Soviet control over
external subscriptions and the early receipt of Chinese
S&T literature. Subsequent increases indicate the
continued upward movement of S&T literature and the
enlarged volumes in free world publications and in
publications from the emerging areas of Africa, Latin
America, and Asia which must be screened to identify
information of intelligence interest.
Growth in Receipts of
Individua it a pen i era ure Items
1950 1960 1966 1970
English S&T 2,100 4,700 5,300 6,000
English Soc- 5,300 10,100 11,900 13,700
iological
Foreign S&T 1,900 8,500 9,600 10,800
Foreign Soc- 6,700 16,200 19,100 22,000
iological
Understandably, translation of open source foreign
language materials has followed a similar trend, showing
a steady rise beginning with the relaxation of Soviet
control over subscriptions, and continuing with the rise
in interest in published materials from the emerging
areas. Changing requirements have brought entire new
areas of foreign language materials under scrutiny,
including those of Africa, Latin America and Cuba, and
South and East Asia. As an indication of the shift in
emphasis, our Foreign Documents Division, which in 1960
had only two translators assigned to Spanish-language
exploitation, now-has 14 people working on Latin American
and Cuban press materials and is issuing a daily Latin
American press exploitation repor t
Our total annual page production in translated
materials has doubled in the past six years, and it is
anticipated that by 1970 this output could reach a half
million pages a year.
Approved For Release 20 'Q,~~~DP80B01139A000300110003-6
 Approved For Release 2001/~/'Z2~"`C'I~~P80B01139A000300110003-6
Page Production of
Translated Foreign Language Materials
1950 Estimated
1960 1966 1970
59,000 212,000 425,000 500,000
To radio monitors, the information explosion is no
myth. A few examples illustrate the dramatic prolifer-
ation in broadcasting,
In 1956, Radio Moscow broadcast about 115 hours a
day to foreign audiences; this year the figure is 213
hours. Ten years ago Moscow broadcast in 40 languages,
today it broadcasts in 68. Then it carried about 77,000
"commentaries" or long items each year; now the annual
figure is 177,000.
Radio Peking's output also has skyrocketed. In
1956, it broadcast 47 hours a day in 15 languages, today
it broadcasts almost four times as much (166 hours) in
twice as many languages (32). Ten years ago the Chinese
broadcast 25,000 commentaries annually; that output has
now nearly quadrupled (97,000).
This increase in volume has been accompanied by new
demands for coverage as one-time Soviet "satellites,"
whose media could largely be dismissed as echoes of
Moscow, have become independent forces to be reckoned
with and as new problem areas in Latin America, Africa,
and South Asia have focused fresh attention on radios
in these areas.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/1'f/ZZ -CiA~[7P80B01139A000300110003-6
In the field of reconnaissance, there is in fact
an information explosion affecting the processor,
producers, and consumers of photo intelligence. Among
the various reasons for such an explosion, the-most
important perhaps, are the increased use of our ever-
improving reconnaissance capabilities. an increased
widening of interests into areas previously neglected or 25X1D
ignored by intelligence.
Prior to 1956 the primary source of this data was
captured World War II photography. With the advent of
means to collect aerial photographs of those areas of
vital interest to us the situation quickly shifted from
one of lack of photography to one of properly exploiting,
on a timely basis,the wealth of photography available. 25X1D
Volumes currently processed are many times the 1960
volumes. For example, during fiscal year 1965 an excess
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 200TT1'f/2~ : DTA~DP80B01139A000300110003-6
One factor not accounted for here is the rise in
copy distribution. Our documents staffs are disseminating
to more points than ever before. About 35,000 classified
intelligence documents flow into our dissemination center
each day (in about 12 copies each) and are disseminated
to more than 200 offices or individuals. In the case of
open literature the copy volume in 1966 exceeded 1.6 million
(1,237,000 daily newspapers; 270,600 journals; 114,500
books). This is a three-fold increase over 1950. By 1970
the figure will be two million.
In addition, figures indicating receipt; or production
of unique items do not reveal the increase in number of
pages issued yearly. There has been a steep rise in page
production for certain series of reports. Illustrating
this trend is the page production of translated materials
cited earlier and the output of "impressions" (individual
page reproductions) from our current intelligence office.
This latter figure rose steeply from an annual output of
seven million in 1955 to more than 30 million on 1965.
While much of this latter increase represents production
for enlarged readership rather than contributions to the
information pool, a good part represents additions in the
form of analytical commentary (see following chart).
One curious development, not usually taken into
account, is that the computer itself is currently aggrava-
ting rather than alleviating the information problem.
The brute-force power of EDP has in many cases been sub-
stituted for adequate analysis of a data processing
problem. This has been the case for many new computer
applications. It is only after the second or third expen-
sive iteration that a sensible perspective is obtained.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 25X1A gpproved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/~TZ2~`tTi~f2bP80B01139A000300110003-6
We see instances where basic input data is replicated,
reformatted, resorted, to the point where the output
volume exceeds input by perhaps 100 times. It seems
evident that the same methods applied to harnessing the
information explosion elsewhere in the intelligence
process must also be applied--perhaps more carefully--in
.the computer environment.
All in all a significant part of the information
problem is of our awn creation. The overall magnitude
for the American intelligence community of the informa-
tion handling problem was given some clarification in a
survey of information inventories and flows. This study
estimated that we annually produced or handled 20 thousand
individual series, in ten million issues, published in 150
million copies. While the amount originated by intelli-
gence agencies is only a fraction of these totals, it is
still considerable. For example, one component of one
agency alone produces some 44,000 information reports
annually. An average of 90 copies of each is distributed
initially fora total of four million copies.
Add to the above the fact that the advent of efficient
office copying machines has encouraged secondary reproduc-
tion by recipient offices to absurb proportions.
It is little wonder that intelligence analysts have
come to feel that they have in the space of a few years
moved from a condition of having too little to having too
much information.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
S'ECR ET
 Approved For Release 2001/11/22 ; CIA-RDP80B01.139A000300110003-6
Where is the Wisdom we have lost
in knowledge?
Where is the knowledge we have
lost in information?
T. S. ELIOT
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 200TMfirZZrG7A~DP80B01139A000300110003-6
II. IMPACT
At this point it is well to make some distinctions.
Not all analysts are equally effected by changes in
information availability and receipts. Moreover, the
demands made by policy makers and planners upon produc-
tion analysts are themselves subject to considerable
change. The relationship between the causes of change
in the demand for intelligence and in the availability
of information upon which that intelligence i.s based may
be a significant one. But this has not been the matter
of serious study and little about it is known.
It is important therefore to resist the tendency to
view the aggregate receipts from all sources---teletype,
radio and TV, ground and overhead photography, COMINT,
maps, books, newspapers, and serials--as equally affect-
ing all elements of .the intelligence process. Analysts
throughout the intelligence cycle are affected by the
changes in the information flow according to their area
and type of assignment. It is not often the case that
an individual analyst requires all types of incoming
materials to do his job. One of our economic research
areas, for example, indicates that open source and recce
information comprise the bulk of the significant data on
which their analysts rely. Specialization by area and
topic also tends to reduce to manageable proportions the
volume of sources a analyst must examine. In addition,
the production analyst is shielded from the full force
of the information flood by an elaborate filtering and
requirements process, and by a wide variety of indexing,
abstracting, collating, and reference services.
Information deriving from new collection means or
from open literature has had its greatest impact on the
military, economic, and scientific/technical areas with
the main emphasis falling in the Soviet Bloc;. However,
with the exception of open source information, the bulk
of the new material arrives as raw data of which only a
fraction expresses itself in the form of published materials
through which the production analyst must make his way.
It is probably a safe generalization to say that, with
certain exceptions, many research or production analysts
are not acutely aware of living in the midst of an
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/i1~2CZ:fA=RDP80B01139A000300110003-6
information crisis. Indeed in some cases we may find an
information drought in an area of plenty, largely for
organizational or logistical reasons. For example, the
main reason for liberal copy creation-and dissemination
is to get the word to those who need to know. This has
to be done through set channels of communications, The
resaurces devoted to the operation of these channels must
not be disproportionately high to the resources allocated
to the substantive use made of the materials transmitted.
A point is therefore soon reached beyond which the system
will not carry additional traffic. In this squeeze
dissemination policy is reviewed, "marginal" accounts are
dropped, and with the passage of time a given analyst will
not be receiving what he needs. What's more he won't know
that he isn't.
This indeed emerged as the main conclusion of a study
undertaken to find ways and means to staunch the flow of
paper to the analyst. We thus have the paradox of too
much paper for the analyst to handle but not all that he
ought to see to be on top of his job.
The scientific analyst is one who is all too well
aware of the difficulty of staying on top of the S&T
literature. Where, formerly, he screened the literature
himself, he now relies on translation services to scan,
select and abstract material for him. By 19'70 the volume
of S&T material will be overwhelming, and mechanical means
will be required to scan and select important items and
collate information.
The full farce of the flood is felt by analysts
working in collection, data processing, or in reference
activities. It is from these quarters that the greatest
anxiety about an "information explosion" has been voiced,
for it is these processors (intelligence support analysts)
who face the huge data reduction problem or must cope with
the cumulative effects of high volumes of information on
their ability to transmit, to disseminate, to store, and
to retrieve.
