PROJECT TREASURE ISLANLD EXPLOITATION OF OPEN SOURCE SLAVIC PUBLICATIONS

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP81-00706R000100230035-8
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 25
• Document Creation Date: December 23, 2016
• Document Release Date: December 11, 2013
• Sequence Number: 35
• Publication Date: October 12, 1948
• Content Type: MEMO

Pw- elir-roi-A- 11-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
IA-R D P81-00706 R000100230035-8 j MP EteelitiVe Registry
es- /3/f
OCT 1 2 1948
itSgOitati.alg FOR: Director of InteLligence
Headquarters, ',Tilted States Air Force
SUB.h;CT: ereject i'ressere eriend (Fxpletteelen er leen
Source elavic eublications)
1. Refereace your eemorendum, saae ueject, dated 20 September,
the Treasure Island project has been reviewed as reveeted.
2. Althotteh your three quetitio.ls are concise ane logically
acLiressai to tele Agency, the ?mewled plan of operation, co contained
In your enclosure, isnot sufficiently specific to eereit either erief
or categorizal replies to your curies. The comments attached as Tab A
point out certain ineoneistencies in the project, ane sone or the
prOblOMS WhiCh Maild arLOO if the plan were implemented in its present
form.
e. The following provitional answers are submitted for your
consideration, subject to the qualifications of paragraph 2, above.
a) "Will the Air Force be duplicating similar projeete
being conducted by other agencies?"
If the scope of ehe project is as extensive as paraeraph
4r, page 10 of the brier would indtcate, the enemy is in the
affirmative. Ir the definition of the pertinent iftformation to
be extracted is limited, as etated in paragraph 44 of the brief,
to "the fields of economic and iudustrial informaeion related to
strategic veleerability, technioal air iftformetion and air arm
and civil air information, biographical air irforustion and air
geography, *the answer is still in Ube affirmativo eince Foreign
Uocueents branch, CIA, is ereeently extracting irom current
Russian neweeapery and periodicals information-wherh-falls into
these categorise.
If it were presumed that the 7reasere leleed PrOect
would, on the other hand, make use of roroign Pocuments ':e.anon
erase ?attracts and the Joint Press eeadieg Service Presa Summary
iiriRpIcifation of the Foeiet prees, althouge not under Alr
igenog control, then duplication would still exist in the proposal
(paragraph 5) that the Army Library handle current periodicals and
that the Library of ecngress handle "government periodicals."
Inasmuch as all Soviet periodicals are eovernmant periodicals, the
distinction is somewhat illoeical. eevertheless, since Foreign
Doc.attents Branch is at present abstracting all periodicals that it
is possible to get from the rasa, duplicatioa would necesearily
result.
Declassified in Part - Sanitized Copy Approved for Release a 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 LILU 1L
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8) IIP
Although the extent of air geogriehe ifs not clearly defined,
it is believed that in the exploitation of balite information on
this subjectthere would be some duplication vith the work done
tYY the Array fte Service.
rt in possible that processing oe the baoklee =telt delicate
some of the work done oy the Inter-departmental Committee for the
keluisittnn of 'oreign euhlicatione end the tore. en economic Admin?
istration during the ear, and that ehionLems done ee the Liffice of
tntereetional Trade, Deddrtment of Commerc-, and the Intellieence
Devi/Lon, V006, especially am contained in the latterve war-time
publication entitled eInformetion from the europeen eress.e German
material, already processed kV CIA, also conteins the results of
exploitation of etwever b]vie.publications and captured doeumonts
for ineustrial and SCOnOlda inforsevion.
einally, it is Assumed that tne proposed project would be
coordinated wiun erojtot 307-t ef the esrausiot1cs1 Churt ervics
nad with the extousive abetraetine et fersien ate techeical
puolications as carried un by e-2, AMC (Wright )ield), to avoid
duplication within the Department of the Air Poses.
erfth the Air Force be operatine in its field of dominant
interest?"
In *newer to this, it muet be stAed that the exploitatioa of
Glen seurcemotertel Involving tile eeleetion and extraction of zilch
beeio informaLion as exists inehreign publLeeiene.eannot, with
certain to.naisal exceptione, be said to lie within eey ono field
of dominent interest. A pub1iano4 1teaon tin, construction of new
sten factory in the ussit in of Interest to C/A and other intell-
igence ageecies as well as to the Air Force. Deminent or deoert-
tentml interest lies in the interpretation, analysis and use of
the raw informat4on. Aecordingly, it o*i. only be stated that by
itr very nature, this tyee of operation must to of cencerm to all
the IAC seencies.
o) VA final queetien asks: wWill such a. project lead to
unheeltey coapetition for trenslatorsiet This is gleaned by a state.
sent, in pareeraph 3 that only lo* order non-technIcs1 translators
are envisioned.
tn ecnnection wite the latter, a queatien ariees as to how
technical. air xnformatIon (paragraph los of tee brief) will be ex-
tradited by non-teohnical traaelators. It is believed that experience
will show theL tee *Aployment oi really low order tronelatore is
inefficient, time.constaaa, QZ3 costly. This has teen recoenised
in the preperetion or the tentative T/0 (Annex /) mach cella for
22 P-40e, 14 P-ale, etc. Trailalators of such grades are considered
by CIA and the Civial eervice Commies-tor: to be of extremely high
calibre. Thus, the administration of the proposed project would be
recruitieg the best-trained lineuiste it could find, which then put
-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11 :
CIA-RDP81-00706R000100230035-8 unkipvio Mu. n
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
in the field another competitor for experienced and qualified
linguists and translators. Experience has shown that for every
low grade translator (1.e.? CAP-5), one qualified and experienced
tranelaeor is needed to check and correct the work, and in the end
less time wonlA ec spent if medium grade translators were hired.
