THE INTELLOFAX SYSTEM

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP84-00951R000300040002-5
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 93
• Document Creation Date: December 16, 2016
• Document Release Date: September 10, 2004
• Sequence Number: 2
• Content Type: REPORT

Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
THE INTrn LOFAx SYSTEM
Introduction
The history of the Intellofax System encompasses
many facets of the infcrriation and storage retrieval sy-
stem of the Office of Collection and Dissemination (CCD)
and its successor, the Office of Central Reference (OCR). Two
Divisions., the Machine Division (Mn) (Central Index of the
Reference BranchIdOffice of Reports and Estimates RE], until
May 19)48 and then Machine Methods Division [mil of OCD until
September 1951) and the CI.A Library (Intelligence Documents
Division, ORE Reference Branch., until May 1918) were responsible
or the development and operation of the Electrical Accounting
Machine (EMI) upported document storage., reference,' and
retrieval system. The office reorganization of November 1956
added a third layer of responsibility-a new Document Division (DD).
This history covers all aspects of the Intellofax System
from 1947 until its demise in 1967: equipment developments and
improvements, including microfiiidng, print service, and fast
transmission of data; classificstion input scheme; and retrieval,
A project that had great impact on the intellofax System but was
not adopted-Minicard- is also discussed in detail0
The Intelligence Publications Index (IPI)s the printed
index of finished intelligence documents, is historically
part of OCD/OCR's information storage and retrieval system and
therefore appears in this chapter with the Inteliafax System.
s* For the Intellofax System as it related to the Graphics Regis ter,,
App @wdf&"glease 200 :cQ1$ 951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
The effect oft the intellofax System of the Library Consul-
tants' S,zrvey of 1957 and the resulting Task Team Reports of
1958 is discussed in Chapter (the CIA Library) of the
office history becaus' of the overall impact on the Librar .
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 13
Approved For Release 20U4/09/23 : CIA-RDP84-00951 R000300040002-5
1. Early Developmental History
a. Objectives
In providing a central reference service to the
Central Intelligence Group (CIG) and its successor, the
Central Intelligence Agency (CIA), as well as to the
intelligence community, the early managers of the Agency
recognized the need to develop a machine capability for
indexing and retrieving a staggering quantity of intelli-
gence documents. The resulting Intellofax System was
unique--no other government agency, no university or
library, and no c omnercia 1 firm had anything of its
type in operation. The name was coined in 19419 by Drs
the first Assistant Director (AD) of
OCD, to d.eocribe the system that combined IBN and facsimile
reproduction techniques for intelligence documentation pur-.
poses. Later, Intellofax became a household word not only
as an adjective (the Intellofax System and the Intellofax
files) but also as a verb form (intellofaxed and intellofax-
ing for the indexing aspects).
The actual authority for establishing the Intello-
fax System appeared in July 1947 in ORE Instruction 31!171,
1
entitled "Functions of the Reference Center, Thomas Babbitt,
AD/RE, charged the Central Index and the Intelligence Docu-
ments Division to:
(1) index, by bn.siness machine procedures,,
the subject matter of all available reports,
and other documents of a foreign intelligence natur
(2) classify and catalogue all intelligence documents
of a foreign intelligence nature available to CIG.
Approved Fo Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
ORE Instruction 31-7., 15 Jul 47 sub : F cti ,U of th
Referents enter, Oki", pages aria o S. `a.le : CD His or,~ 1y~
5rr- WA
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
b Early Equipment Needs
25X1A9A I I Chief of Central Index, was given the
responsibility for organizing and developing the initial essential
steps toward establishing a central indexing and filing system,
in conformity with an earlier Interdepartmental Coordinating and
Planning Staff (ICAPS) recommendation in March 19117. It soon be-
name apparent that no existing equipment would be capable of meeting
the needF envis gai Although an IBM punch card offered
great flexibility and speed in the handling of thousands of cards,
each of which would represent a particular document, no card would
carry enough printed data to supply the researcher with titles
and descriptions of documentse
During 19)i7
met with top management of Finch Communications, Inc'to discuss
the possibilities of the use of standard Finch Telefax machines
and the adaptation of these machines to the documentation problem.
A/ice resident of Finch said that his company would be willing to
cooperate with IBM in adapting the Telefax machine t o automatically
reproduce bibliographic and subject abstract data typed on IBM
cards onto any type of paper that included a duplicating
medium. This would answer the problem of preparing,, accession lists
and lists of abstracts requested.
*Management originally planned for a daily accession list 6f `thee e
documents received and indexed, all of which would be abstracted.
This pla}.-vas given up in 191i9 as entirely impR ctical and uneconomic
Metro, -sting Chief Reference Center, CRE, to Chief Central Index
Approved 9/2358 P84 00951 R000300040002 5 Filing Procedure* U,
JW. afthkO f
dog-~,-""ar't
 Intellofax 5
Approved For Release 2004/09123 : CIA-RDP84-00951 R000300040002-5
After numerous meetings with Finch and investigation
of other companies, such as RCA and Eastman Kodak., the machine experts
opted for Finch, and a contract was let in January 1948. By July
Finch had produced the first of the Library Recorders and had
Y
completed the final design for the IBM card scanner. Both awaited
CCI) approval. Experimenting and testing contindod, and in January
25X1 A9A V
go 1k l91G9 I reported favorably on the equipment0
Progress repots were prepare, peraoazca.= ~;yr ' utuvu6laVUV
e ?c3
r00
"T
va Fa
19L
061'',
43
br-i
00 4)
+O * 0
~W
c39
ed -H 25$ 9A
o
n
.-
i.4 rl 0) I 1 ?ci
the first 6 months of 1949; test runs were made during June, and
the equipment was finilly accepted in July. The Projects Review
Committee (PRC)
contract, which
6
$203.,0000
The
an IBM punch
on 27 July 1949 approved an amendment to the origingLl..
had been for .$100,000, to increase the amount to
Intellofax Card.,
card of standard
face up to 200 words
or Faxcard:(see Figure 1) was
shape and dimensions, which bore on its
of printed information- the so-called
bibliographic data: source, country, date9 title, possible abstract.,
pagination and seempity classification. The corresponding coded.,
Punchedand interpreted data appeared at one end of the card. The
cards were sorted., selected and arranged by standard 13M machines
and the printed information on the selected cards was transmitted
and reproduced by facsimile process.
The equipment delivered in May 1950 was the second proto-
type resulting from the developmental engineering begun in Nay 1948.
)( hakedot 'tests were still being conducted in mid-1951 concurrent
W;Yl AAA
with actual usage0 ., an Office of Communications
A~~ fr~~a~ 2 i8S/ P 4-611 '1 90=b?4bQ(BZt Travel to Fin
telecommunications. line. V.U0 (in Machine Division 1917-5r'-) 1)
5-k 9VS
 Approved For Release 3004/09/23 : CIA-RDP84-00951 R000300040002-5
employee (and formerly an, ngineer with Finch), was on
temporary dutyt with OCD and placed in charge of the Faxcard
25X1 A9A?Y-3
equipments He wrote to (chief of
eon- since September 1950 0 that since the equipment
was not standard, additional development was anticipated before
the equipment's stability could be placed in~a class with
that afforded by existing teletype machines.