Certain changes in the climate in which analysts work
tend to create a further effect of large-scale expansion
of the information load. Among these changes are the
emergence of new topics or world areas of intelligence
Approved For Release 200?j~1~2~~.EITRDP80B01139A000300110003-6
 Approved For Release 2001/1'f/ZZ Z:iAlZDP80B01139A000300110003-6
significance. Fifteen years ago the. world's sovereign
states numbered 80. Today there are 128. A rise in
topics of interest has accompanied the growth of nations;
a decade ago we were not concerned with a French or
Egyptian missile-program,
A gazetteer on ntarctica is about
e pu is a wi over 11,000 place names. Hundreds
of new personalities now play important political roles.
There is a tendency to spread analysts fairly thin
in assigning work on new areas. In addition, new re-
sources, and some resources diverted from established
activities, have been required to man new collection,
requirements, and data handling activities. Technical
innovations, particularly in communications, have brought
the reporting and response to events into a "real time"
mode. Events are reported as they are happening, and the
report from the field not only informs but often generates
requests, creating an information-request-information
spiral.
In general the production analyst has made few
changes in his traditional approach to his tasks. In
areas burdened by manpower shortages or by heavy document
or ad hoc production loads he may tend to read more
casually, to treat mare information as background, to file
fewer document s, and cancel subscriptions. One office: has
placed its filing activities on three levels of activity:
active research files, files which are maintained but on
which no research time is expended, and, for areas not
being actively followed, files which include only the
finished reports of others. In some instances analysts
will cancel receipt of material on inactive fields but
will expect the information to be collected and filed by
central reference against the time when the field again
becomes active or when basic studies must be updated.
While there is a tendency to place more reliance on
central reference functions, particularly on such
specialized services as those provided by biographic and
installation files, there is apprehension about the lack
of an all-source capability and about the inability to
retrieve materials by subject and area. In most cases,
the analyst continues to regard his active file as the
best, relying on central reference service for backup,
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/~/~2CCIA`RGP80B01139A000300110003-6
for developing information on new areas and topics, or as
a check on dissemination. Reliance on summaries and hand-
books produced by central reference or other research
areas has increased, and these products fill a need for
basic reference materials.
While analysts have been active in changing require-
ments and in making more strict definitions of their
missions, they are often called upon on short notice to
answer questions on matters which they have not been able
to follow in depth. Nearly all production offices complain
of the disruptive and time-consuming nature of the current
re~,uirement to divert time from production o:r analysis
into ad hoc activities. One component has indicated that
some analysts are concerned with work on ad hoc problems
to the point of falling out of touch with their research
fields. This situation indicates a need to review mission
statements and organizational philosophy in the light of
current response requirements.
Approved For Release 20~ECRET RDP80B01139A000300110003-6
 Approved For Release 2001/11/22: CIA=RDP80B01139A000300110003-6
...it takes all the running you can
do to stay in the same place.
If you want to get somewhere
else you must run at least
twice as fast as that:
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/?'f/Z2 `"~C1A`R~P80B01139A000300110003-6
Thus far we have seen that an information problem
exists, and that its impact upon analysts varies, but is
considerable, especially upon analysts in the support
services. Let us now look at some of the techniques
which we actually employ or are devising to control
information.
A. Management Controls
A number of basic and direct techniques are always
at hand to control the data flow. These are in the area
of exclusive or arbitrary controls, e.g., cut manpower/
increase manpower; cut budgets/increase budgets; apply
output limitations; reorganize data processing activities;
eliminate editorial controls. All of these have been STATSPEC
applied at one point or another in the intelligence cycle.
essful abnlication of such an arbitrary control is
Reduction of manpower and budgets is a:n often used
.control. It, also, imposes high selectivity standards
and has the advantage of requiring the victim to look
sharply at his mission description, not solely from the
paint of view of what his mission is but also is not.
The expansion of the all-source concept has provided
gains and still shows considerable promise. The increase
in cleared personnel has allowed a free exchange of views
and documents, has made it easier to obtain collated
information, has bridged the gaps caused by compartmented,
limited source activities. It offers an overview of what
is known on a subject from o 'al all the
way through the products of
systems. It has been of particular value in assisting
rapid evaluation of recce data.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/T1T~2C~TA~DP80B01139A000300110003-6
Reorganization is a way of life in the intelligence
community. All too often, however, reorganization is
inspired not by an examination of existing responsibilities
and work patterns, but by the emergence of new requirements
and priority areas. At present our production directorate
is examining a number of activities to seek more effective
combinations of functions and areas of assignment of
responsibility.
B. Selectivity
One major organizational advance was the formation in
1963 of a Collection Guidance Staff (CGS) to serve as a
focus for coordination of collection-activities, particu-
larly those involving high investment, technical methods.
One feature of the CGS is an all-source requirements
registry which can match information to intelligence
needs, recall reporting on a subject or area for evaluation
by analysts, and identify reporting which is not related
to stated intelligence needs. The coordination of require-
ments and output from various collection means has enabled
us to limit collection in many areas. In addition, the
gain by the user in technical knowledge of collection
capabilities and limitations has resulted in more refined
and meaningful guidance to the collector.
Advances in communications technology, tteletype and
long-distance facsimile, have not only made information
more timely but have enabled the analyst to get to the
collector much faster with much more information than
ever before. This improved dialog has made collectors
more responsive to guidance and has made analysts more
appreciative of good collection guidance. There has also
been increased initiative on the part of the collector in
reviewing requirements and reporting needs.
In photo collection, attempts to exclude margina:L
information at the source have centered on trying to hold
down the number of reconnaissance targets. These efforts,
however, are complicated by the desire of each area or
subject specialist to regard his particular set of taY?gets
as valid, necessary, important, or critical. In addition,
there is a reluctance by all consumers of photo intelli-
gence to permit even marginal, poor quality photography
to go unexamined and uninterpreted. The main technique
currently being used to place both the collection and
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 200' 11~!'~~-'~DP80B01139A000300110003-6
exploitation (interpretation) of reconnaissance photogx?aphy
in order of importance is to accept the requirements of
an inter-agency committee or working group (where avail-
able) which is concerned with a particular problem rather
than to accept those of individual analysts. While
effective in providing priorities,this has had little
effect on the total volume of material collected in this
relatively .new area, though it has been more successful
in dealing with other sources.
One of our offices has made good progress in controlling
dissemination of hard copy materials by encouraging its
analysts to accept search responses in microform. Aperture
cards, containing up to eight microimage document pages,
are sent to requesters who must return the cards when no
longer needed. This system has had the advantages of
economy and security, of reducing the number of hard copy
documents being filed, of reducing actual file sizes, and
of good customer feedback--by punching out a hole in the
aperture card_the user may request hard copy or indicate
that the document ought to be purged from the file. Other
forms of microform service are under examination, includ-
ing the possibility of reducing reference and indexer aids
to rapid-lookup microform systems.
Approved For Release 2001'~~8-~DP80B01139A000300110003-6
 Approved For Release 2001/'I''fi'Z2~C;GA=RDP80B01139A000300110003-6
Various types of non-mechanical dai;a reduction
techniques are being used with good results. One of the
simplest of these is t:he use of title listings or
announcement bulletins to inform users of tYie availability
of materials, avoiding the need for scanning, abstracting,
translating, or dissemination of large volumes of material.
We have a system through which the tables off' contents of
selected domestic and free world scientific journals are
copied and sent to users in accordance with their estab-
lished coding requirements. The Russian Book List gives
translated titles of incoming Sovie oo s, i- n~orming the
analyst of their availability and enabling trim to determine
which ones he wishes to review or have tran,~lated. The
Consolidated Translation Survey is a monthly record of
trans ate items or i ems in a process of translation.
It.serves as an announcement of availability of trans-
lation of items of interest and as an antiduplication
device. Accessions lists of maps, photos, and other
limited copy or hard-to-disseminate items ar.e in wide use.
The application of a large array o:~ indexing
schemes is used to control large document collections and
to reduce reading and search time. Machine-controlled
indexes guide the user to information in COMINT and
collateral documents and provide special controls over
biographic and installation hard fact information.
In the area of published indexes, the Intelli-
gence Publications Index provides the user with a monthly
catalogue of finished intelligence publications and
reports.. Published indexes to selected foreign newspapers STATSPE.C
or radio broadcasts serve not only as guider to further STATSPEC
exploitation but as press emphasis indexes and as a
reference to important speeches or articles.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 200'f1~22''C~ARDP80B01139A000300110003-6
Simple indexes of "keywords" which can be searched
in combination are useful guides to large data collections
and are also of value for dissemination. Keyword title
listings are used to provide minimal control over the
large volume of COMINT data and are valuable in freeing
the data from the rigidity and ambiguity of a classifi-
cation schedule. While it is unlikely that users will
accept indexes as a substitute for dissemination in fields
they are actively following, indexes may be accepted in
place of documents on lesser priority topics, as a check
on dissemination, or as a desk reference.
One such index, which will be issued in 'the near
future, is the SKAN (Subject Keywork ANnouncement) index
of incoming ChiCom documents. This keyword-aut-of-context
listing will be the first product of the computer system
being implemented in our central reference facility.
This index can be sorted by topic and area and special
issues, of interest to :individual analysts, g>repared, and
distributed.