The answer is therefore mainly affirmative, in that *competition
for translators" would result. nether or not this comeetition
would be "unhealthy" would neeend on whether a coordinated recreit.
:sent plan could be worked out with other agencies.
It is believed that the competition for translators might be
less serious than the competition for sources and source materials,
La view of the difficulty of obtaining even one copy of many current
USSR publications in Moscow.
4. The opeortunity to comment on the proposed project is appreciated
by this Agency, and if you should desire further discussion or clarification
of any af the points sent_ened above, a conference between the appropriate
branch chiefs of the CIA end the Directorate of Intelligence can be
arraneed at your conveeience.
BRIGADIER GENERAL. USA
4444444. DIRECTOR
LAD:abc
Di04464140n
cc: Signori:: Copy
Central Records
00 Please return to 2 A 116,1ards te Docks
FDB
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 t
50X1
_
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11 :
CIA-RDP81-00706R000100230035-8
Director of Central Intelligence
Assistant Director for Operations
Air Force Project "Treasure Island" (Exploitation of Open Source
Slavic Publications)
1. The "Treasure Island" project, as proposed by the Department of
the Air Force and forwarded by memo of the Director of Intee3igeoce, USAF,
dated 20 September 19, 48 has been studied by this Office and Foreign Docu-
ments Breneb.
Although the memorandum of transmittal limits the project to fit
program to extract basic air intelligence information from Slavic publications
in the Library of Congress*, and proposes "not a super extracting agency,
but a modest effort in the air intelligence field", the Treasure Island brief
discusses the larger problems of coordinating the proposed effort with other
government agencies which are active in the field of processing foreign
publications for intelligence information. It also suggests an allocation
of sources and division of labor between the Library of Congress, the Army
Library, and CIA, and mentions other organisations, such as the Aeronautical
Chart Service and the Britith JIB as possible participants in the over-all
program. It recommends the etahlishment of a Basic Information Section
in the Directorate of Intellieence USAF, to supervise a service for the ex-
traction of information from Slavic open source materials, authorised to
negotiate a contract eith the Library of Congress for the processing of
current and backlog material. In addition it states: "this administration*
to be located in the library of Congress, would control procurement of
materials, maintain Liaison with the Acquisitions bivisioc in ie Library,
keep a central file, and publish the extracts. The administration would
control the various extracting units, lo-ated in the Library of Coneress
and in other government agencies. The work would most advantageously be
divided op asach a manner that each participating 44e.ncy would handle
One maaor category of open source 'materials."
It is felt that the functions of coordinating and controlling the
various abstracting and publications procurement programs already under-
taken by the IAC agencies and other government departments cannot properly
be assumed by a new and see:ordinate unit of the Department of the Air Force.
The copplex problems of inter-agency coordinations control of procurements
and allocation of effort in regard to open source material require planning
and action on a higher level. Under the provisions of Section 102 (d) 2
of the National Security Act and NSCID 3, Para, 1 a (Basic Intelligence)
such action could more appropriately be taken by CIA.
2. Considered from the point of view of an operating unit (omitting
the function of coordination), the mission of "the Treasure Island" project
is not Clearly defined. It discusses air intelligence requirements (pare
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
_
Declassified in Part- Sanitized Cop?y Approved for Release 50-Yr2013/12/11 :
"
CIA-RDP81-00706R000100230035-8
--.- it
2
4e, p. 10) and yet mention the ailvisability of indexing all information in
order to achieve cLIniacto coveraiie st. that this would be available to another
agency wthieh might wish to participate in the program in the future (pare 4f,
p. 10). It suggests an oilAnization of 76 people to accomplish this uncertain
mission within five years, without relating the proposed workload to the
figures quoted for the present backlog of publications and th,. current and
Mere accessions of Slavic matlrial. It does not define certain tares which
it employs, although any estimate of workload dependa upon the accurate defi...
nitions of each terms. For /wimple, it uses *abstracting* and *extracting*
apparently interchangeably and without differentiation. It also speaks of
Slavic publications of ail kinds without olariiying whether this includes
material iroo al Satellite countries or Slavic countrioa only (omitting
Btak:ary end Finlane). It farther mention intelligence exploitation throutths.
out, yet would place the major portion of the operation under contract with
the Library of Congress, a non-intelligence agency and the second largest
share of the operation under the Army Library, also a non-intelligence unit.
? 3. Accordingly, before an eetimate of the work load can be made, the
following premises must be established in line with the general tone of that
accomrlishment is Oesiredt
a. That all intelligence material will be extracted except political
and sociological..
b. That extracting or abstract.tng as used in thc Audi means the
reporting in ieiglish or all information of intelligence vAue appearing
in all availaLle Slavic lancuage publications.
c. That Slavic publications are presumed to include only Polish,
Ukrainian, Bulgerian, Serbian, Croatian, Slovenian, Slovak and Bdhemian
in addition to Ritalialls and will exclude publications in the some one
hundred. and tea (110) nowatilavic languages used in the MM.