The Intellofax tape, as it was known throughout the
entire Intellofax history, -as originally a 4-inch-wide
tape prepared by the facsimile process. The Intellofax
punch card was fed into a transmitter, which optically
scanned the printed information. ,receiver received signals
from the transmitter, the printed information was impregnated
into a chemically treated tape, which was dried by a heat
process.
The early OCD managers had hoped to electronically
transmit the Intellofax information to requesters in their
own office locations. As of 15 May 1950, six transmitters
and 12 receivers had been delivered (see Figure 2).Experi-
mentation continued throughout the summer months' and
'via s
the first t ansmissionnst icttAy local--transmitter and
M
nT _ PST , ,. _ -, _ n ? ^ fl.'-?,~,o
receiver side by side in
0
.?J
One receiver was placed in K Building in the Branch Library, but?
security considerations and technical problems of transmission
were responsible for not continuing with what seemed like
~riemo,
Machine Division 190-r; i'
f
401\ 1 A * trans erne ,T5V
25X1vectMbr
zcNi;~: l?!1' 1'~p 095X9
;0 ~rrnt. Uo (in
n
~Zp
.
a
lu~
d to the Management Staff
an transmission phase. The completed folded tape s
Gt
i
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
M. Coding Schemes
(1) The Intelligence Sub ct Code
In conformity with the wishes of ICAPS., the Central
Index also took steps to prepare a unified subject classlfica-
Lion schemed Acting Chief of the Reference
9
Center$ wrote to in July 19147
Although the Reference Branch has taken the
initial steps in the direction of establishing
central indexin and filing procedures,, any
unified acceptance of the end product of these
inve s ti 2tions will depend upon joint action
of lAB Intelligence Advisory Boar and CTG
representa tives and the agencies' final
acceptance of the system decided upon.
On 14 July 1947 F 5X1A9A I entered or, duty as
Chief of the Classification Unit of the Intelligence
Documents Division to work ww ith the Central Index in
1C
developing a classification schedule for C ft
tt was soon evident that the War Departments
Basic Intelligence Directive (BID) devised during
World War II for collection purposes (although it had been
used for classification of documents in the G..2 Library in
Vienna immediately after the war) was not adequate. The
subjects listed in the BID were not sufficiently compre-
hensive to cover the wide range of subjects in intelligence
documents because it had been devised for Arrmj purposes
only. The economic political , 2nd sc"?nti: =i_c sections
were wti eful7,y weak., It as decided to prepare a list of
subjects that would include those contained in the BID,
of Addicott memo (2 o ab ov?e )
Zt Reference (;enter Library Monthly Status Report 26 June-
28 Jul.;; 47. Sa File : Ljbraiy 1911.7.18 Job: 158-93/1
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 5/'
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
the Navy Monograph Guide,., the abridged Dewey Deci-
mal system used by the State Department, and for
scientific subjects, the Voge Classification, pre-
pared and used by the Joint Research and Development
Board (JRBD). the Librarian,
and made visits to the parent organizations
using these classification schemes.
5X1A9A
three people, with the assistance of a classi-
fication specialist from JktD, had completed a general
framework of an all-inclusive classification schedule.
The major subject categories included Army, Navy, Air,
Political, Economic, Sociological, Scientific, Geographic,
and Biographic. On. 22 August a familiarization meeting
was held with duly appointed representatives of the three
services. The participating IAB agencies agreed to de-
velop and/or revise their respective military categories
in the BID. To tho 4-00951 R000300040002-5
3 Procedure ~j''''aannua is 159 above
~j/Intellofax Chronology- Appendix
 r~Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
1 (I) "'Flash"
An indexirg economy r'sveloped !4as the so--
331 /"
called "Flash's procedurtTZ By January 1949 classifiers
.reconized that there was a sizable segment of documents
1 i covered the same subject matter each time and
were issued periodically. "Quarterly Mil .tary Review ?sl
"6Jeekly List of Infectious Diseases' "Semi-Annual
Railroad Statistics" were a few examples. It was a
waste of indexing and key punching effort to code.
these documents every time they appeared? The .
such a documnnt was encounter6dk,.it was coded and
abstracted in a table of contents moaner; the Intellofa :
card bore the usuel bibliographic statement but without
report nii: for or date. The w ord "Flash" ~npe red on
the ThM card. A master Transmittal Sheet on which the
abstract was prepared was filed in the "Ulash* }ok
in the Analysis Branch. :Every time a similar document
was received, its report number and date were entered
into the FFlash" book` When a
subject retrieval request turned up the original "r7ash's
card., the librarian or classifier knee that he must
refer to the "Flash" book to find all the documents - had.
been published and receivedi This "Flash" record was the
only means of determining documeht numbers in order to
retrieve the material , the document file*
rocedure Manuals ( l above)
FAT
Approved For RelPasP 2004/09/23 - CIA-RDP84-009518000300040002
 i d4 uaa
S"ID? Fry
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
periodically distributed
hts it
/ i
ciently c9 ered by tY codes ! or gi-hal3,y
covveragg' in the s erio did nectssita e added codeA.
The "Flash" system finally outlived its usefulness.
Irregularity of ytssuance of certain seriess the manual
labor involved in maintaining the s1Flash" book
V f or a period one cow was maintained in the Reference Branch
also, the nuisance of not having a source card for every
document, anO the advent of the revised Intellof x MY I-V
~ystem--all combined to its demise in 1960a
./ ' Memo n C/DD'LY and NE) to AD/CR "Elimination of Flash
Entries in Intellofax Systems" CIA Internal 18 Mari- 60 -(
(in Library 1959-60 64-341A)
the ISC asp Droll as *ider subject
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
The Abstracting Program
was very proud of his central reference facilities.,
and by late 190 he felt that OCD was on the track of a "far
more effective solution to providing analysts with quick re-
trieval of intelligence information than has ester, anywhere,
been achieved before." He devoted 32 pages to a definitive
description of the Intellofax System, entitling his monograph
- -2 And
Central Referenc9. - Facilities. Status (l. November 19
3 ii h'became a class uay in LU
'ectiygav .He had prepared this paper at the request
Ob7
r
of the Chief, Coordination Operations and Policy Staff
Staff proposal that the Library classification functions be
3~
decentralized and placed in aubstantive units of ORE and OS
In his usual dynamic ataph-, had,briefed
STAT
specialists in ORE on the potentialities of the Intellofax
System and how analyst participation in the growth of the file
would benegit the. system and therefore the whole Agrncyo in
other words, he was asking analysts if they would like to contrive
bate abstracts to the file. He stated that only the specialist
could decide which were the important documents bearing upon
his field of specialty and only the specialist could write com-
petent abstracts of such documents. In January 1919 the Turkish
desk began selecting an-4 abstracting; the most important documents
on Turkey. OST also commended operations on a trial basis in
Approved For Re ease 2d 4/09/2 DP8 Ob 914000300040002-5
34 C/Pk? 7 & CIA -o arian to AD/GIG "Comments on Mara ement's
Proposals Relative to Abstracting and Codin of Intel'_Ligence
Information" 10 Nov 1,9 SEC ~,'T (in LY 4 Lo, . j?-5 i
 3 d
February and thel of ORE followed shortly thereaf taro
A so-called contributor code was punched into colurm 21 of the
IBM card so that if a specialist ever wanted to retrieve only
his on abstracted material., he could do so 0
(rhis seldom occurred.);
OCD rieantiahile continued to write abstracts of a table of con-
tents type for publications covering a wide variety of subjects
and areas, Much as he had fostered and approved of the abstracting
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
became concerned about the rapid. growth of
these files in Intell.ofax, By the end of 19149 there were
l8, OIi 7 abstracts (OCI),w28L1.7; ORE,-62b5: OSI 8955),, and they
3
were growing at a staggering rate. He warned that each extension
of the system to a new group of specialists involved difficult
adjustments., revisions ) and expansions in the classification scheme,
required more typists)and could be accomplished only by day-to-
day hard work. How trues
spent a large proportion of her STAT
time in these early years working on changes and expansions oft ho
ISC to satisfy the needs of new contributors to the abstract program.