3. Synthesis
Certain types of data known to be of general
utility, particularly those reported from many sources,
can be effectively synthesized and reduced in the form
of listings, indexes, and summaries. Often these are
prepared for rapid publication and distribution. An
example of the publication of selected data elements is
publication of
division. in s e intelligence publications are he
most significant product of synthesis, with the NIE
reflecting the largest reduction ratio.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/1''f/Z2 "CIA'Rd'P80B01139A000300110003-6
4. Abstracting
There is an expressed need for the production of
more information containing abstracts or prepared in the
form of abstracts as an aid to rapid review and identifi-
cation of data of importance. Producers of reports provide
service to screening by including introductory abstracts.
These often contain enough information in the way of con-
clusions to stand alone or to save reading the entire
report to determine its interest or value fo:r filing.
Abstracts of foreign literature, particularly books, are
also required. In the vast area of S&T literature an
active contract program, coordinated between CIA and the
Department of Defense, arranges for the preparation of
abstract translations of items from Soviet Bloc journals
and monographs. The results of this load sharing are
published together in topical serials (Bio-Medical Science,
Chemistry, Electronics) and-are made available to the
scientific world through the translation service of the
U.S. Department of Commerce. Not only do these abstracts
overcome the barrier of many languages and technical
vocabularies, obviate the need to maintain large numbers
of individual subscriptions, but they enable the user to
screen large masses of data rapidly and with ease.
In the central document files an abstract system
known as DARE (from Document Abstract Reproducing Equip-
ment) has been implemented. ~1'his system, by copying and
reducing the first page of a document onto a card which
can be sorted by EAM equipment, eliminates the need for
tedious and costly human abstracting and often enables
the user to determine at a glance whether or not the
document meets his requirements without the need to work
toward the document by way of an abstract statement.
Since an estimated $1% of all information reports consist
of a single page, while 34% are limited to two pages,
savings in time and effort of both the processor and the
user are significant.
5. Formatting
Preparation by the originator of data in a form
suitable for immediate file input can be a major step in
shortening reading and filing time. Good progress has
been made on this front and a wide variety of programs
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001M''I'/2~'`aCI~F~DP80B01139A000300110003-6
exist in which data is formatted in such a wa.y as to be
ready for filing upon receipt by the user. Our China
Plant Data Report is an example of a collection format
which structures significant data elements according to
the logic of the user's file. Plant names are highlighted
for rapid identification and screening, and the inclusion
of only "hard fact" or meaning-bearing words (locations,
production figures, floorspace) eliminates tYie need to
translate and. read much superfluous text.
There are a number of formatted souY?ces contain-
ing biographic information which are received on contract
or from other government agencies. Hard fact; economic,
technical, and geographical information is similarly
prepared for Recce collection, in addition
to making requiremen s more specific, circum;~cribes the
collector's response within rigid reporting formats.
Highly formatted data lends itself readily to
computer processing. Key punching is simplified or
flexotape can be prepared at the time of processing for
later conversion to magnetic tape. A number of appli-
cations, which will be discussed later in this paper, are
in process in this field.
D. Division of Labor
In a period of tight budgets and expanding requirements
it is essential that duplication of effort be eliminated
to the greatest extent possible, that cooperative ventures
be developed whenever feasible, and that we rely on the
work of others to accelerate the pace of our research and
ease the burden of our personnel.
Obviously a clear delineation of mission responsibility
and the elimination of overlap by organizational approaches
offers promise in this area. The grouping of "like"
functions into single organization entities is useful in
providing more integrated production and less duplication
of effort than is required by coordination processes.
The bringing together of experts to comment on a given
problem also offers opportunities for gains in time and
efficiency. We have had some experience in this area with
still another type of committee--the task force--which is
often formed during a crisis or in anticipation of the
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/"I'1722'`Z:1P~RDP80B01139A000300110003-6
25X1 B ~
25X1 B
emergence of a critical problem. Such task forces have
.been in operation on Cuba, Vietnam, the Dominican Republic,
China, and the Soviet succession issue. These bodies
bring together experts from all intelligence disciplines--
collection, analysis, reference--and provide a highly
integrated, fully informed attack on the target problem.
The degree of success which these operations have achieved
probably has strong organizational implications and should
be studied from that point of view...
tion and analysis. A recently translated publication
There are other, more obvious forms of load sharing,.
Handbooks, statistical yearbooks, the finished work of
Free World economists and the publications of research
foundations and corporations provide reliable basic
materials which eliminate the need for additional collec-
dictioxiaries, gazetteers, anal city
direc ories, particularly in the Free World can be used
and filed as received. Even Soviet
can be relied on to give at least the outlines of a
which could not easily be pieced together from
e co lection of many identifying items. Internally
25X1 B
25X1 B
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECR E T
 Approved For Release 2001"RI~F~DP80B01139A000300110003-6
produced handbooks- geographic, general area
survey--are widely usec as asic aids to research,. thus
saving the time of many analysts in screening documents,
maintaining files, and searching for information.
E. Optimizing the System
This catch-all can be said to cover a wide area of
techniques and methods which could affect tYie environment
within which information is handled, the re]_ationships
between different parts of a system, the form of the
information itself, or man-machine relations, Most of
the techniques are not unique in their application to
information handling, and I do not propose to dwell on
them at length. They are of importance however. They
can be used with expectation of realizing efficiencies
in the activities studied.
1, Systems analysis
Techniques for optimizing a system would include
systems analysis and operations research to study flow of
data patterns, to locate bottlenecks, to identify over-
lapping functions, to eliminate unnecessary ;steps, to
simplify channels of communications. Systems training for
personnel is required to qualify them to bring about
orderly change--and these should if at all possible be
members of your own staff. Managers need such training
also to enable them to make informed decisions about the
applicability of automated techniques to the~_r operations.
The high volumes of today's information, and the
rapid processing and response speeds required, indicate
an increasing dependency on exchange of data in 'the form
of tapes and graphics, the ability to rely on and remotely
query the files of others, and the capability to collate
data from many sources in a single product. This requires
standardization, that is, speaking the same language and
agreeing on the meanings of forms and terms. In addition,
the requirements of computer processing are compelling us
to seek common formats and languages to facilitate
computer input and interface, Moreover, standardization
permits a wider application of the concept of division of
labor.
Approved For Release 2001~~'~~'gDP80B01139A000300110003-6
 Approved For Release 200'I~'~',;~~'~bP80B01139A000300110003-6
The difficulty of standardizing on a broad front
is illustrated by the failure after 20 years to agree on
a standard Cyrillic to Latin transliteration system,
Standardization problems are among the issues on which
USIB Committee on Documentation task teams have been
meeting over the past year. Goals include standards for
geographic and installation codes, for film storage media,
and for report identification.
3. Program and organization planning
With program planning we hope to achieve a better
job of allocating resources to information handling
programs responsive to priority needs. By flexible
organization we endeavor to anticipate requirements for
organizational change.
4. Analyst communications
Analyst communications is another very important
part of the system. Analyst files represent a refined
research product ready for response on special problems.
Analysts themselves represent a community of available
experts competent to provide information or informed
judgment. We need to improve the channels of communi-
cations between these experts and files, the means of
locating what is known, who knows it, and how to bring
the full range of available expertise to bear on a given
problem.
In the realms both of sharing the burden and of
analyst communications, U.S, intelligence has initiated
an experimental Community On-Line Information System
(COINS). This will permit an analyst in one agency,
through a remote desk console, to query a fi:Le stored in
the computer of another agency (see following schematic).
5. Upgrading the analyst
Upgrading the analyst is another promising means
of handling information more successfully, The production
analyst can be upgraded by more specialized training
(including foreign language training which would reduce
the load he places upon the. translation facility), or by
being given support (human or machine) designed to relieve
him of the less productive aspects of his task. We are
doing both.
Approved For Release 20019'~~~-~'DP80B01139A000300110003-6
 Approved For Relea a 2001/ 1/22
! l_ L
CIA
TERMINAL
COMPUTERS
IA-RDP~0B011 ~9A000~}01100Q3-6
DEFENSE & STATE
TERMINAL COMPUTERS
COMMO
STORE and FORWARD
SWITCH
Schematic of
Community On-line Intelligence System
(COINS)
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/~T1'2~CT~RDP80B01139A000300110003-6
We have long provided massive reference and
support services to our production analyst. The present
thrust is to make reference personnel a more active
participant in the production process and the production
analyst an active participant in a system design to
support him.
In photo interpretation, for example, we have
had good results in teaming the PI with the all-source
analyst in work on specific problems. Our
support service has also worked well. The question is
no longer whether there shall be both a political analyst
25X16 and in support, or a single intelli-
gence officer to 0 o jobs. Rather the question is
to determine the optimal relation between the two. Nor
is the question whether the shall
or shall not receive machine support, but how twat support
is to be organized.
A recent user survey, conducted in the Department
of Defense, provided statistical evidence for certain
intuitive views we have held for some time. First, that
specialists in information handling can not only fulfill
user needs but often anticipate needs before the user is
fully aware of the requirement. .Secondly, in 68% of the
DOD sampling, the user preferred to obtain a detailed
analysis or a specific answer to his request rather than
a set of abstracts to documents or, presumably, documents
themselves. These survey results have meaning for our
own activities in how we use our reference staffs; we need
to gain a clearer view of the function and potential of 'the
information specialist in creating new information through
synthesis and correlation of available data, and of his
responsibility for monitoring the user's process of informa-
tion acquisition, use, and generation.