4. That all information of intelligence value will be indexed,
orosaisindeaced and filed.
e. That backlog figures as given in the study aro accepted.
f. That all information extracted will be published and disas
nominated.
g. That uncleared personnel would be authorised for use in the
operation.
4. Based on the above premises, the following figures can than be com-
puted for the involved,
a. Baeklog plank
(2) Newspapers (100,000). Experience has shown that the average
extractor can cover five Slavic language newspapers per day. There..
fore, for 10Q300 newspapers over a five-year period 17 people would
be required.
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
-;IA-RDP81-00706R000100230035-8 .?_
_ 4?
Declassified in Part - Sanitized Copy Approved for Release 6-50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 r
- 3 ?
(2) periodicals (30,000). The average Slavic periodical
is estimeeed to contain 80 pages. To read and extract the in-
telligens* information from the average periodical *old require
eight days on the Weis that the extra:tort:an cover 10 peps a
day. To: the backlog of 30,000 periodicals than 200 extractors
mould be required.
(3) Books and othermblications (20,000). Meperienes has
shown that prObtay only _101: or 2,006 of this backlog would be of
intelligence interest. Since the average book mast be considered
to have 250 pages and for the books selected approximately 50% on
the average is worth extrecting, the average linguist should
savor a book in 25 deo. Therefore, h4 translators mould be
needed for this phase.
b. Current publications
(1) leespapersi For an estimated 80 aliment Slavic news..
papers on the basis of five newspapers per tur per person, 18
extractors would be required.
(2) Periodicals: ale estimated that at least 140 Russian
and 100 Giber Slavic 'language periodicals would be processed of
'high 25% mould be a-monthly. On the sane basis as used inies(2)
above 76 linguists would be required,
(3) nooks and other publication's An estimate of current
available Sfavfo Looks and publfostions of intelligence interest
would plum the Liguria at 75 a month. On the basis used in 10(3)
above this phase would require 85 people.
44 For the other phases of the operation, experience nae sheen
that the Winging numbers of personnel would be required. Supervision
and direction $9 intelligence *Moors. Rough..draft typing 50
typists. Screening analysts to sort all publications - 20. Sditeriel
and reproduetion 104. Indexing, orouwindexing, filing and searehi.
ing - 80. Administration - including personnel, supply and services
and correspondence files 25.
6. In the fallowing figures
show that a total of
summation,
770 people would be required*
Linguist extractors
440
Supervisory
59
Bditorial
104
Typists
SO
Files
80
Warners
20
Administrative
778
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 r c ii
AFOIR-CO-6
AEMORANDUM
SU3TeCT:
HEADkOARTERS UNTIED STATES AIRORCE
4AS.qIN3TON 20 Sep 1948
FOR 2.HE DIRECTOR OF CENTRAL INTELLIIENCE
Project Treasure Island (Explitation of Open Source
51avic Publications)
1. Considerable preliminary discussion ha a taken place in
connection with a program designed to extract basic air intelligence
information from Slavic publications in the Library of Congress. The
Air P'orce has assigned the name "Treasure Island" to this project.
2. Inforual discussions have given rise to the following
questions:
a. Will the Air i-lorce be duplicating similar projects being
conducted by other agencies?
b. Will the Air Force be operating in its field of dominant
interest?
c. 4111 such a project lead to unhealthy coelpetition for
translators?
Z. In view of the fact that this project is designed for the
procurement of air inforfaation it is believed that the answer to 2 b.
above is in the affirmative. However, it is considered that the Central
Intelligence Agency is the proper medium through which to arrive at
answers 2 a and 2 c. In this connection, only low order, nontechnical
translators are envisioned. This nucleus of translators would tend to
broaden rather than narrow the field. Yoreover, the intent is to sat up
not a super extracting agency but a modest effort in the air intelligence
fle1d.
4. A brief of the proposed project is inclosed. It Is requested
that such comment be made, as you consider appropriate, in connection
with paragraph 2 above.
1 Incl
Treasure Island Brief
C. P. CABELL
Yajor Jeneral, USAF
Director of Intellience, Office of
Deputy Chief of staff, Operations
_
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11 :
CIA-RDP81-00706R000100230035-8
7IDeclassified in Part- Sanitized Copy Approved for Relea-ser?d 50-Yr 2013/12/11 :
A-RDP81-00706R000100230035-8 r
ie il ol?
5. It is apparent from the above cesenstatione that the 50 abstracters
proposed for the projest (para Ise) or the tentative T/O of 76 persons
(Annex I) would be inatiesteate for the task. The 50 abstrastors egad
seareely do more au sake bibliographies and translate headlines old
tables of contents in five years, and to upset them to handle ourrent
books sod government, reports sinultanseuely would be unreallatie.
6. It nut be coneluded that the projost would be limited to *the
fields of erronsate and industrial information related to strategic valnerer
batty, technical sir Information? air ens and civil air Informations bie.
graphleal air infuriation, and air geography,* although the next paragraph
(4f, p.10) inetteates a asch wider scope of subject natter. With regard to
the statement, It would be a good walling point to have all the iatormation
indexed and crose-indexad se that extrasts of any kind can be made, two
quartieu arises
a. Does this OM that information would be indexed previous
to the extracting, is., using linguists for indexers?
b. Does *all the informations neon
ill all information from the limited eategories above?
2 all information of intelligens* wane?
3 all information?