Area codes for three of the Near Eastern countries were expanded
to take. care of provincial divisions.: And as it turned out they were
never used for retrieval.
The abstracting program mushroomed from its beginnings in
On pages 23 and 25 of the Intello Iax study. ,I
provided samples of OCD ORES and OSI abstract cars.
3 ,atistical Reports (2 ~28 above )
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
19499 reaching a peak in the early 1950's. Other desks joined
the program in 1950 and 19510 Some of these desks even set
up their filing systems according to the ISCQ ORcontributed 169558
abstracts in 1950 and 109695 in 1951. OSI contributions amounted
to 209186 and 209715 in those 2 years.
However., a dovmward trend began in 1952453 because specialists
were not interested in spending their time learning how to use the
ISC correctly for in-depth indexing. The_ Physics and' Medicine
DIjvisions of OSI continued as heavy contributors but. finally ceased
In 1958
abstracting in 1954, and 1955, respectively. ORR/shipbuilding became
the last ORR component to stop input*
OCD management had always hoped that these contributions
would preclude the indexing of the same docaa.ments b:r the OCD classifier;
This was never possible, for the seialists frequently extracted and
coded only th part of the document that interested them. So so,-re
duplication continued; both IBPN cards would turn up on a machine
run if the same codes had been used (and there was no guarantee
of this). In such a cases the OCD classifier or the librarian in
screening the ccr ds before an Intellofa'x tape was made would
A study of "Information
Processing Needs in the CIA" refers to. these noble if somewhat unsuccess-
ful efforts of OCR to harness the specialists' indexing skill.. The
25X1A5A1 0 consultants regretted that the problems coui.dr .ve been soled
without abandoning the total effort and that therefore the indexing
capability of subgt ntivo specialists was no longer substantively
25X1A5A1 exploited by 00 L0 II "Information Processing Needs in the CIA
A Preliminary Appraisal" 10 March 1961 SECRET CRS Historical Files)
3 'Statistical Reports ( (2,'b above
39 ~ PAppend:lx
I
~~ J
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Abstracting 3
'Z--.
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
pull out the nonabstracted card as being less meaningful]
A great danger to the Intellofax Dystem
input t o the
surfaced after the research offices had stopped tF}e. program.
References td up that could no longer be retrieved., such
A as articles in Russian scientific journals tndex}'by OSI and
later throve away. No limit a 's to source material had been
imposed on the specialist.,, some of whom even extracted
or prepared abstracts fromhthe New York Times, d
In the mid--1960ts (exact date not confirmed) all IBM
cards with the contributor code were pilled and destroyed
they added hothing to the Intellofax file but instead created
document retrieval, difficulties.
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 pproved For'Release 2004/09/23 : CIA-RDP84-00951R000300040002-5
(6) cables
25X1A9A
OC1D management earJr in the Intellofax history
decided that cables and telegrams would not be considered
a part of the central reference system for stihject/c rea
retrieval. There as no question as to their current
intelligence value' but CCD did not consider them
vital to rgtrospective searching and therefore felt it
was uneconomical to index the enormous cable flovro
This philosophy carried throughout the years of the
Intellofax System, although there was a brief
,r- a cable indexing experiment in 1963/64
in his November 1966 study~e titled
"Choosing the Director's lentral Me S,, `?e rt=F said
Thi most ioteworthy Pmtsatom to th
of ma j ar intellig ce input is not th
deffciencc it might,at first sight 'App
substantie content f most cables Is
valueeo. . j.The curre t intelligent
particular
in h
di ort ret
The su1 jec
technic.
He could
eval time
will des
deasible
file holdings
his omission
s erious
ar. The
f transitory
i .rst is
t on keeping c,ablbsclose at hand
ng the usefulintplligence life of
not risk r eliance
that current int
rve reexamination
and economic ly
analystq to l interroggte central f
query stations and recover needed
o4 OCR for the
Uigenve requires*
'a4en it becoms
on remote
nts wiyhin the
time frax I }e In which/current intelligence
cument Division'
See his ter ' t e0
6
25X1A9A
operates,
GROUP I
Excluded from automatic
downgrading and
declassification
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 CIA-RDP84-00951 R000300040002-5
(a) T6chln auapi~t -
(11 The Batch System
In August 1~r8 OCD management decided that a
review of the e;{'-i sting ?contemplated document handling pro-
cedures should be made with a view to determining if such
procedures could be modified to e xpedite the distribution,
machine
classification., and indexing of documents pending the.
25X1A9A fulfillment of T/O requirements,,
25X1A9A
and Management Office worked closely with
of Administration
in setting
M. a detailed procedure for the use of multilit; Bats in controlling
and indexing intelligence documents,, This was the beginning
of the so-called t'Batch S'ystem," which was modified many
tires during the next 19 years
A bitch usually contained in one envelope 15-20 documents of
like source. (This was similar to the organization of the
dissemination function in the Liaison Division. The classifiers
were not organized by source breakdown until April 1952.)
Put into operation in December 1948, the system
included the complete processing cycle of dissemination, distribu-
tion, indexing, key punching and final filing of the documents in
the Library document files,. The 'bibliographic information for
a given document was typed on a multilith mat,which had on the lower
half Es a pre
-printed distribution ladder or dissemination points. (See Figure
This mat was the basis for the preparation of control slips (a
cutoff IBM card) to be attached to each distributed copy of a document.
__One ontrol cord became a source card (see page 15
'.Analysis of atch stem, _ aep 9 o an lob 50,, C4
(in File : Management Office 1949-510 8-98 j))
vN ,~ M2 81PQf.8' tMQ031~Od ( 5 C. in
Machine Division 19t~.7 Job: k /) 5-8 -x$`13
Procedure for Use of Multili.th Mats in Controlling and Indexing
 Intellof ax-page -4q,
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
all the documents in a given batch were recorded on one shes
The classifier used one of these control slips on which to
write the necessary codes for ultimate key punching. The
typist prepared another multilitja mat to be married with the
punched IBM card, In February 1953 a Code Sheet was
adopted for the use of the classifiers; thus ! he codes for
The duplicate preparation of multilith mats continued until
instead of on a separate control slip for each documentp
1956? when a revised batch system eliminated the
preparation of control cards for distribution purposes.