6. Providing user feedback
User feedback, whether related to collection,
reference, or production, is an essential element of any
system and one of the least satisfactory element of our
own. Often analysts work for long periods completely
in the dark as to how satisfactory their work is to users
or to what purposes it is being put. There exist also
problems of making the full range of services or functions
of one office known to potential users in other office.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001M'172Z"'`"CIA=RDP80B01139A000300110003-6
Our reference staffs have been seriously concerned about
this matter. Conversely, reference staffs have few ways
of determining what the user's problems are except to
the extent this information can be derived from request
patterns or through the tedious and dubious process of
surveying user needs. Often the capability of information
personnel to tailor responses directly to user needs goes
unused because of inadequate communications. It may be
time to think about spending less time attending documen-
tation meetings and more time studying modern retailing
philosophy. The situation between the production analyst
and his customer, the policy maker or planner, is also
unsatisfactory, though we do have cases where unexceptionable
relationships have been developed to the mutual benefit
of producer and recipient.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
Cold blows the wind on my true love
And a few small drops of rain --
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/'I`Tf2`1~CTd`~2bP80B01139A000300110003-6
IV. AUTOMATION
Quite recently the Saturday Review devoted a special
issue to the potential of the autorr}ation revolution and
its implications for oux~ society.? This forecast is an
impressive one. So impressed were the editors that they
equated the computer with the atomic bomb as a technolog-
ical development formidable enough to mark a turning
point in human history,
A Newsweek report,~~ prophetically entitled "Good-by
to Guten e~-also gave readers a glimpse of things to
come in information technology: a photosensitive crystal
the size of a lump of sugar that is capable of containing
images of 100,000 pages; a lensless photographic system
which could lead to three-dimensional home television; a
no-contact, no-pressure printing technique that can write
on sand, print a message on a pizza, and put a trademark
on a raw egg yolk.
Such concepts, and dozens of others are indeed
incubating in industrial laboratories and being translated
into hardware. This progress is being heralded as the
final phase of the U.S.'s $10 billion communications
industry's transformation from the old Gutenberg ways.
Marshall McLuhan for example predicts that books and news-
papers will no longer exist; that publishing will become
an active servicing of 'the human mind througYi research
packages done to suit individual needs.9~
To those of us who have read similar promises over
the last 20 years, these words smack of extravagance. I
do not doubt that amazing developments will continue to
take place. I do doubt that we can count on them for
early and revolutionary solutions to our information
handling problems. The kind of problems to which I have
referred earlier have a habit of staying we17L in front
of innovations.
In CIA we are for example currently upgrading our
computer facilities to third generation hardware infinitely
superior to our initial gear, Yet the contribution of
the computer to the task of producing intelligence is
still both specialized and very limited. We have secure
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECR E T
 Approved For Release 2001~9~~~:DP80B01139A000300110003-6
telephone communications. But these are far from ideal,
the number of instruments are few, the cost of each is
high. Great- strides have been made in our printing
establishment. Still, the lapse between preparation of
copy and its availability to the anxious reader can be
measured in weeks rather than days for nonpriority items.
Reproduction techniques have shown major gains. But
material of intelligence interest received in such poor
quality that it cannot be microfilmed runs in some
categories as high as 20%, We have improved our training
of people and our means of furnishing them with better
instructions. This notwithstanding, 50% of the titles of
some report series need to be rewritten in headquarters
in order to reflect substantive content meaningfully.
I do not cite these findings to ring the changes of
gloom. I cite them as a call to caution against going
overboard in diverting resources from the workable (if
less than satisfactory) old to the untried new.
Having said that, I nevertheless urge that mare
attention be paid by intelligence to what is going on out-
side: As never before we have opportunities to capital-
ize on the work and ingenuity of others as a source of
relief for some of our own problems. Much of the work
done outside is solid and relevant. We ought therefore
to pick-a-back whenever we can.
Let me mention a few of the more important outside
activities I have in mind. Two are noteworthy in the
academic community at large--TNTREXl~/ and EDUCOM.
INTREX, which stands for "information transfer experiments,"
has been called progress toward a dial-a-thought world.
INTREX is setting up an experimental laboratory to test
-ways of giving professors and students instant access to
information. Xerography, film projection, and telephone
communication between computer and user are planned.
Basically experiments conducted will (1) attempt to auto-
mate and rationalize the functions of libraries, and
(2) develop an information transfer network which is com-
puter based (see following schematic).
EDUCOM,11~ the Interuniversity Communications Council,
is an organization of over 30 universities in 20 states.
It aims at evaluating the significance for higher education
of electronic hardware (computers, light pens, graphic
displays), and software (computer programs).
Approved For Release 2009'f~~~DP80B01139A000300110003-6
 Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
9.-it1I1~~~~IC~L~I
~~I:~IbY~:~~
~~1~1-1~l1:~i~~i~Cll-~
3~i p~0~.~1-11:rrCi]1~
~~I~l~-iilljl~i~.
~~~1CIt;-~I:~i~1;
~i~~~"yyyyJlll~~
~.~~~6:1 ~~~~
_. _. _. _
~~~III11111-II~111~
~~~>-~Ii1X~~:~
request
~?~~il'r1111M LL'
~~.~~IIII~~ ~ ~ ,
~ ~~~~~i~l~- .iii.
~~~,~~Iil~lll--1~~1~
replica o~
document
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 THIS TA ~~~r~~dA~l?p`~~~~ l~~( ~~1'~~ t
1139~000300110003-6
TO CAPACITIES OF COMPUTER MEMORIES ~`
World?s total store ofliterature -Year 2000
10,000,000,000,000,000
I~lorld's total store ofliterature
1,000,000,000,000,000
'4 4
All literature in science and technology
100,000,000,000,000
Capacity of fast random-access memory -Year 2000 ~~:P
10,000,000,000,000
Capacity of serial-access memory being developed
1,000,000,000,000
Literature in average field of science
100,000,000,000
Literature in average sub-field of science
Capacity of fast random-access memory -Year 1980
10,000,000,000
Capacity of magnetic disk file
Contents of Encyclopedia Britannica
1,000,000,000
100,000,000
Information in average book
10,000,000
Capacity of fast random-access memory
1,000,000
'rThe scale measures the number of bits, or binary digits, of information that can
be stored in the computer memory, or that are required to store the contents of
the literature. All figures are estimates for 1964 except as noted otherwise.
The capacity of fast random-access memory, now of the order of ten million
bits, ig4~S~21i~R~f~14~N~~~~e2~0~0~1~1'~1~2~' : ~%~~2C'~P80B01139A000300110003-6
 Approved For Release 2001/~~~;'~~~7P80B01139A000300110003-6
A number of individual university libraries have
forward looking programs, Washington State12/ and Florida
Atlantic13/ to name only two. The latter library has the
distinction of being the first in the U.S, to introduce
data processing methods and techniques into its operations
from the very beginning. Washington State on the other
hand is converting from traditional library methods to a
totally on-line library system which offers multiple
remote access to a single library record. Thus by sharing
the time of the university computer (IBM 360/67}, it will
be possible to reduce typing substantially, eliminate
duplicate manual files and have complete control of each
item's location and status in the library system.
The value of these projects to us is that they are
more in keeping with the size of projects which intelli-
gence libraries may undertake. While much valuable data
has been published about ways and means for automating
the Library of Congress,14/ the sheer size of the holdings
makes many of the. parameters of that undertaking of little
practical value to us.
The publications of professional engineers, documen-
talists and specialists in various aspects of the informa-
tion handling industry are also increasingly more solid
and relevant. Particularly useful books, fo xample,
have been ppublished by Licklider,15/ Baurne,~6~ Becker
and Hayes,17/ Greenberger,18/ Feigenbaum and Feldman,l9/
and Borko:~/ The National Science Foundation's "Current
Research and Development in Scientific Documentation'121/
serves to alert readers to on-going experimentation in
information handling techniques. And the proceedings of
professional meetings on this subject ggive first-rate
coverage for those who cannot attend.22/
I could go on to enumerate the many fine papers which
have recently been published in professional journals,23/
but such additional evidence would only serve to emphasize
the point which I hope has already been made: there is on
the outside of the intelligence community much wisdom and
talent which we have neither tapped sufficieni;ly nor used
effectively because we are ill organized to do so. This
work is relevant to our needs: we cannot afford to
duplicate it; we must therefore learn to exploit it in
our own behalf,
Approved For Release 2001/~TI~~IP80B01139A000300110003-6
 Approved For Release 2001/'x'1722''CIA'-Ff~P80B01139A000300110003-6
In the past, CIA has had no organization worthy of
the name to identify this outside work and to relate it
meaningfully to our improvement programs, This gap has
been filled with the organization of our Intelligence
Sciences Laboratory. This facility is acquiring its own
computer and associated equipment to provide an experi-
mental environment closely approximating actual operations.
Illustrative of its areas of interest are on-line process-
ing for analysis, pattern recognition, language and text
processing, and speech and audio processing, We will thus
better bridge the work done outside and our own EDP
related operations (see following schematics).