7. The fellordng minor points in the Treasure Inland study seem worth
mentioning as sontradietory, lasseelarate, or open to disagreements
a. Pare Pls, (P. _6) mantiens the leek if flow of informatien
al *pea sours, but pant to ()), p. 6, OINSPISASS of the volume of
the loseaseerdinated digests* at present issued by various agensiee.
b. Pe_piLN..21,1_2.0_ states that Most of the estrasts are
astually? piasso Mow with air intelligence require-
ments. Many digests, espesially the Daily Reports en Foreign Mite
*readouts, attempt to reproduce current nen and promos.** but
eensistemtly calit intelligence data.* aenitored industrial and
workmate information is invariably insluded FRU in this publi?
cation within the limits of space and with regard to the relative
Imports's** of the information. FBIB is at present providing all
mdisearde* containing such material not published in the Doily
Report direst to the Strategic Vulnerability Branch. The Foreign
Domments Branch, 0Ia, *doh abstracts foreign periodicals and
newspapers en ropiest, his not reseivod any solution direetives
free the strategies Vulnerability Branch stating air intelligens)*
requirements in this field.
0. Para 20014 P. 6) due met take into asennt the resent
astivities cg Foredo Doeumente Branch, CIA, with regard to books,
nem etc.
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
5
ut? Pare ,414(P. 9) is imeorrest In stating that the Library
of Congress iithe 'government's purchasing agent for fornypt publio
aldose*, sines prosuretent of suet, material abroad is handled by
the State Departmont.
*Ai Pea WA fr. 9). The method of extracting by entering the
istormation on u&ouh shoots with s limited number of pertinent as
times is thought pratioable only for target data and by no meal
applicable to the wider field of istelligense. At least 3/14 of the
Wort used in exploiting theme decussate would be wasted by this
method. A ate* strew of information reports flowing into the
hands of wellsqualified anskyste as the result of definite anew.
ties directives is believed to be the met effigies% method of
hendling foreign published material,
8. Conclusions t
a. The necessity for adequate intelligelos expleitottion of
foreign publications to nest the specialised requirements of the
Dugouts's% of the AV Fero*, and 11111p01114.4 those of the Strategis
Vulnerability Branch, is dinonetratod by the isTreaunuse Islands paper.
It ie evident that the situation has hitherto been unsatisfaetory.
If snob requirements had, however, been submitted to CIA end the
desired fors indieatedit it is possible that the oritieism isplied
in pars 2b and a of the brief would sot be valid.
b. The study attempts to find a relation which 101) ovoids the
setting-op of one Andependat sup. ting agency; 2) whines a
meximon of **operation from sedating agensias 3) provides facilities
for the sarryinvent of mwoh extresting sonless as are necessary if
a somplato average of open scum materials is to be ashieved0
e. It seemess the function of interagency coordinatien and
postulates a preliminary survey of Boum" and sours* materials ',blob
suet be made before a satisfactory plan for division of labor between
participants can be evolved. It states 'Silk a prelladmary job mast
seeds be done by the Lilting of Congress whieh would have to be given
a contrast in order to enable thee to eceplete their index of Slavic
materials and to as fell use of the *lion Catalogue*.
d. It confuses the results to be produced by this prelimimary
sump' group in the Library of Camgrese with the body of extracted
basis air intelligens* information required by the Strategis Valuers.
laility Breaseh, end with the solution of the problem as indieated in
Para Lle above..
e. In view of the backlog figures and workl.oad ealoulations
outlined above, it is believed that the proposed Table of Organisation
of 76 people for the Library of Ccmgress Unit is unrealistic.
GG0ALTIVIAD:13C/H13D
nrt: vns
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12711-70. B CAM
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 r
TREASURE ISLAND
THE PROBLEM
1. To organize an extracting and digesting service to exploit for
Air intelligence purposes open sources, including Slavic publications of
all kinds.
HYPOTHESIS
2. The need for intelligence materials and open source information
will become increasingly acute.
DISCUSSION
3. A very large portion of information pertinent to the mission
of the Air Intelligence Division is available in open source materials.
The intelligence mission cannot be accomplished in a satisfactory manner
without a systematic and thorough exploitation of unclassified publi-
cations, such as books, periodicals, newspapers, foreign government
handbooks, broadcasts, atlases, etc. At present, little exploitation
is being carried out although the source materials are available in
the United States. Consequently, a great deal of pertinent information
is unavailable for purposes of air intelligence analyses. (See
Appendix A).
4. The extracting service would have to be financed, yet cost re-
quirements cannot be estimated at this juncture, because it is not known
which of the existing facilities can be brought into the scheme. It is
clear, however, that the investment required to finance Treasure Island
will be small, at least in relation to its usefulness. It Is believed
that, within a few years, it will be possible to say: 'Never before was
so much intelligence gained from so small a cost.'
CONCLUSIONS
6. At the present time the Air Intelligence Division does not
receive basic information from open sources either in sufficient
quantity or in a quality suited to its needs.
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 r IP
itit4?1
6. The satisfactory accomplishment of the air Intelligence
mission requires a coordinated effort by the Directorate of Intelligence
and other agencies for the exploitation of open source materials.
7. The organisation of such an effort requires the setting-up
In the Directorate of Intelligence of a special unit, tentatively called
Basic Information Section. The extracting service can best be started
through a contract with the Library of Congress.