In September 1919 each classifier was assigned an
Intellofax stamp bearing his individual numbero Used instead
of the classifier's initials., the stamp was affixed to
three places: (1)=s-the face of the document to Indicate that
indexing had been completed; (2) the control slip for
the codes, so that key puncher could question a classifierp
if necessary; and (3) the Batch Control Sheet (see Figure
that stayed with the group of documents through the various
processing steps*
As the Intellofax System grew, it became more involved
and procedures were constantly revised and 44 of y impro edo
MD and Library personnel worked hand-in-hand in developing better
and faster methods of processing the document flow and in 1~r ~!
for every exceptn~
taking care of users' needso A procedure had to be written
ft
-:? See history of dissemination in chapter - -t1 Foct Vj)t
Divisio
n
-Lachine Division Intellof o Procedure`s 1949-55 (in CRS
Files
Approved For Release 2004/09/23 : CIA-RDP84-00951 R 03000 2-5
 Intellofax-page 36
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
For example: Extra IBM cards were printed for a number of
offices- NTop Secret Control so that that office could set
up its own Source Card File;
0
cont"I or every T,AT
document -mss so that the Intellofax cards could be
on contact card ( this procedure
u
continued until 1967); Reference Branch of the Library for every
finished intelligence and basic intelligence (encyclopedic
type studies) document for setting up a cumulative index by
subject, area, and title (this stopped in 1953); and 0RE/ORR and
OSI offices that were engaged in the abstract program (see page
*, A special procedure was written for loan documents that
were to be microfil edo If more than 11 subject codes appeared on
see Figure
the control sl.ip$ the classifier wrote 'tMATS" on the Batch
Control Sheet opposite the CIA control number in the 'tcoded" column
to indicate that additional Intellofax cards were needed,
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 ~.rr.rti r. m
Microfilming
q Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
(2} ;T a Microfilm Program
One of OCDts problems was that-of keeping the Library
operating at a maximum peak of efficiency, The Intellofax
System had made available a greater volume of library document
references to a larger number of requesters than had ever
before been possible. The increased output of the System
had resulted in a corresponding increase in requests for the
documentary materials referenced. In wishing to offer maximum
service to all offices, the Library was faced with the dtlima
of coalescing two variations in point of view as to these
services. On the one hand some of the Library users insisted
on an inviolate set of documents in the Library at all times.
On the other handy some. insisted on the availability of
library materials to their respective offices_upon demand.
In answering this criticism by an ORR ana rst: >~ I wrvt-el 25X1 A9A
We clearly recognize the need for ensuring the
availability of a master set of d ocuments;
howevers keeping an original document collection
poses problems of filings spaces circulatiornf
and refersmse which are almost Vrerwheim o
With 617,562 intelligence documents on file (but not
all indexed) in the Lyrrary by September 1950.9 space had also
become a serious problem,* There was no doubt in anyoneis
mind that microfilming was urgently needed*
Mamoy Librarian tol 6 April 1951 Excluded from automatic
SI downgrading and
subs Techniques and Methocuo d
eclassification
U. (in Library 1949-51
File: 58--98/1)
Approved For Release 2004/09/23: CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
25X1A9A
25X1A5A1
in March 1950 the Library began experimenting with a
microfilm and print procedure$ and by mid-1951 it had begun
to microfilm all si le-copy material (10 percent of the receipts)
on 35mm reel film.
MD and the Library cooperated in developing the best sort of
system to solve the document storage and retrieval problem. In
Jiuary 1951,
and his Deputy,
examinee 5X1 A9A
equipment at where
16nm microfilm was mounted into an IBM aperture or window card,
This system allowed each document that was microfilmed to
become a separate entity in itself and not just part of a reel, as
was generally the case in most microfilm applications up to that time.
25X1A9A In October 1951
25X1A9A
problem was urgent and proposed that the I.d.ary microfilm all
incoming documents, keeping a copy of the document as well as
4FY
ofa.1ra' he latter would be available at all, times both
the micr
for viewing purposes and for reproduction in cases where the
requester wished to retain a copy. On 19 December 1951 the
PCrn.
authorized "microfilzring of al? signifi-
cant incoming intelligence documents" and approved funds in the amount of
$16,700 for the initial purchse of equipment and in the amount
$62,2885 for the annual costs of personal services and supplies'
.Memo, C. CIA I brary to AD/CD.9 .20 July 1951, sub: Status Reports 19yy-
CIA Library Microfilm Program' Co (in Library 195L
File : 98' I
) Memo, G,CIA Library to AD/CR, 16 Oct 51, sub: 'UTA Library Services'
So (in Library 191i9-51' Tile: 58-98/1)
mo, to Acting DD/A,. 15 Oct 539 sub : Microfilm Program
of the Office, o Collection and Dissemination' So (in DD/A 19-54
58-98/2)
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 INTELLOFAX*PAC 39
Microfilming Approved For Release 2004/ P84-00951 R000300040002-5
44041~~ from the Deputy- Director/Administration's
Organization and Methods Service surveyed the proposed OCD Microfilm
Project because of Jurisdictional questions raised by the Uffice of
General Services. In her final report a October 19 3
A
X1A9A
recommended that the personnel required for operation of the system
Ua~J~
be on the -- T/O and under t supervision She outlined what
the proposed system would achieve by 1956s immediate film inspections
immediate customer service on the entire collections el3minat ,tracing
of documents and repeal borrowings from other agencies, about 9(3
econor in apace and filing equipments better utilization and
conservation of personnels pera.ft redact -on in size of researchers'
files and expedito chain routing of document Ithe
mid. 42 2bove)
Approved For Release 2004/ DP84-00951 R00
GROUP 7
Excluded from automatic
downgrading and
 Int e1,ioiax .L.[J
Approved For Release 2004109/23 : CIA-RDP84-00951 R000300040002-5
Research and testing ctntinued throughout 1951, 1952,
and 19530 At a CIA Budget hearing in October 1953 - rn Andrews stated
that ND's prim job was to keep pace w ,th new developments all over
the countryp'but that the most important research then at hand was to
develop a microfilm processer that, in conjunction with the Inte]lofax
equipment, would ensure that the Library would give to the
abaly:t the documents he a sked for.
At no expense to the Agency, but at OCDts suggestion,
the undertook the development of
a new machine to r eproduce prints of microfilm negatives mounted in
IBM??Dilmsort (aperture) cards. Because no ollhe chine existed at that
time that could perform the necessary task for OCDts microfilm project,.
Dr. Andrews requested an expeditious procurement of the automatic
one-step reproduction .equipment (Photostat cpe#tgr) at a cost of
approximately 3,950oT-
Microfilming of incoming documents began officially in March 19511
with the following equipment: two Recordak Model D Planetary microfilm
mameras for the purpose of making initial microfilm reels, three Diebold
film processors, one Microtonics Film Printer for copying the original
reels (one copy of the reel was placed in the Vital Documents Repository) y
three Filmsort "mounterd" to cut the microfilm reels and install sh e
frames in the aperture cards, and one Photostat Printer-.Processor to
make positive prints from the aperture cardso
/O.C.D. Topics to be iscussed in CIA Budget Hearings, 22 Oct 53. S.
(in Chrono 1953. 58-9810
'
I
Memo, AD/CD to Ch ef, Contracts Branch, P & SO, 20 Nov. 53, s tb a Re..
quest for Purchase o otostat Copy--Card Filmsort Type Machine.%
Api*AI2d or eQeas19 g 23 8CNIMP84-00951 R000300040002-5
 Jnte.! B orax-L i.