While CIA pioneered much automatic information
processing activity with its punched card equipment, our
experience with general purpose computer operations is
short of six years. In those years we have very consid-
erably expanded our use of these machines.
The use o~ 'cne computer
in intelligence processing has in the last three years
expanded to the point where we see applications in almost
every element of the intelligence cycle. Some of the
applications and techniques which we consider particularly
useful or promising are summarized in APPENDIX B.
A word on how we determine suitable applications may
be of interest. There are essentially two basic
approaches. In certain large data handling areas, par-
ticularly where the information is numeric (military
economic costing data, agricultural data, including rain-
fall and soil moisture statistics) or formatted, computer
applications have been initiated by the production
analyst with good results. Since data preparation in
most cases is done by the EDP staffs, these applications,
particularly in the economic area, have produced high-
yield products which .require little investment on the
part of the analyst. A wide range of applications of
this type remain to be tried, including computer control
of Soviet S&T literature. Projects of this kind, involving
primarily collaboration between a production analyst and a
representative of the applications division of our central
~ computer facility, can be characterized as special projects.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECR E T
 Approved For Release 2001111122 :CIA-RDP80B01139A000300110003-6
SECRET
AN ILLUSTRATIVE MODEL OF THE INTELLIGENCE PROCESS
COLLECTION
DEVICE
DEVICE
DEVICE
DATA
MODEM
IMAGE
RECOGNITION
DATA
FILES
IMAGE
FILES
P. I.
DEVICES
MACHINE
SOFTWARE
DOCUMENT
TAPE
SCREENING
PATTERN RECOGNITION
MULTI-SENSOR
CORRELATION
SECRET
i
----- CONTROL ---------------------?---_-
52323 9-65 CIF
(Sheet 1 of 21
Approved For Release 2001111122 :CIA-RDP80B01139A000300110003-6
 1
t
1
i
SECRET
CHARACTER RECOGNITION
TEACHING OF COMPUTER
TTY
DATA LINK
HARD COPY
Approved For Release 2001111122 :CIA-RDP80B01139A000300110003-6
MESSAGE
DISPLAY &
BUFFER STORE
MESSAGE
DISPLAY &
BUFFER STORE
AN ILLUSTRATIVE MODEL OF THE INTELLIGENCE PROCESS
PROCESSING -ANALYSIS -SYNTHESIS
~----- 1
I I
I FILE TYPE I
FUNCTION
APPLICABILITY
NON-REDUNDANT
PRIORITY
COMPARISON
PATTERN RECOGNITION
SYNTHESIZE
STATISTICAL CAL.
TECHNIQUES
LANGUAGE
PRESENTATION
PRINTERS
PLOTTERS
PROJECTED DISPLAY
CATHODE RAY
REAL TIME STATUS BOARDS
I
i ~~------~--.----- i
DIGITAL I TAPE DISC ~
I I I
I---------~--------i
I
STATISTICAL CALCULATIONS
SET PROCESS (CONTROL) LIMITS
THRESHOLD SETTING
TRIGGER PROGRAMS
COMPUTER I REElS I FILM I
PLATES I
I CHIPS I
DISPLAY ~----___-~--------I EXPECTED
DEVICE I j l UNEXPECTED
1
ASSOCIATIVE MEMORY I
I
PARALLEL DISCS ~
HARDWARE '
ESTABLISH 1
UPDATE
MERGE
OUTPUT
PURGE
ROUTING
DISSENJNATION
52323 9-65 CIA
(Sheet 2 of 2)
SECRET
I ESTIMATING
REPORT
GENERATION
FOR SPECIFIC
INTELLIGENCE
OBJECTIVES
CONTROL ------------------------?----~-------~----------------------------------------- --
- ?-
PUNCH TAPE 8
COMPUTER (DEVICE)
Approved For Release 2001111122 :CIA-RDP80B01139A000300110003-6
 Approved For Release 20~'f/'f'f/2'#:-CTA~DP80B01139A000300110003-6
At the other end of the scale is the ambitious
attempt to change in a fundamental way the present method
of doing business in a major component of the Agency.
Our Project CHIVE illustrates this. In one way or
another this general project will change the way hundreds
of analysts are now working.
There are many unique features of the CHIVE approach.
The CHIVE team is a fully integrated group drawn from
production, reference, and computer components of the
Agency, and includes contract personnel as well. In-house,
experienced operators of information systems who have been
given training in advanced techniques are in charge. The
prospective user of the system is drawn in as active
participant. An important geographic area, China, has
been selected as the first increment of the system, but
will be conducted as a pilot operation in parallel with
the old system to permit experimentation before going
operational.
The need for CHIVE arose from developments over the
past nearly 20 years, during which we evolved a number of
special-reference services to support the production
analyst. The multiplicity of classifications, of indexing
tools used for control, of formats employed to collect,
disseminate, store-and retrieve, made it increasingly more
difficult to meet customer needs. The problem of hetero-
geneity was compounded by the increase in volumes
received, and, with the passage of time, the volumes in
file. Moreover, the output expected of the production
analyst of today calls for greater sophistication within
shorter deadlines. Project CHIVE is designed to help him
meet that challenge.
Same of the parameters and characteristics of Project
CHIVE are summarized in APPENDIX C. Only because computer
technology, capability, and capacity are what they are
today and will be tomorrow do we dare count upon the.
success of this project.
And risks there are. This we know. In this costly
field it is well to seek ways to reduce risks, but after
you have done all you can they are still onsiderable.
We recently had a series of meetings wit
during which he shared
Approved For Release 2001/e1s1'ECRE ~P80B01139A000300110003-6
 Approved For Release 200'f/~/~RCTA~DP80B01139A000300110003-6
with us the experience of others in using computers. He
reported that of 27,000 systems installed in the U.S.
today in some 16,000 installations, 40% are unsuccessful
in the use of their computers. This means that 6,500
organizations are not deriving economic benefit; or are
.not achieving their objectives. In JO% of these cases
schedules and bud ets have been exceeded. The main
reason for this i iew is that people
using the machines aY the technology. In
short, machines are much more capable than people are
capable of applying them,
My own rule of thumb in the application o:f: machines
to non-numeric problems is this: expect half as much in
twice the time at twice the cost. If you get :it you can
count yourself lucky.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
If we begin with certainties,
we shall end in doubts; but
if we begin with doubts,
and we are patient with
them, we shall end in
certainties.
BACON
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/117ZT:'Cif'-'1~bp80B01139A000300110003-6
V, CHALLENGE OF ASSUMPTIONS
Having searched and studied all possible sources of
assistance which techniques, procedures, and machines can
offer, the most important job still remains undone, for
it is insufficient to improve present means of doing
business, or even to find superior means of supplanting
present approaches,
What we most need to do is to challenge our assump-
tions. Reason will prevail and progress be made only
when the assumptions by which we live are cultivated,
examined, destroyed, and replaced, Assumptions form the
skeleton of any information system. They determine its
operational role. For that reason we can never ask "why?"
too often. We are assisted in this challenge o.f assump-
tions by the variety of motives, good and bad, which
different members of any organization have. And also by
the variety of assumptions held about a given operation.
Those who have not viewed an operation favorably will
welcome a chance to bring about change. Their operations
may be enlarged ar improved in some way at the expense
of an operation they may view as archaic. Those who have
not known much about it may welcome a chance to broaden
their knowledge and experience, which in turn may lead
to more rapid personal advancement. Those w:ho have a
stake in the present operation may welcome a chance to
attract greater attention and support to it than it pre-
viously enjoyed, particularly should the validity of the
challenged assumption be confirmed in the mind of the
examiner as a consequence of his close scrutiny of it.
As we all know, in the impersonal catacombs of the
bureaucracy, outmoded notions and routines are likely to
be jealously guarded. This is human, and indeed manage-
ment tends to foster such attitudes. We build morale by
impressing upon subordinates the importance of diligent
discharge of particular tasks performed in particular
ways. An abrupt challenge which disturbs these patterns
can be as startling as a slap in the face. We need
nevertheless to provide for change. Genius lies in doing
so in a manner which builds rather than destroys morale,
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 200'f/T'f/T1 : ~TA~DP80B01139A000300110003-6
Here, as in other fields of human endeavor, a sense
of participation is the most promising approach. Making
the production analyst, for example, an active rather
than a passive participant in bringing about change is an
educational method of the greatest power. His interest
is aroused, he shares in the program's development, he
supports its implementation and operation.
The inquiring mind skeptically turned to a situation
of long standing can yield commendable and at times even
surprising results. We had, for example, a collection
program which by 1964 produced 40,000 reports annually.
Collecting, typing, editing, processing, recording, and
disseminating these sa preoccupied case officers that
minimal effort went into guiding the collectors.
An analysis of this situation disclosed many useful
and interesting things. However, the most significant
attack upon this particular citadel was to challenge the
assumptions that quantity of reporting was what was
wanted; that quantity of reporting best filled information
gaps; that quantity of reporting was the way far the post
and the collector to get the highest marks; that standard
written guidance sufficed as the medium of communication
between analyst and collector.
We set quality as the standard, brought collector
and user closer together, and relaxed the formalities far
reporting management-type information (e.g., statistical
data). The disabusing of falsely held notions and antic-
ipations had immediate consequences. In one year's time
volume was down 36%. In the second year volume dropped
12% from the new base. Thus .in two year's time the number
of reports decreased by 44%. This saved time and paper-
handling. More importantly it greatly improved quality
of reporting.