RECOMMBNDATIONS
S. That a Basic Information Section be set up in the Directorate
of Intelligence.
9. That the Basic Information Section be instructed to set up and
supervise a service for the extraction of pertinent information from
Slavic open source materials.
10. That the Basic Information Section be directed to negotiate
? contract with the Library of Congress in order to have the Library
do preliminary indexing and searching, and in order to start the unit
processing current books and the backlog.
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
APPENDIX
INTELLISENCE FROM OPEN SOURCE MATERIALS
INDEX
DISCUSSION.
The Value of Open Source Materials
Current Exploitation of Open Source Materials
Remedies
Results of Discussions with the Library of Congress
Open t4uestions
Basic Information Section
Academic SOUTOOS of Intelligence
ANNEXES
Annex I - Chart, Open Source Procurement
Organization
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @_50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
APPENDIX PIA"
INTELLIGENCE FROM OPEN SOURCE MATERIALS
SMEARY
1. Open source materials are of great value to intelligence.
A great deal of valuable air information can be culled froa publics.
tions, such as books, periodicals, newspapers, broadcasts, and gov-
ernment reports.
2. At present open source materials are inadequately exploited.
The Air Intelligence Division does not obtain the information con-
tained in such materials nor Is there an organisation which is
extracting air intelligence data that are available in such materials.
S. There is a great need for reorganising methods and procedures
for the procurement of open source intelligence. Such a reorganisation
requires the setting-up of a Basic Information Section in the Direct*
orate of Iotelligenoe; the setting-up of an Administrative, Piling
and Publishing Section in the Library of Congress; the establislhment
of procurement unite in the Library of Congress; the creation of an
extracting unit in the Library of Congress; and a coordination with
extracting and abstracting units in the army, the CIA, and other
government agencies.
7fritsipur
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
)eclassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
IA-RDP81-00706R000100230035-8
?
Now
APPENDIX "A*
INTELLIGENCE FROM OPEN SOURCE MATERIALS
DISCUSSION
1. The Value. of Open Source Materials
a. Intelligence work is based on two different types of
materials: open and classified. Classified materials include those
which are originated or processed by United States governmental
agencies and which are kept under strict security restrictions. Some
these materials are obtained by secret means. In most instances, they
report on facts and developments which foreign nations want to keep
secret. Among the open sources are those which are originated by forel
governmental or non-governmental agencies, to which the originator does
not attach security restrictions. These materials can usually be bough
In the open market or obtained without too many difficulties. They in-
clude magazines, newspapers, books, government reports and documents,
broadcasts and news picture..
b. In customary intelligence routine, great stress is laid
on working with classified materials. The importance of the classified
materials cannot, of course, be mlnimized in any manner; a large per-
centage of significant facts cannot be gathered except through secret
sources. Yet, the two types of source materials complement and supple-
ment each other; taken singly none of them gives the complete picture.
Despite the fact that classified intelligence documents will be most
profitable if they are correlated with open source materials, the value
of the open sources is often underrated. Only small efforts are
made to procure and exploit the unclassified type of intelligence sourci
materials.
Declassified in Part- Sanitized Copy Approved for Release @50-Yr 2013/12/11 :
r?IA ID nosz I_nn7fARM)100230035-8
Declassified in Part- Sanitized Copy Approved for Release @50-Yr2013/12/11:
CIA-RDP81-00706R000100230035-8 st(RET
C. It is sometime
felt that open sources do not cnntain
eny eateries that are worth gathrring. At least, this is supposed
to be true with respect to the open anurces rubliehed by dictatorirl
countries, the cenror allegedly cutting out anything that may be of
use to foreign intellieence services. However, experience with the
Carman censorship during the war showed that even very important
military information can be collected from newspapers. /or examP141,
it was believed that German PIT boats were armed with 20nm guns.
Tnterpretation of a. photograph published in a Carman newspaear stowed.
that these boats actually carried 40mm guns. On the authority of
General ronovan, it appeare that mort of the inforrrtion about the
Carron oil industry err Fathered from newspapers. It may be added
that strategic and tactical doctrines, too, were freely discuesed in
Oman technical publications. For example* the joint employment of
aircraft and-tankr was described as frrly as 1934 by a Omen General
in a book which was easily obtainable all over the world for 02.50;
the tactics which he had outlined were applied by the Germaun through-
out tha aerly caspairns of the second world War.
d. Rowever severe, censorehip in never fool-proff. Censor-
ship 7ernonnel is rarely too intellirert,, ancialmont sever w,11
in-
forred. Very often items are published of wh!eh the sienificance
eecapes the censor. Dercriptions of cities and plants %many give
away some good diner. Obituaries often contain inadvertent references
to person's place of work; advertisementa or propaganda indicate the
output of a elven plant; technical diacuseions of production processes
ecestimes permit deductinns no- to locationa, 'witty and output. The
Russian rrens offers, moreover, the specific advantage that (1) the
labor crews of factortee frenuently send telegrams to Stalin; (7) praise
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Al\
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
COS
is given to specificelle? nnmed rlants whieh exceeded their output
assignment; end (3) criticism is often voiced of production ecbieve-
ments in named fectories. Altb-ugh many of theme itens ere camouflaged,
and incomPletely reproduced, additional informetion from claseified
well tee open ores ueually peri i the cameletion of the etory.