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
The aperture card (see Fie was a punch card that
had information identifying the microfilmed document printed across
the top and 16mm film images of an intellofaxed document mlunted in
apertures (openings on the right-hand side of the card). Aperture
cards varied in that a single card night contain one, two, three,
or four apertures. Each aperture contathied a maximun of two film images, each
image being a page of a document* The basic procedures of microfilming
documents did not change materia]J,y until 1968 when 3~k= film replaced
16mm.
25X1A5A1 A screening committee of librarians and machine people)
chaired by of the Analysis Branch,yras established to scan all
documents at the intake point and to work up proper procedues. The
first documents microfibmed in March 1954 were State Department despatches.
These were followed in April by Air Force reports (from an original ozalid
copy, which had to be returned to the Air Force within 48 hours); by
Army and Navy reports in July; and by the last segments-CIA reports--in
fleptemberQ
With full-scale microfilming in effect? the Library and MD
decided late in 195 not to microfilm NOD & documents because their con-
tents did not meet indexing standards. In April 1955 this decision was
amended so the microfi] ing would occur only for those NCD` S that were
single copy-, required further routing, contained enclosures, or were of
CIS: originj thus ensuring an inviolate copy in the Library, In the s t by.
step batch procedures established for the flour of documents, microfilming
* Detailed procedures of microfilming were outlined in CTA/CR 25-3. '
5,01 CI:A/CR 25-3, December 1959, szA-c Refg nce Aid on Machine Support
7-~ Z..= ""'?
ScrErLcass C. 6n CRS Historical Filn4
.
Ttsellofax Chronology -Abpendix
Approved For Release 2004/09/23: C4A-RDP84-00951 R000300040002-5
 " L/
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
occurred after indexing: so that N determination could be made
I O~x
first, This had one big disadvantage in that the microfilm of an
indexed document was not an file until after all other processing
had been completed*
Print service from the aperture file became the responsibility
of the Circulation Branch of the Library where they were filed by
control number assigned to the document. 35mm reel film was used for
documents over 50 pages in length. Bulky and oversized docurnnts were
not microfilmedm Any equipment developments or problems were the
responsibility of ND# such as experimenting with improved aperture
card positioners for Filmsort viewers and with better methods for
printing documents from the microfilm viewers.
See page 45 - for microfilm designators assigned to control numbers.
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Intel lof ax-45
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
(3) Machine ?rocassir
After the microfilm prograri was implemented,, ND
inaugurated a system of using work cards to prepare LWellofax
cards and aperture cards. A work card was a bunk brown punch
card. For the preparation of Intellofax cards, YD punched
one
(from the code sheet) one subject and gain area code into each work
card. It w as possible that several work cards were prepared for a
single docttient-? the average being four.e The-I loi-t o=' :e.
When t1 e printed white TatelloF'ax cards (made by PSD from the
multilith rat) were received,, the punched codes were transferred from the
work cards to the core: ospondi ng
L ru. I nitt el ?oSsyfvVaxcarde s,
Off es uti -tt i - e
containing the bibliographic citation. After the above processing
was completed,, r) used the Intellofax pork cards to prepare
statistical records and eventually sent the cards to the Vital Materials
Repositoi7 (,121) at the rate of $66,COO annually. These cards 1~
could be used in ~~emergency Visas- to reconstitute the
they corit4 ed
;xr' the Intellofax File $ but .,,only the punched information and not
the bibliographic citations Ono work card was also prepared from the Batch Control Sheet
for each microfilmed document, .Coded information (docur nt control
num;er,9 date of fil ing,t number of pages in the do3-ument,y the
sequence miirber and the reel number) was punched info a work card
first
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 IntFellof x-)411
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
Consumer s Prsric e to
rather than directly into a blank aperture card because it was
more efficient to punch one work card for each document,, select the
number of aperture cards to cover the total number of pages in
the docur;nt and then transfer the punched codes by machine from
appropriate
the work cards to the,,,ipperture cards. The document control number
printed across the left end of tho-&p each aperture card
faciliated the manual sorting of the cards for faliai; by document
control. number in the Circulation Branch, Libras
was provided on tb- intellofex rile in the form of Tnteilofax
tapes ,tea e = o The tapes turpre lists of bioliograpitic
citations on the indexed documents., covering subjects, areas.,
dates., seciril;3r classifications and, other details specified by the
requesters The tapes were arranged in the sequence specified by
the requester--that is, by subject., area., itIVI source, or dates
unique
The Tnt?,llofax tape was prspared by the47Cfatsrile process descr taed
`F b
1, , lL"y '-j++ . N+ Y F.Lf.U+.s. .1 BI.bw~ (iL. LF.~ ozi gage , ' u:iu l 19519. developed nnhou&a much
faster%Card,L st,Canera. By this machine the printed data on the
cards were photographed on a I.-inch strip of photo-.sensitive paper]
at the rate of 3,600 cards Ter hours r:e a a om the cards
The tape was then inserted in a modified Photostat Expeditor (had no
exposure xanit)., a machine that developed, fixed.,, washed, and dried the
tape in a singl operations The machine could process tiro tapes
s7 multaneously--350 feet of single. tape hourly-or a riaxi um of
700 .f^~ f tape hour].,.
Appl e r Re~er 2VU4/?9/23 D 8 (~51 ~3000~4 00 - a, t~ j
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
f. The Source Card File
An important by-product of the Intellofax System was
the establishment of a Source Card File. The Library early
recognized the need for a card catalog of document sourcesy
similar to the author catalog for books. In the first coding
efforts of 1948 prior to the Batch System., the classifier wrote
the ISC codes on a 3"x5" form on which the typists had typed
the bibliographic data. With the inauguration of the Batch
System, I,?) provided the Library with a "cut" Intellofax card
or control slip for the Source Card File.
This File served several purposes: (1) inventory of
document holdings., (2) identification of a documentsnci 131 loca-
tion of a document. Requesters looking for a specific document
often did hot have the document number but might know the source
and/or year. The cards' were filed by source of origin with
further divisions by year of publication followed by specific post
or agency breakdown and document number. A brief title description
of the enclosure and a notation whether it was received or
no jreee_4~;Q4 microfilmed or notm d, appeared on the
card. After the inauguration of the microfilm program in 195Li.,
the approach to the aperture card was only through the
document control number that appeared on the source card:
cure 4' control number-on 16mm aperture card
'tC't + control nut.ber---on 35mm reel. film
'IV" + control number--not micro t'ilme d and in hard copy
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
The source card that was prepared for-111 4 5 contained
only an abbreviated bbiliographic entry., that is., sources
document entry., date., and security classification, The ti:ble and
country were not entered, This abbreviated notation saved typing
time., but it created problems when librarians searched for document
identification, By March 1961 a full source card entry w as made for
all N P> that were microfilmed., the so-called. Pt#T'-M category---
primarily CIA documents that w ere nedexed.
For further discussion of the Source C . rd File see
25X1A2G page I_._, the /xnpact of
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Y7
.11 Ic.
ge The latelligence Publications Index
p,v
25X1A5A1
Equally as important as the Intellofax System in
the history of OCR's document and indexing s chemes was
the IPI0 Because the !PI was actually a corollary to the
:Intellofax Systen, its history is dealt with here*
Although the lute?? ofax System had included finished
intelligence documents since early 1949 with a special
"fin intent` punch added to the IBM card, OCI made a strong
plea in 1951 for a separate printed index of finished intelli-
gence studies? The request was all-encompassing and included
intelligence studies and featured articles in intelligence
,periodicals, up through Top Secret, published by Intelligence
Advisory- Committee (IAC-successor to TAB) agencies, certain
organizations ? ..