In another area, an assumption of long-standing has
been that given certain facilities designed to support
him, the production analyst will in fact use them. And
if he doesn't he should be instructed to do so for his
own good.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 200TfTT/22-LTARDP80B01139A000300110003-6
Here we are entering a complex area involving the
characteristic inertia in human nature, the futility of
legislating human conduct, and relative factors of con-
fidence of an analyst in his own file versus one
organized by others in his behalf. Don Swansor~~ has
suggested the application in such cases of the principle
of least action; namely, the design of any future informa-
tion service should be predicated on the assumption
(whether true or untrue) that its customers .will exert
minimal effort in order to receive its benefits, Further-
more, they will not bother at all if the necessary
minimum is higher than some daily low threshold,
In our planning, or in reviewing assumptions, there
are of course certain constants which have profound
implications. One is the rate at which infarmation
can be read by any one person, sometimes referred to as
"individual channel capacity." This taken together with
the fact that recorded knowledge accumulates through the
years means that each individual will see a decreasing
fraction of available information. If it i~~ necessary
to blanket the accumulation, then the response must be
increasing specialization. This leads to "twigging," the
phenomenon which occurs in the endless fractionation of
interest and knowledge in various fields. The twigs
remain the same size but the tree gets larger. In this
situation, can we safely assume, as we often have, that
adequate coverage of information can be provided by a
staff whose size is constant?
Take another situation. Production analysts often
request documents five, eight, or even over ten years
old. We have assumed that a request of these is an
indicator of value, and that we therefore dare not risk
purging old material from ever growing holdings.
Just as we have considered it unwise, as has many
times been suggested, to use senior production analysts
as a filter to thereby reduce appreciably the present
flow of information into the system. Should not this
assumption .too be challenged. indeed the assumption
that lots of paper is in itself bad, and that volume
reduction of any kind, is in itself good, needs to be
under constant scrutiny.
Approved For Release 2001/ ~~GC~A~,P80B01139A000300110003-6
 Approved For Release 200?T~/22` ETA=RDP80B01139A000300110003-6
Should we not be more daring in our search for the
real edges of our new world? We devote vast treasure to
R&D programs which promise gadgetry, accepting willingly
the risk of possible failure. Isn't it time we just as
willingly seek the promise of experimentation in the
softer information processing sciences. And in doing so
incur no greater opprobrium for failure in this field
than we incur in the development of hardware.
Recently, at a seminar on "The Computer and the Policy-
25X1A making Community," concluded that we had
perhaps been taught "...t a e vices of computers are
our own."25/ It is time we learn more about these vices,
and, as with intelligence information get these under
better control.
Approved For Release 2001/115/ECRE~80B01139A000300110003-6
 Approved For Release 2001/11/22: CIA-~2DP80B01'139A000300110003-6
Knowledge is of two kinds.
We know a subject ourselves
or we know where we can
find information about it,
SAMUEL JOHNSON
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 206'fM~1T22'L'1~RDP80B01139A0~~~~03A6
25X1A
REFERENCES
1 The guiding of intelligence
co ec to MC Proceedings, November 1958.
2 On processing intelligence
informa ion. IMC Proceedings, November 1958.
Advanced techniques for processing
. information. IMC Proceedings,
October~1962.
25X` 3 .
~..
25X1A
4 Introductory remarks to the
emons ra ion of machine translation. IMC
Proceedings, October 1962.
5. PRICE, DEREK JOHN DE SOLLA. Science since Babylon.
New Haven, Yale University Press, 1961. 149p.
6. BUSH, VANNEVAR. As we may think. In the Atlantic
Monthly p.101-108, July 1945.
7, THE NEW computerized age. In Saturday Review
p.15-37, 23 July 1966.
8. GOOD-BY to Gutenberg. In Newsweek p.85?-88,
24 January 1966.
9. Mr. McLuhan's book, UNDERSTANDING MEDIA--The
extensions of man, is of interest to intelli-
gence.
10. PLANNING CONFERENCE ON INFORMATION TRANSFER
EXPERIMENTS, Massachusetts Institute of
Technology, 1 65. INTREX: report of a
planning conference on information transfer
experiments, 3 September 1965. Edited by
Carl F. Overhage, and R. Joyce Harman.
Cambridge, Massachusetts, MIT Press, 1965.
276p.
11. MILLER, JAMES G. Interuniversity CommL~nications
Council, EDUCOM. Unpublished manuscript.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
-. SECRET
 Approved For Release 200'IT1'172-DTA~DP80B01139A000300110003-6
12. WASHINGTON STATE UNIVERSITY. Proposal for the
on-line automation of the Washington State
University Library, submitted...to ttie
National Science Foundation, Office c>f
Scientific Information Service. Pullman,
Washington, 1966. 26p.
13. Using computers in a new university
American Library Association
Bulletin :823-826, October 1965.
14. KING, GILBERT W., and others. Automation and the
Library of Congress. Washington, Library of
Congress, 1963. 88p.
15. LICKLIDER, J. C. R. Libraries of the future.
Cambridge, Massachusetts, MIT Press, 1965.
219p.
16. BOURNE, CHARLES P. Methods of information handling.
Wiley, 1963. 241p.
17. nd ROBERT M. HAYES. Information
retrieval. New York, J. Wiley,
1963. 448p.
18. GREENBERGER, MARTIN ed. Computers and the world
of the future. Cambridge, Massachusetts, MIT
Press, 1962. 340p.
19. FEIGENBAUM, EDWARD ALBERT, and JULIAN FELDMAN, eds.
Computers and thought . -l~ew York, McGraw-Hi-I~
1963. 535p.
20. BORKO, HAROLD, ed. Computer applications in the
behavioral sciences. Englewood Cli:Efs, New
Jersey, Prentice-Hall, 1962. 633p.
21. U.S. NATIONAL SCIENCE FOUNDATION. OFFICE OF SCIENCE
INFORMATION SERVICE. Current research and
development in scientific documentation. No. 13,
November 1964. Washington, 1964. 486p.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 200~~22 ~~CI~JA RDP80B01139A000300110003-6
22. CONFERENCE ON LIBRARIES AND AUTOMATION. Airlie
Foundation, 1963. Libraries and automation;
proceedings, a ited by Barbara Evans Markuson.
Washington, Library of Congress, 1969:. 268p.
23. Illustrative of the type of papers of which I
speak are the following:
ASHWORTH, WILFRED. Librarianship and other
disciplines. In ASLIB Proceedings
18:152-159, June 1966,
. BAR-HILLEL, YEHOSHUA. Is information retrieval
approaching a crisis? Washington, Office
of Naval Research, 1961. llp. (ASTIA
no.269881)
FIOCK, L. R., jr. Seven deadly sins in EDP.
In Harvar~Business Review p.88-96,
May/June 1962.
. FLOOD, MERRILL M. The systems approach to
library planning. In Library Quarterly
34:326-338, October~964.
HEILPRIN, LAURENCE B. On the information
problem ahead. In American Documentation
12:6-14, January~961,
KNOX, WILLIAM T. The new look in information
systems. Washington, Office of Science and
Technology, Executive Office of the President,
1966. 18p.
. OETTINGER, A. G. An essay in information retrieval,
or the birth of a myth. In Information and
Control 8:64-79, February~965.
SALTON, GERARD. Progress in automatic information
retrieval. Reprint from IEEE Spectrum
vol. 2, no.8:90-103, August 1965.
SHAPIRO, E. B. An evaluation of computers in
text editing. Menlo Park, California,
Stanford Research Institute, 1966. 31p.
(SRI project 5849)
Approved For Release 2001/1~~ ;,,~Q~p~80B01139A000300110003-6
 Approved For Release 200~TT/~ : ~TA~DP80B01139A000300110003-6
SIMONTON,_WESLEY C., ed. Information retrieval
today; papers presented at the institute
conducted by the Library School and the
Center for Continuation Study, 19-22 September
1962. Minneapolis, Center for Continuation
Study, University of Minnesota, 1963. 176p.
U,S, FEDERAL COUNCIL FOR SCIENCE AND TECHNOLOGY.
COMMITTEE ON SCIENTIFIC AND TECiiNICAL
INFORMATION. Recommendations for national
document handling systems in science and
technology. Washington, 1965. 20p.
Appendix A--A background study--vol. 1.
Washington, 1965, lv.(various paging)
. U.S, NATIONAL ACADEMY OF SCIENCES, NATIONAL
RESEARCH COUNCIL. AUTOMATIC LANGUAGE
PROCESSING ADVISORY COMMITTEE. Language
and machines: computers in translation
and linguistics. Washington. In preparation,
announced for publication in October 1966.
24. SWANSON, DON R. On improving communications among
scientists. In Library Quarterly 36:79-87,
April 1966.