C. eenrorship recuiremeets are thangeeble. Items that are?
eecret now were freely revealed e few years Apo. From ties ts tire,
facts, yuch ts locetions rd typ-e of facto tee, remain more or less
conetentt present-day censorship is not necessarily ea obstecle to
collecting pertinent information. For sweep's, the Rneriane do not
now nublirh infOrnation 'b'ut uranium deposits. Tat the litereture
contnine Trey reference? to uranium depoeito in the raviet Union. Time
an sgain, it was poosible to evaluete claesified reeorts on the dis-
oovs-T of uranium on the beets of previously -published peoloeicel re-
ports; in some instances, it vas forecast where the Ruestnns would soon
start digging for the uranium in Czectollovekie. For that natter,
cnrrent Ruesian eutlicettons occerionelly contein references to 'rare
minerAlse which, nometbeel turn out to be uranium.
f. /Is the specifie field of eV intelligence, there are
severel terg-t cetegorice rhich must needs be worked up on the basis
of open materiele. It to not implied that now useful information on
these categories is being attained from classified sources, but the
fact remains thnt the over-ell picture of these cateeories hes been
gained thraueh open eaurcep, includine tease eerman-captured documents
which hed been veTked up from Emden open publications.
fr. Ile the railroad category, rail trackeee, stations,
bridges, =meth/aline. verde, over-ell tornAgee crried? tyrer of com-
moeittee cnrried, tynee of rolling stock need, caeacity- of locomotives
end freight ears were obtained from open sources. There iv also avail.
- 3
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706 R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
SECRET
able a great deal of informatiot ,bout railroad operations. There
is no information on a few secret rail lines, nor an freight train
time tables, but open information is adequate to list mort of the
importent rail targets and, if necesoary, to pormit the SV !ranch,
without benefit of classified inforattion, to set up rail targets
the destruction of which would greatly impair $ov!et rail capacity.
h. An estimated 75 per cent of the electric power stations
listed in SIT files were idertified on the basis of open materials.
At the present time, the Soviets do not oven to censor news about
power stottone, with tha eoceotion of info ,.ition on grid systems.
The Soviete? usually, announce 'hen o power atation starts operations;
In Fon, instances they go as far as revealing details about generators,
including the places of production, capacitso, plant layout, etc. Neg.
"Alva clues can be obtained about ares which are not mentioned ?Or
example, the Chnsovoy hydro?electric program which figured orominently
in the news un to ebout two years apo, aaparently is no longer commented
%roe. Since . there is information 1-out variouu important ataeric develop?
ments in that area, Soviet ollonce providew sone kind of a confirratior.
I. It is most eurprislno that a great amount of open source
infor-a*.lon is obtained on highly claasified industrial ctogories.
The diecovery of new oil fields is usually reported in the Soviet press.
The British, by oetient research in paperv and periodicals, were able
to ieentlfy about 40 nircraft plants, their locations and
numbers.
3. Soviet papers contain frequent mention of new cities. It
'Will be recalled that, acoordIng to Marshal Rundetedt? the Wean army
ear ereetly handicapped because their maps failed to show newly erected
cities. If the Germans had ruffictently exploited the open sources or
information, most of the gaps in their knowledge of Pussian towns would
have been closed.
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706 R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
Slab
CIA-RDP81-00706R000100230035-8
k. A good illustration of the value of open nource mateeials
is provided by the work done by Jan Scott, the Aaerican engineer who
spent many years in Rusaia. Early during the war the Board of Economic
Urfare put him on the job of collecting open source inforration on
Russia. lith only four female asslatantel Scott produced an 8l5-page
volume itRussia.test of the Volga*. Practically every industrial cate-
gory And all minerals were covered in the document, which also listed
dosene of new cities, raft today, the Scott Report is often used to
great advantage. It contains information ab-ut many plants and places
concerning which no other data are available in SV files.
1. Open sources are indispenseble for over-ell economic and
induetriel analyser. laisleadina and incomalete th:veh Ruseian statis-
tics are they are nevertheless the only available basis for statintical
evaluations of the soviet war potential, tha productivity of individual
industri-s, and the general organisation of the rtviet economy. nth..
out the periodic reports on the progrets of the rive Tear Plans, no
over-all pirtuee of Pordet capabilities and vulnerabilities cool.: be
obtained, despite the fact that no Ansa= statement can be accepted
at face value, but must be thoroughly analysed and checked with other
data.
m. Open materials are equally molded for the analysis of
atrateeicelly vital categories. ?Or example, without unclaseified
basic geographic data such as climate, freesin conditions of rivers,
water depths chemical content of eater, elevations, roads, geology,
etc., it would be vary dirficult to crack the secrets of the
atonic program
r. Russian open sources cantain a great number of photoerap s
and other pictorial materials, ineudina plant layouts. There are
quite a few industrial pictures which could be used for targ.t folders.
There are Also pictUre nroduots which would give clues as
e of Pussien p
5
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 Mir
to output of specific plants and nuality of Russian tochnolog7. There
are numerous pictures of outstanding architectural structures which
would be very veluil3le 8 9 id 911 ..i.flabl* landmarks.
0. It is core'uded that the exploitt,tion of open eources is
not only a very velu'Ille but actually en indisnenneble nart of air ir-
tellirence operv tiers.