25X1 C8A
subordinate commands, and intelligence
Library personnel made trips in November 1951. to the
Near York Tim s and. to H.W? Wilsom Co. to r eview their x?espective
Issued in October 1952, it listed as its purpose:
procedureso was recruited from Catholic Hniver,! t,,
where she had been involved in the preparation of a similar
Readers' Guide type of index. One of her first tasks as chief
og' the Editorial Section of the Book Branch of the Library was to
prepare a prospectus to the Index to Intelligence Periodicalso
To e stablish a current., continuing, curaulative,a
subject index to articles and studies contained in
a selected list of the more important intelligence
periodicals heretofore not con red by cumulative indexing
Frequency was to be monthly, with semiannual or annual cumiAations.
A full. set of the IPI is kept on file in the CRS Document Library
4 WA4~A25 ,sub: rip
I ~ qdCFp%W ~~ 0 ~ P MC ?AR
to New York, Nov 1.3-114, 1951a Jo File : -
14i 8 Lb 01. _A Q-Q") Z473~
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 q~ r ved F r Release 2004/09/ 3 : CI,- RDP 4-0,09551 R000300040Q02-5
Reactiono the proposeddex throuou gencya.,
the intelli enee comnrznity was most favorable, # 'valuable
suggestions received on coverage and format.
Issue No. 1, dated January 1953, was entitled
Intelligence Periodicals & , @ I - changed with the November
issue to Intelligence Publications Indexp when, after much
urging OCI, finished intelligence studies were included,
I
The IPI was a
conventional library index, follawing".American Library Association
rules for bibliographic enties. Arrangement in the enlarged IPI was
by subject and area in one alphabetic sequence. The IPI staff
opted for a sem nnual cumulation as easier to prepare than an annual.
In January 1955 an important change in format occurredc
The index was arranged in three sections: a classed subject index with
area subdivisions, thus putting related subjects into a compact section
of the index; an area index with broad classed subject subdivisions; and
a list of series by source and titled The area subdivisions of the
subject index, the countries of the area index}and the bibliographic
sources of the monographic series were arranged in alphabetic sequence.
-- ----- - - _). Beginning with the January /?S1)
issue, a black border appeared on the edges of the area index for easy
identification.
The security classification of the IPI changed several times:
it began as SECRET in 3.953; from November 195355 it-vas CO'IFJDIEMTIAL
f
Exc1 ed famatie
GZR".
oy'rn . d
-dassifie6flon
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 p ~ b
verl nr Release 9nnd/nQ/23 ? r_IA-RI,P6AII'N1Q51Pnnn-annnA nn9-5
1tlN 1' /1"Llj MrMr- Llvlu' ..LJ;JV -Wr 1.V WCta L)' LLJi ltlv v1LLQ? 9
device h ttshingledn' typed cards onto a large board. '', " },~
An IPI issue was assembled by means of an aligning 11""- I
l
The assembled boards were returned each month from
Printing Services Diviside aauxi the cards stripped from
the boards to be interfiled with each monthly group of cards,
ready for the se , nnual cumulationo In 1957 a Justo-
writer Model L Recorder te, a Justowriter L Reproducer and a
)
Line Fender were procure". A punch tape containing a record of
what was typed in the form of a code for each letter and function
was inserted in the Reproducer and run continaously until 10
cards, a standard number to take care of the.r
three indexes, were produced for each entrya Justified
composition was produced on a continuous strip of perforated
card stock. The arrival of this equipment made possible the
assembling of a monthly issue for printing within the first
week after close out date.
Beginning in 1960 the June and December monthly
issues were omitted) and the entries of those months appeared
with an asterisk in the respective cumulations. This saved about
$460 a month in printing costs and provided for faster
publication of the cumulationso
Document coverage for the IPI evolved throughout the
c
Rules for f1oe7 and
years as it did in the Inte7lofax System*
selection oriteraa were made simultaneously with those for the
Intellofax Systea. The c.,-sequence of selectivity
.a~wneroam9 aid
atTecle5sifi284~$n
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 5
Approved For Relea M /2 ,pQUksRJP O25469N&%NO4&QRCO oEzr
TheNational Intelligence Surveys (NIS)
(see Appendix t5), were indexed from 1953
until 1958; classified US research and development
fv/7
publications were included from l957 ,r58; PD]) publications
were selectivity indexed after 1955; selected non.
NPIC photo interpretation studies were included after
25X1A9A
25X1A9A
25X1A9A
1960, In 1956 a request to index the lI Daily 25X1A7B
Reports in the IPI was rejected as not feasible, be-
cause they did not fall into the criteria of finished
The
intelligence, (IPI staff was constantly making decisions
as to that constituted t?tin intell,") At the request cf
the DD/I in 1955j, the IPI staff began to provide citations
of all new monographic intelligence, including the
NIS's, for listing in the back of the OCI Current
Intelligence Weekly SUMWr7*
3 ~ $ { ytiwrub X1 1 K , zL.,v. 4-to had been designed Gtez a six-character co o The seven major
chapters were: 100 -Governments Politics, and International Activities
and T~atitutea, 2OOZo-vial and Cpltnral Structure and Institutions,
300-Zeience and Technology, Engineering, 1j00- Commerce, Industry",
Finances 504 Transportation and Comiimunioation Sy?stems, 600-Resources,
A
Commodities, Weapons, 700--Armed Forceaa The combining of military
71,/CODIB- 13s 9 June 59 (in CODIB 1959 64-341/2)
, /Intelligence Subject Code 1960 (in I511 Historical Tntellofax Files)
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 6(
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
organizations and activities in chapter VII and the
establishment of a single list of military weapons and equipment
a
in chapter VI simplified the use of the ISC by the i].itary
services (and was actually the Air Force's sugges7ion)o
Eighty-nine modifiers of three digits each were available
in combination with the subject coded The Introduction ex--
explained that these were "a facing device which can be combined
with certain subjects to mpeoify actions or states which affect
these subjects,"
173/
The Area Classification Code had been prepared under an
25X1A9A
irman,
rl~cing Group, of which was chairman,
AB IP/ ODIB o
The Code was designed to be used with a six ndigit hierarchical
? digit non9lierarchical. numeric notation (for CIA primarily: if the
Intellefax System continued), The Code was organized into nine incisor
area divisions, The last position of the notation wee called an
"indicator" showier either political affinity. or a geographic
unit. Far ewe Communist China was ACH00C or 1013; Wept
Africa, AOOOCW or 1006. Only USSR and CmmYninist China were
or provincial
subdivided completely to administrativeylevels,
requested by the
alphabetic notation ( Air Force)
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
25X1A9A
senior classifier
25X1A9A
normal disseminaticn cmVleted. is d ocunent _waa, the~ ~``afore, net
(cozwers1on 'took time)
available in the aperture and Intellofax files for 8-9 weeks as
(1960 and 1961 respectively) increased processing time and
manpower in CIA to orenar9 Air Force reports for Intellofaxo
Ible Air Force sent-the 'coded document (for conversion) almost 6 weeks after
had been
1f
Ewen bef of a the final acceptance of the revised
ISC, CIA and the Air Force agreed to exchange their coding
effortso The former sent its Intellofax code sheets to the
latter). Air Forse's "Okw code sheets were accepted for
conversion into the Intellofax Systemo This exchanger enabled
to the Of,
the Anea]tpsia Branch cease coding Air Force M's and td free
classifiers to support the iD testo it
also set a precedent for cooperative processing in the intelligence
April,
7 ~
commlunity. Beginning in June 1959, this exchange lasted until 1962.