25. SCHWARTZ, JUDAH L, Computers and the policymaking
community. Livermore, California, University
of California, Lawrence Radiation Laboratory,
1966. 19po (AEC contract no.W-7405-eng-48)
(UCRL-14887; TID-4500, UC-32)
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
,SECR E T
 Approved For Release 2001/11/22: CIA-RDP80B01139A0003001?0003-6
Nothing can come out of nothing,
nothing can go back to nothing.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 200'fl'I'I/2'T-CT~RDP80B01139A~$(~~~~(~00~-6
SELECTED COMPUTER APPLICATIONS
AND TECHNIQUES
Collection
The computer is now being used in a rudimentary
fashion in collection management and in determining
collection system performance. We see, for example, the
"bookkeeping" functions of a requirement registry per-
formed by the machine. Mare sophisticated applications
are on the horizon which involve the testing of hypotheses
which a manager of intelligence collection normally is
concerned with, i.e., given certain collection targets
and priorities and specific collection sources, how might
these collection resources be allocated against these
targets? At. this point, however, the computer is not
used as an integral part of this optimizing process;
rather, it provides an answer to a specific computational
translation of a collection hypothesis.
Data Reduction
A broad view here requires a look at areas which we
consider normally under data reduction--signal analysis
and the like--as well as other activities which not only
filter data, but present them in a form that a "lay" analyst
can use effectively. The computer has played an important
role in the former area for quite some time. We see,
however, demands for more sophisticated processing of
higher volumes of raw data--better precision in wave form
analysis, ability to perform more sophisticated spectral
analysis and correlation, a wider variety of emitters to
identify, and so on. In the latter area, we should con-
sider topics such as automatic language translation, auto-
matic filtering of redundant textual material, automatic
dissemination, and so on. We see little operational
impact of computers in these areas aside from long-term
experiments with limited chance of payoff.
Retrieval and Analysis
Again, we can cite two subcategories: computational
and non-computational data processing; and, as before,
progress in the former is considerably advanced over the
latter (this is independent of any value judgments on the
utility of the product involved). For example, the
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/1-1-/2~ Z:IA=Rd`P80B01139A000300110003-6
computer has been used very effectively in filling in the
complete picture of a missile or space event with only
fragmentary data available. Secondly, where a particular
activity of intelligence interest can be reduced to a
mathematical model, some successes have been achieved--one
example here is our ability to analyze a particular air
defense system.
In the area of non-computational data processing,
the computer has been used principally as a file updating,
sorting, searching, and printing device. Examples here
include files on aircraft movements, personnel travel,
facility status, and so on.
Some success has been achieved in assisting economic
intelligence analysis, which is a combination of both
types of data processing. Some of tYie first Agency appli-
cations of computers to intelligence were in this area
and continue to be of some importance--computation of
Soviet military expenditures and of Soviet gross national
product.
The process of providing intelligence material in
a form suitable for wide dissemination involves consid-
erable expense and in some areas is amenable to computer
assistance. The two particular examples are cited here.
The production of high-quality, printed material is now
being assisted through 'the use of a computer-generated
tape which is fed to an automatic photo-composing device,
New editions of the. National Intelligence Survey will be
printed using this technique without sacrificing the
aesthetic qualities of this publication (the use of
multiple fonts and other complexities in print composition),
Secondly, the computer is now being used on a.n experimental
basis to provide assistance to cartographers. It is
intended that a large bank of cartographic information be
machine stored and selectively retrieved and plotted to
provide the basic map over which specific intelligence
information would be superimposed. Considerable flexi-
bility would be available in terms of scale, projection,
degree of detail, and number of geographic features
shown on the map.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/1 'T/2~ ~~80B01139A000300110003-6
A basic criterion that might be applied to determine
if the production analyst is making effective use of the
computer is, "Would there be any disastrous effects if the
computer were suddenly withheld from any of these appli-
cation areas?" A candid answer in most cases, unfortunately,
would be no. A basic reason for this is that the analyst
has maintained separate backup facilities in many cases--
perhaps because of his reluctance to rely so heavily on
the computer. Secondly, several of these applications
are in the "why don't we try using a computer" category--
without appropriate analysis of the costs, effort, and
ultimate payoffs involved. There is some cause for
dptimism, however. Notable successes have been achieved
in cases where (1) the analyst initiates the requirement,
(2) he understands the capa i ities as well as the
limitations of the computer, and (3) he permits the com-
puter specialist to get deeply involved in the early stages
of problem analysis.
Input
Considerable attention is now being given to the
concept of "source data automation." This implies the
keying of data as close to the source as possible. In
same circumstances it implies direct entry of these data
into the machine. A technique of interest here is
optical character recognition, where the emphasis has
recently shifted from an interest in reading text to the
use of character readers in the place of punching paper
tape or cards. Where reasonable quality control can be
applied to the typewriter font, paper contrast, and
character registration, available print readers are now
competitive as an input device with punch card or punch
paper tape readers. This input preparation technique has
several obvious advantages; the use of standard office
equipment for keying, decentralization of input preparation,
the ability of the human to read exactly what the machine
reads, ability to implement a "turn-around document" con-
cept, i.e., computer printouts can be modified and then
fed directly back as input.
Progress is also being made in design of. economical
secure tape typewriters, In the next few years, we expect
to see such devices available for under $5,000. It is
also reasonable to expect that the cathode ray tube or
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 20011'f'fT2~Z:T1~~P80B01139A000300110003-6
typewriter terminal will be used as a direct data entry
device for large volume inputs. This would provide a
facility for direct feedback of detected error conditions
which could be serviced on the spot.
Processing Units
The basic design of computers has remained unchanged
since the invention of the stored program concept over 15
years ago. It was designed for arithmetic operations and
as such is not particularly appropriate to the solution
of non-numeric data processing problems. However, attempts
to design a machine which better fits these problems have
not been very successful. For example, parallel search
memories where all cells are interrogated simultaneously
have not progressed beyond the laboratory. Attempts have
also been made to build a machine which processes materials
sorted in something called an "information processing
language" (IPL), where data elements are stored in a tree
structure and search progresses up and .down branches of
the structure. Another example of a special-purpose
machine is one currently installed in the CIA Computer
Center called the Automatic Language Processor, which
employs a large-capacity disk called a "pho~tostore" which
is used as a dictionary for lexical processing. Simula-
tions of all of these special machines have indicated that
the general-purpose device which is now available is an
effective and relatively inexpensive--if unsophisticated--
facility for the solution of data processing problems.
The principal reason for this is that the market for the
general-purpose computer is extremely large and competitive,
and the cost for a given amount of processing continues
to decrease.
Files and Storage
The comments made above are applicable here as well.
That is, storage devices normally made available with
general-purpose machines cannot be considered optimal for
most intelligence data processing problems. But the
capacity and speed of these devices makes them economically
feasible. While direct access devices such as storage
drums and disks have been available for several years, our
ability to use them effectively is still behind the
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 2001/T11Z"2~~TA~DP80B01139A000300110003-6
electromechanical technology, For example, we find that
the various segments of disk storage are really consid-
ered to be small magnetic tapes--data within each segment
are usually searched serially.
This term was coined about four years ago to identify
a concept which implies that many of the facilities that
must be available to users of computing equipment are
appropriately implemented through a set of programs rather
than through specific hardware logical features. For
example, while a computer contains a control unit which
automatically determines the sequence of its operations,
more sophisticated automatic control features, such as
automatic scheduling of major jobs on the computer, have
been assigned to programmers for implementation. Similarly,
more sophisticated languages for commanding the machine--
other than the basic instruction set which includes functions
such as add, subtract, move, and so on--have obvious
advantages to the user. These again have been implemented
through programs which translate these command languages
into the elemental instruction set available directly to
the hardware.
This concept has mushroomed to the point where buyers
of computer systems demand a wide variety of control and
utility programs as an integral part of the ;package to be
rented or bought from the manufacturer. The software con-
cept, along with the increased capacity and speed of com-
puting devices, has resulted in the notion o:f the computer
as an information processing utility. As a :result, we
hear the term "time-sharing" quite often when we talk to
computer people these days. In this kind of an environment,
the computer user sitting at the console has the illusion
of the complete computer being made available to him
continuously. In reality, however, he is only given short
bursts of the time available on a computer. These time
slices might be several milliseconds for each of perhaps
100 users. This area is similar to the method of inter-
leaving telephone conversations in a modern communication
system. If this concept is workable--and experiments thus
far show promise--the computer user, whether he be a
programmer or intelligence analyst, will have more direct
access to computer facilities.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 Approved For Release 200'I"T/ZZ-LTA~DP80B01139A000300110003-6
Analog Devices
The digital computer has for quite some time
surpassed the analog computer in its speed and precision.
However, an increasing amount of data which must be
processed in an intelligence environment is collected
in analog form. If the digital computer is to be used
for the analysis of these raw data, more efficient means
of converting the analog signal to digital form must be
available. The principal limitations of conventional
converters are speed, and quantitization precision. In
addition, it would be desirable if the analog portion of
such a converter would have feedback facilities for auto-
matic retuning of the device to enhance the quality of
the signal being processed. Converters which incorporate
some of these advantages and overcome some of the speed
and precision limitations previously encountered have been
designed and will be used soon experimentally. The results
of these experiments should be followed closely because of
the vital role that analog collection systems promise to
play in intelligence in the future.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECR ~ T
 Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
In the library of the future, man
will continue to read books, gain
insights, think, and make discoveries.
But the library will do ,most of the
searching, transforming, interpreting,
and checking of information he need's,
and thereby will free him for more
creative uses of stored information.