2, current rxoloitntinn of Or -uree,iaterials
a. Preliminary surveys have shown that a very large percent-
age Of total Iberian open-rource :',7ate,r-tals are availe:ble in the United
Stetes. In the Teehineton area alone, the Library .,41 Congress plus
veriour othPr !?overurPilt A arenci such as the CIA, Army Library, and
State Depart:I-et have been receiving, during the 13t five to ten
years, at least ?-ne-half? or more, of the Ruesian output in publica-
tions. Large university libraries, such as Harvard, Columbia, Berkeley,
and the New York Public Library also posses conaidernble collections
of 'Russian materials. There are innumerable small but nevertheless
valuable Slavic libraries in the hands of minor universities, special
research organisations, industrial fires and nrivate individuels.
The Library of Congress is now trying to obtain all current Soviet
publications on a regular bAsis.
b. Several g3gencies have been set up to produce bibliogra-
phies, ebstractr and digests fror, these materials. The Strategic
Vulnerability Pranch receives about a closet of such publicati-ns, not
including technological abstracte oroduced by private agencien. some
inforeation ito extracted by FV analysts from English-language publi-
cations issued by the rcrriet government. Prom time to time, additional
infore4tor is Mtn er, fro' th dtgeste published by the 13ritieh Joint
Intelligence Posrd. Recently, the CIA startad the Sovi.et Press Ilbr-
tracts which hold out some ororliee, but, notwithstanding this seeminvly
?6.
Declassified in Part-Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 r ' 4-VIA.Ktri 11.
, . .
. '
large quantity of open source materials there is no flow of information;
there is at best, a trickle.
o. The reasons for this bielly unsatisfectory state of afters
are the following,
(1) The extracts and digests are nosey made for purposes
which are totally different from the mission of the
Air Intelligence Tivirion.
pest of the extracts are actually made by people
unfamiliar with air intelliaence resuirementa. Rally
digests, eopecially the tally Reports on Foreign
Radio Broadcasts, attempt to reproduce current news
and propaganda, tut coreictently omit intelligence
data
(3) The confOelon in the abstracting field haa had the
deplorable coasequence that while many items are re-
produced in most of the publications, there is not,
ba a long shot, a complete coverage of ell the materi-
als that are actually available. Neverthelens, the
sheer volume of these unorganised and uncoordinated
digests is so large that it is very idfficult for the
SV trench to handle them properly.
(4) !title every abstracting service tries to include supra-
sentatire Soviet publications, such as Pravda and
Isvestiva, little attention is given to books, rovern-
wet publications, (for example directories), pictures
armee, Pence a great many of the valuable enurces
are not exploited, thile other soneees are exploited
reverel times over, but each time in a manner not
suitable to the Air Intelligence Divieion
(2)
Remedies
a. A preliminary survey of the situation as well as dcuatons
.r
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 lor
?
With specialists in the abseractine ficle have shown that it would
not be pracLical to organize one big independent super-abstracting
service which could serve all the interests of all the agencies. even
if such a service were practical the setting-up of a epeciaI agency
would require a very 10 .etime. eover, there weule be no certainty
that air force intelligence requireente woule be saLiefied unless,
OX course, the air force woUld be able to exert control Over operations.
b. It would, therefozei be neceesary to make use of the ab-
stracteee units that aireade- exist. An arreleeaent with euch units
would be profitable not only for budgetary reasons but alee because
some oi the units have control over source meter/ele-
c. At the some time, since present extractine staffs are
smaU and overworked, and since a greet deal of tee' reeeired work le
eot being done, the establishment oe new extracting units cannot be
averted.
d. Hence, a solution must ee found leach (1) avoid the
setting-up of one independeat super-extractine aeency; (2) achieves a
maximum of cooperation from ox tin. aeenciee) and (3) provide
ties for the carrying-out of such extractine eervecos as are eeceeeary
if a complete corae of open eource materLais le to ba achieved.
0. Before a satisfactory plan for the division of lanor amoag
the various wouldebe participants in the prot7ae can be evolved, IL is
neceseary as a erellaenary, to determine -viten eources oueet to be pro-
cessed ane to locate the eource materials not now in eacenington but
available elacweere in Lila United States. euch a preliminary job inlet
nee as be done by the Library of Congress *Lich woule; have to be given
a contract in order to enable teem tie coeplete their index of Slavic
atriale and to make full use of th Union Catalogue.
4. Hesitate of Discussions with the Library of Cengress
a In informal discusaions with representatives of the Library
8
Atz4
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 Y 11.
of Coit was determined that an eetractine service could not
function satisfactorily without a aearching unit permanently operating
in the Library, or withoue a routLee channel to the Acquieitions aivi?
sion.
b. It wee thouht adyinable that the Library of Congress
participate in ay extracting scheme that aieht be evolved. The Lie
brary is the government's purchaeing eeeat for foreign puelicationa
and it poesesseaeuhat is probable' the country's largest Ljavic collec?
tion. The Library is the organizatioa moat suited for locating and
loaning operations and it is the beet medium for the camouflage of tbe
entire undertaeing.