CIA's the Air Forces
Subsequent changes in the Intellofax and, 2I 1 ARD System
opposed to 2-3 weeks for other sourceso
Division _
The Docw.ierrt $ ent its senior personnel, o von
in 1960,1961, and 1962 to train
request, for short periods g service personnel in the use-
of the. revised ISC s who had been responsible for conversion of M code sheets to Intell-
ofax) to the Air Force Strategic Command in Omaha, and
to US Air Forces in Europe. Nine separate training sessions
were coducted at CIA headquarters fob Air Force, Army, Navy, and even
State personnel,
,~ ~/.Y!! 7V ~ .l/ l CIIC r . ~ bv'G7
M
G p~~w l~r 2 -~
7l~ -CGS _.
App o o~`f sl asgg 4
tvi~ ~, R4 7 Y2 : P 'x DR8 - '9 1, Dfa
P
 Intelloffax-page 63
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
(11 Purpose and Developr e rt -
The Machine Diction's charter to keep abreast of the
25X1A9A
research and development in machine technology led
0
25X1A9A
25X1A9A
25X1A9A
(with the Management Staff since 1950) in 1952 to investigate
confidential new equipment being developed by Lastrnan Kodak for the
Air Force* Wit s1gnr en in September 1%5 to Join the staff of
Eastman Kodak, his successorl ascontinued
the close rapport with the Air Force in devising compatible methods for
joint Air'&C I A use o f a m u l t i , - - ; m i l l i o n d o l l a r system k n o w n as M 3'A Dp
Four and one-half years of effort in overcoming problems in opties,
photography, mechanicsp and electronics ensued with Eastmano OCR's
M TQA'RD Project wall OCR's biggest system design effort in the
office's history since the innovation of the Intellofax System?
From its inception in 1917,
the storage and retrieval capability of the Intellofax
System had been increasingly strained by the flow of information until,
by 1955? storage, retrievals and costs problems were considered urgent.
The increased growth of the film bad been accomplished by multipli-
cation of IBM equipment rentals, storage units, and personnel* The
rationale for CIA's interest in. M~ was spelled out T Dr.
Andrews in a 1955 memorandum. to the l ro ct Revie~x Committee'
He recommended a contract for the purchase of M Ai equipment at an
estimated cost not t o exceed $330,000 in order to c onduct "an early
and large scale test of .. odata handling equipment. ..which it is
believed capable of substantially improving CIA's Intellofax Syetemo":I
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
/~1 aop AD/CR to PRojeet Review Committee 25 Aug 55' sub: M
Project. Co (DD/I 1955 ) e -l3~
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R00030004,0002-5
Eastman had developed an iAformatiorA storage retrieval
systan utilizing the micrephotographio medium in the form of a
piece of 16 mm x 32mm, 04led the 1N? 94 t D. (see Figure 7)
Eastman .claimed that' this provided for the first time a storage
and retrieval system that could be searched electronically and that
combined the mobility of punched cards with the compaetnese of
microfilmo I' MO 't)S representing millions of pages of material could
be machine organized in one file cabinet In such a manner that
rapid retrieval was possible. The PARD System called for
the retrieval of document images either for viewing or for the
reproduction of hard copyo Retrieval would be by the same
search parameters as in Intellofax .b j ect and area codesp source.,
date' and security classification* eaatase~-the '~s s' i
a nwimt `oi``fi~zcte sr`~ of ?lixzari` Ui tati-OW, -with I Bid (Wid '
the ?~~oulc~bee' or-a-dacanuant~. bat~I1.- ate- their Ar-Fvr
hoped-?that-.e 3.derarrle ~1eri~h~rYess'~ cow~].d, hs_.acha~d byr^rotzs
OCR anticipated thht M j& A RD would hasten and expand
adoption of corsrmon data handling procedures throughout the intelli-
gence cormiunityo The Air Force and CIA maintained a continuing
degree of close cooperation on an almost daily basis with the
mutually agreed objectives of achieving maxi.mu m compatibility' as
a result of which thgr both anticipated the exchange of M GM &S
in an ongoing system. Because there was no limit to the amount
of index data (no 8O-columh limitation as with the IBM card) that
could be used for any doeumehts, both CIA andthe Air Force hoped that
fi~,rn
Approved For Release 2004/09/23.: CIA-RDP84-00951 R000300040002-5
 ~pproved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
considerable "richness" could be achieved by generous
indexing of names of people, organizations, installatior~,
and locations. MD and OCD/OCR managers made frequent
trips to Rochester to follow the developmental progress of
the MIi1 2JW equipment which was in the blueprint staged
25X1 A9A In the spring of 1956 I I was detailed to the Pentagon
25X1A6A
to assist the Air Force at its request in establishing
detailed indexing, plans and procedures, and three members of }ID
spent several months with Air Force personnel in
their stutbr of the experimental equipment*
2 Doubts
The it orce negotiated anccpansion of its contract
with Eastman to enslire the manufacture of a second set of
equipment for CIA he equipment arrived in dtages, but the conc+lete
C,
set was finally delivered in November 1958 in, specially air--
conditioned rooms ca the ground floor of Installation
was completed in February 1959 During the period between the
order date and delivery, several major modifications were made
in the equipment. CIA was aware of them but had no legitimate basis
for objection because CIA's program w as appended to the Eastman
Kodak/Air Force development program. The changes did, however,
invalidate the earlier space and personnel estimates. Operating
speeds on the duplicator, sorter, selecter, and process6r were
substantially reduced0 By the time the test phase arrived,, the
1,1
.ZY ' /Memo, AD/Cf to Mai Gen Millard Lewis, DDT/USAF, 3.7 May 55#
sub: CIA Participation in Air Force's MIIi CARS Contraoto a.
(in Chrono 1951-55 ~1) //
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 4
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
earlier expectations from M P had been somewhat modulated,
There was some feeling in CIA that extreme miniaturization
whit e] iminated manual access might prOTe inferior to the
16mm aperture card systeno The combination of codes and images
equ prlcnt with the _aa^abil?h ;
OCR's tasko Still prior to the tests
senior classifiers expressed the thought that subject coding?
the slo est phase of Intellofax, would be even slower for M QAfD
because of the increased coding complexitieso The pros and cons
already,
continued, but the final recommendatiii was that hav1n;hmade a
substantial investment in M COM and in spite of the rsserrat s
that had developed? OCR sh&uld male its o r ct evaluation and
Test: Taut and Retriev4
in the saws card w as questioned, as was the loss of the bibliographic
Intellofax tape, The changed equipment specifications stimulated
OCR.ts managers' belief that additional purchases would be required
at-an estimated cost of three to five times the cost in the
original orders with an estimated equipment delivery of 1-2 yesre.