BOLT BERANEK & NEWMAN
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
 Approved For Release 2001/1"1722C~~RDP80B01139A000~(~~~$~6C
CIA's information processing problems are being
addressed by a combined team consisting of representatives
from the Offices of Central Reference and Computer Services
on a project basis (Project CHIVE).
This project is a two-pronged effort to achieve an
improved capability to store and retrieve intelligence,
through the combination of machine and human techniques
for managing and controlling massive volumes of informa-
tion. First, a new concept of Agency-wide information
service is being developed with the emphasi;~ placed on a
fast integrated response to analyst queries using a reason-
able balance between the intellectual capabilities of
intelligence analysts and the data storage, selection,
correlation, and display capabilities of electronic equip-
ment. The data base for the all-source system will be
comprehensive, concentrating on named object or "hard"
intelligence facts which will be used for document
retrieval, information file building, and (eventually)
for automatic correlation--or the retrieval of data
through inferential logic.
From the user point-of-view, the CHIVE system will
have unique characteristics which should provide a number
of advantages. Basic features of the system are:
. All-Source Retrieval--Establishment of an inte-
grated, all-source file system covering every
type of printed document of use to the Agency,
including maps and photos, at whatever classi-
fication level. Here, the significance to the
user would be that he would be able to get a
total system response, assured that all avail-
able sources have been tapped and their relation-
ships have been exploited,
Single-Point Service--Reorganization of the
gency s centra information retrieval service
under a radically different management umbrella---
that is, by geographic area--with each geographic
element exploiting all available source material
and using common indexing and procedural
standards. As a result, the user would need to
contact only a few people to bring the full power
of the system to bear on his problem.
Approved For Release 2001/1~,~~C~q;~~Qp80B01139A000300110003-6
 Approved For Release 2001/'f~/~2~ClA~DP80B01139A000300110003-6
Literature Searching--Fulfillment of the basic
unc ion o ,serving as the Agency's institutional
memory by responding to such requests as: "Give
me all books, documents, reports, etc., that
concern Soviet anti-ballistic missile research."
The response in this case would be relevant
documents, not the answers themselves.
Retrieval of Basic Facts--Answers to specific
questions derived rom the accumulation of data
stored. The latter will include a host of facts
concerning foreign personalities of all kinds,
organizations (political, cu tural, scientific,
educational, etc), installations and other fixed
facilities, and activities and events (:For
example, trade, subversion, miss-~fir:ings, etc.) .
Thus, the user will be able to ask, and have
answered inside of 15 minutes, such questions as:
"Who is the head of the Soviet Rocket Forces?"
"When did Mao Tse Tung last appear in public?"
"What operational ICBM sites have located in
the Baku area?"
Counts, Correlations, and Trend Analysis--Answers
to questions, oth or management control as well
as substantive intelligence purposes, such as:
"How many papers on solid fuel rockets were
published by the Chinese in 1965?"
"How many CIA intelligence reports wE~re
issued on Haiti in July 1966?"
Given the ability to make counts, we can also
observe how such counts change as a function of
time and thus estimate trends. For example, we
might observe that there is a marked deviation
from the norm in the number of appearances of
Chinese leaders, and so forth.
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECR E T
 Approved For Release 2001/11/22 C~~6P80B01139A000300110003-6
. Detection of Redundancies
and Inconsistencies--
en i ica ion o up ica ive in orma ion in the
system store as well as apparent contradictions
which might suggest follow-on research to be
undertaken by intelligence analysts.
Automatic Inference-Making--Assuming we can devise
a system that can s ore a wide variety of facts
about people, institutions, and activities, and
can then "operate" on these data, it should be
possible to infer information that man x has
worked on thermonuclear fusion, and if man x is
currently living in city y, we might infer a
certain probability that thermonuclear research
is being conducted in city y. The variety of
problems to which such a man-machine capability
might be applied is largely a function of the
imagination and ingenuity of the intelligence
analyst.
Machine-Assisted Language Translation--More com-
prehensive and rapi exp oitation of the totality
of information received in foreign language
documents through (a) actual machine translation
of Russian language documents and (b) a machine
system for rapidly converting oral translations
of other languages, via the Stenographic process,
to printed documents in English.
Analyst Referral Service--Establishment of a
centra directory or profile" of human and other
information resources scattered about the Agency
possessing expert knowledge in specialized areas.
Such a tool would be used to supplement the CHIVE
system by identi:~ying analysts as well as intelli-
gence problems.
Micro-Image Document Sto.e--Installation of a
system or storing t e vast quantity of Agency
documentary receipts (up to 1.5 million documents/
yr.) in a miniaturized form thus reducing space
requirements and providing a random-access, rapid
retrieval and reporduction capability. While
initially the storage medium will be aperture
cards, it is planned that the system will be con-
verted by 1970 to a large-capacity, document
storage device on the order of a second-generation
"Walnut."
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECR E T
 Approved For Release 2001/'I'f/Z2CCIA~dP80B01139A000300110003-6
Remote Querying--Provision of remote communi-
cation capabilities, coupled with a :Flexible
file structure and machine command language,
which will enable users to interrogate and
maintain data files directly and on an up-to-
date basis. Special-purpose files maintained
by analysts could eventually be included in
the central system. Ultimately, we would hope
the system would lighten the individual analyst's
burden in file housekeeping and serve as a more
effective extension of his memory.
We see the development of an improved information
storage and retrieval capability as extending over two
five-year phases. During the first five years, we will
concentrate on implementing the full CHIVE system in
house through a planned evolution by geographic area,
beginning with Communist China. During the next five
years, the goal will be to consolidate the CHIVE system
with other specialized Agency information systems with
the idea of achieving an integrated Agency information
network. In parallel during the second phase, we expect
to achieve integration of Inter-Agency EDP information
systems and to provide a world-wide communications net-
work in support of U.S. intelligence activities and
operations.
Illustrations
Current System
CHIVE System
CHIVE Functional Overview
CHIVE Information Language
Approved For Release 2001/11/22: CIA-RDP80B01139A000300110003-6
SECRET
 CURRENT SYSTEM
Data Base
Documents
ved Fo~ Relea~e 2001/1/22 : ~IA-RDP~0B011~,9A0003~01100~-6
File Structure
Indexes
Books
Intel. Reports
Open Lit.
Cables
Special Intef.
PI Reports
Ground/Person
Photos
Films/Video
SR
Book Catalog
Tab Cd. Index
Tab Ctl. Index
Tab Ctl. Index
Tab Cd. Index
Tab Cd. Index
Manual Index
Document Files
Book Collection
Ap. Cd. Fi I e
Dossiers
5 x 8 Cds.
Dossiers
5 x 8 Cds.
Ap. Ctl. Files
HC File
Fi Em/Photo
Collection
Querying
Maps Map Library Maps Catalog HC Collection Map Library
Ap. Cd. File
Multiple {Vocabularies Multiple indexes Multiple
procedures document stares query points
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 CHIVE SYSTEM
Data Base
Appro~ed For ~2elease~2001/1'122 : CI~-RDP8gB01139r400030~110003(6
Gooks
Intelligence Reports
Open Literature
Cables
Translations
Special Intelligence
PI Reports
Ground/Personality
Photos
Films%Video
Library
File Structure
Indexes
Book
Catalog
I~ap
Library
o Common o Vocabularies
Procedures
~-- Library
Document Files
Book Collection
Document
Image File
Film/Photo
Collection
Map
Cata[ag
Map Collection
Ap. Card Fi I e
o Master Indexes,
Document Stores
o Few
Entry
Points
Querying
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Appr~ved Fo~ Relea~ 2001/1/22 : C,~A-RDP~0B011~A0003Q01100Q,3-6
RESEARCH
OFFICES
i i
~ EDP INPUT PAGE ~
~ FILES' PROCESSING READER ~
i ~
----~
' GEOGRAPHIC DESKS '
ALL
SOURCE
DOCUMENTS
HEADER
INDEXERS
Title Expansion
Recording of
Bibliographic
Data
INFORMATION
ANALYSTS
Selection
Summary Files
Retrieval
CONTENT
INDEXERS
Indexing
Dictionary
_ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~1
MASTER
IMAGE;'
R
E
SING
" FILES
I
IMAGE
P
OC
S
DATA
TRANSCRIBERS
Typing of
page reader
form
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Approved Fo~Relea~ 2001/1/22 : ~A-RDP$0B011~A0003~01100~-6
? Doc. Ident. No.
? Source
? Security Classification
? Title
? Date of Information
"'Soft" Subjects/Concepts
? Plans
? Trade
? Propaganda
? Scientific
? Research
Persons
? Name
? Location Data
? Birth and Ceath
? Affiliations
- Org. Name
Position Title
? Travel
? Education
? Occupation
? Etc.
? Political Attitudes
? Cultural Movements
? Education
? Commodities
"'Hard?' Subjects/Named Objects
Organizations/Foci I ities
? Name
- Preferred Named
- Ident. No.
? Location Data
? Function
? Subordination
? Products
? Associated Persons
- Name
Position
? Etc.
? Communications
? Transportation
? Finance
? Agriculture
Locations
? Name
? Country
? Geocoordinates
? Telephone Number
? Street Address
Approved For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
 Ir~'i~`For Release 2001/11/22 :CIA-RDP80B01139A000300110003-6
-~