0. There was no argumene about the need for a new unit if
the backlog or books and other publications an well a the current
Eleet3810115 of beoke are to he handled. The backloe is estimated to
compriee 20,000 booke, more than 30,000 periodicals arle more than 100,W0
newspapers, not to mentioa the backlog of ot at waterials, such as broad.
caete, for example. It hae peen. tentatively eetimated that euch aunit
would requira aoout 50 extractors and indeetre an 10 CAFte. A unit
of that else could process the backloe in 5 to 6 years, and also handle
current acquieLtions of Cooke and covernment reporte. euch a new unit
voule work most efeicieetiy in the Library of Gongreza. (see Annex 1).
d. . With respect to the actual proceeaing of the materials, it
'would be moat effective if extracts were done on check sheets liettng
a. limited member of i:wrtieent queetions, aeon as those an the terget
inforeationsheets. At the end of each cheek sheet a akort dleeet of
the proceeSed materiel would be appended. The extxact enacts would then
be oalleetoOd and consolidated and pub into a publication, to es issued
at regular, intervals, preferably weekly. The laforeetion 'ovule also be
9
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
Declassified in Part - Sanitized Copy Approved for Release 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
SECRET
44iit .3 'Amami ilk.
so. The discussions suggested the advisability of limiting
the scope of the extracting effort. If, for example, political Jo--
formation were processed, the =tire project meld have toll's very
large in sise, while the expenditure would not be worth the result,
certainly not in so fir 46 the Air Intelligence Division is concereed.
The extracting service could be profitably limited to the fields of
economic and industrial information related to strategic vulnerability,
technical air Information, air are and civil air information, bio..
erashical air information, and air geography.
f, It woe recognised that even within this ilimitaticoe
economy of effort amid be practiced. For exanple, a great deal of
information .may be available on the paper industry, but there might
be no current need for this information; hence, it weued be wasteful
to process and reproduce it. However, since intelligence needs can.
not be anticipated, and since it is conceivable that, for one reason
or another, even the paper industry mey assume inportance, a central
file would have to be maintained in such amanner that, at a monentle
notice, any inactive category can be worked up. It might even be ad-
visable to maintain the file for subjects which hardly ever will become
of interest to the Air Intelligence Division, simply because another
over:upset agency mmy want to participate in the program at a later
date. At would be a good selling point to have all the information
indexed and cromeoindexed so that extracts of +Lay kind ean be made.
g. For various technical reasons, it is believed that the
Library or Congress would be the logical agency to put in charge
of both the actual publishing and the central file. In order to get
the ex:a-acting effort going it would simply be necessary for tne eir
10
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 A
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 r glr
. ?
force to make a contract with the Library of Congress ehich then woad
set up the various activities and units tlined above. It is believed
that once the extracting service has been started, cooperation from
other agencies will be obtained more easily.
5. Open Geeetions
a. Up to this pcee-,t, there is not much roon for eisagreee
meet about the organisation of the extracting servleee. Di ficelties
arise, however, it rasp ct te the eays and rxans by which other
government agencis an private abstracting firms .could, or should*
be bleught into the echeme. It is relatively easy to conceive o.. a
set-up ey which the labor would be efficiently and equitable -Avided.
F,r example, if the Library of Congress would handle current hooka,
government periodical*, and the backlog of periodicals and neespapers*
the Army Library, mhich operates a periodicals abstracting service,
could handle current periodicals. The CIA, which already has made a
start in the field of newspaper abstracting, coeld take over current
newspapers. The CIA also has facilities for radio monitoring and for
the procerement of industrial doc,ments. Contracts coeld e made with
private firms for the erocersing of tec%noleeical and sc1-ritific mere
t rials. Specialized oieanieatioes, such as the Photo Library and the
Aero Chart Service, cold profitable participate in the processing.
The extraeAlne services would have te ??e coordinate with the Air Re-
search Unit at the Library of Coneress. /t mead also be advisable to
arrange for a division of labor with the British JID,
b. Although such an overeall or anization can be cencetved
in theory (See Annex I), it must be anticipated that various adminis-
trative and "political" obstacles will arias. It is obvious, at any
rate, that the setting-up of a coordinated effort on such a?scale will
require a great deal of wore. Lenethy negotiations will be required.
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8
IDeclassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11:
CIA-RDP81-00706R000100230035-8 r
_ IIP
?
Wen after contracts will have been ooncluded, adjustments will Al nicer
gory quite frequently.
6. In sum ups in order to get ? fully satisfactory service
speoialising in the extraction of intelligence from open source materials
and, at the same time, to derive benefits from the research done at
military and eivilian schools* it is necessary to set up a Basis In-
formation Section in the Directorate of Intelligence. This &motion would
maintain liaison with military an civillossehools and control the
administration of the =treating services. This edlirdstratior., to be
legated in the Library of Congress, would control procurement of
%oriels, maistaiaIisison with the Acquisitions Division in the Library,
keep a central Moo and publish the extracts. The administration would
eontrol the various extracting units, located in the library of Congress
and in other government ageneies. The wqrk would most advantageously
be divided up in mush anions? that each participating agency mould handle
one major category of open scums materials
7e She total manpower and money requirements of the scheme Gannet
be estimated at this jamstum bosses* it is net knotnatsbleh =tett%
facilities can be brought into the *shams. An exact budgetary estimate
must wait till the time Then nagotiations with various government agencies
trill haws bees engaged. It is clear, however, that the Investment required
to undertake the Treasure Island scheme will be rolutivaly mail and that
the mousgorial dittielati?? and costa will be minor. The gain in term of
information obtained will, by contract, be very largo. It is more than
probable, indeed certain, that within a few years it will be said that
never before in the history of American intelligence was so staple and in..
expensive an effort needed to produce a, :peat and so valuable a return.
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2013/12/11 :
CIA-RDP81-00706R000100230035-8