Despite the increased pessimism, M, though untested, seemed
to be the only alternative to 'A?4
of handl in; -
A. test of the equipment and of several systems apppica-
tions began in January 1%9. In the meantimaj MD had been photographing
n
codes and the related documents for about 64-
7* o
f the CTAI
information reports that had been pi;c.ssed into the Intel.lofax System.
25X1A9A i . _l
s esker-ef1managed the
era-men
emo, DAD/CR to AD/CR, 22 Dec 58, sub: M TCARD Testo_C
0003X4000
 +r~a}j
. 'i O
o~ a q>
4)4)4
to ra W
~~oy
6?
Approved For Release 2004/09(p, SZ A-RDP84-00951 R000300040002-5
As of Ap*i1 1958 there was- an e sti sated 13,000 master MIfboAl I
in the special file mechanically converted from IBM cards,
The full line of rjofC equipment was operated parallel but
separate from the Intellofax program. AMli~ Project Test
Mannal. (dated January 1959s revised August 1959) spelled
out the objectives and stated that the test was intended to
find out the quality of the systems, as interpreted by both OCR
and its customers and as compared with the Intellofax System,
A test corpus consisted of a copy of each document
distributed during January and February 1959-u-approximate] 25,000
information reports containing a norm-a mixture of document
categories by sources, format enclosure variety, and with every
kind of problem in subject coding and ph6tographys, including
paper types, inkss, size W"color differencess, and varying
legibility? The documents were processed into both Intelofac
and 4lXT,AN),,with codes and document images stared in separate
but related M#TXP D'So (The Air Force had already experienced
difficulty in placing codes and document images on the sine I I A~RDd}
A coding variable was the use of the revised ISC (the 1960 ISC in
draft) for the Mt,AD test and the old ISC for the Inteellofax
input and retrieval., (This also provided an unusual chance to test
the revised ISC. Several weaknesses ware corrected before the final
printing.) This test required the services of 18 staff members for a
period of approxnately o set year, The indexing staff;-made up of
soma of OCR's most e xperienced people ( four classifiers from
the Analysis Branch and two from the--Speelal-Regis-teas pus one
reference librarians, operated in spade contiguous to the Analysis
25X1A6A
Branch in th gall supervision was under tbo
Approved For Release - 84-00951 R000300040002-5
7W A mo, DA.D/CR to Indexing Staff (by name) $ ii Apr 60, sub: Commendation,
MR) Ch->r'ek An A),-, On Asi, i/._ z vdia
 page /6d
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
the aegis of OCR's Automation Development Groupo_ The
significance of the test lay in unique MARD coding
techniqueao This included rules for linking codes in phrases
and for using clear texts b ;of which wr4-aat possible
in the Intellofax System at that time.
The retrieval test began a year late in February 1
9(}
after input was completed. .,-test questions included both live
and simulated requests suggested by retrieval test personnel
(two groups: the Composite Intellofax Group and a M 6ARD Composite
Group) aIftd by other OCR$ other CIA, and non-Agency personnel tdto
were Intellofax users. Also involved were requester interviews
to define and redefine questions and to check customer satisfaction
with content and format.
)Findipgs and Conslusions
The test clearly demonstrated that subject control of
information and . tbie pro ei .tme a -tployedrthere Q` were the principal
areas to cohsider in the development of a successful system. The
overall retrieval tests showed an appreciable qualitatIre advantage
for M ,r,G,D. Both systems combined retrieved 997 specific
references of which M CARD obfiined 788 or 79 o a the total
and, Inte7lofax obtained 6449 or 65f,bf the total. The retrieval
groups concluded., however, that this 'superiority was attributable to
25X1A9A
ADC was organized with from MD as chi of in June 1958
after Task Team No. 32 on MIWTCARD 'rea mrrended that OCR should have a
continuing high-level. planning and management staff to Investigate new 25X1A9A
ideas. This staff reported directly t? 'a Team No l2/4
in Task Team Reports., 1958. - H tom; ; cRl =, ~ n ~ 3 -
S'~pe al=Reg 4at had for-nanjr .,rears in its indexing - stem been using
clear text for abbreviations of or an3.zatio
Approved For Release 2004/09/23 : CIA-R P8 -00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
the coding and procadural techniques used and not to the
IY
The Chiefs ADG submitted a detailed study outlining
equipment and personnel utilization for the 1 C, tD test
project and requirements for a fu l-scale MtiI4. D System. H.
recommended that the MINICARD equipment not be used in OCR*
LJA 1/-,%Z? in his final report to the AD/CEiave the following
reasons for rejection ywith which P' . Baxel concurreds
aoThe retrieval tests clearly showed that any
qualit4tive advantage of MIlICARD over Intellofax
could be introduced into the Intellofax System
with only minor revision of camm" routined and at
a minimal cost.
bo A substantial standbyequiprnent woun be required with
MINICARDa
c. It would be necessary to operate the MOARD ahd
Inte llofax in parallel for not In s s than 7 years
s;eIntellofas could not be converted feasibly to
M C4BDo 7
do The technology of storage and retrieval of information
has advanced so rapidly that *"t -be a "` Mfl IPJRD,
even in its latest model, ,;to be obsolescent.
e. The economics of the Mt RD System are prohibitivb*
f. Personnel required to operate t h e M t Y I C A S
would number 47 more people than w ,y usto
operate the Intellofax Sys temm,not including the extra
/(Memo? C,ADG to DAD/OCR.. 1 Apr 60' sub : MINICARD -Test Project, go
(in MINICARD Folder V -,3 yI
Co (in MINICARD Folder R I) j c( - 3 y/
Memo. DAD/CR to AD/CRS, 8 Apr 60 sub: Final Report on MINICARD Testa
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 OApproved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
analytical personnel demanded because of the improved
coding technique.
g. The original space requirement was of the order oZ
3.00 sq feet per MINICARD installation, .endings
indicatttthat 3#007 square feet would be required for an
insta3.1a tion of the size necessary to handle ow load DD/S concurree-in OCR's conclusionedirected they
AD/CR to proceed with plans for the disposal of the equipment . ;
and supplies on hanct, /
It was incumbent upon Mr. Borel to explain OCR's rejection of
the MINECA`RD System to CODIB. ft s summary and decision appeared as
follows
Our findings are negative, This conclusion is based only
in part on our findings that the M GrAY D System would nod
enable us to give substantially superior reference service
over that possible with our present system,,*... expectation
of econsies in the Agency from a common community program
for one-time processing., common code, and identical equipment
and proceJres has not come about and from all indications
would not c one about through MMARD, whether this Agency
adopted it or not. Technological advances since 1955 have been
such that alternative systems have been presented and no
other Agenc in the US33 other than the Air Force plans
to use M'IN 1D, One of the aims of CODIB has been the
stimulation of compatibility of systems considered, and this
goal has by no means been reached: to hope for identical
systems is just not realistic. * ,As a result of the tests,
the Working Group concluded that ICAAD did not live up
to what had been hoped for in terms of our own problem,
There were demonstrable advantages, but important disadvantages
were also discovered, Ad a consequence the Working Group did
not recommend a conversion from Intellofaz to MX=k9
but recommended instead the modification of Intel.].ofax to
incorporate as many of the advantages of IIIN'ICARD as were
technically and administratively feasible. OCR management
took these findings into accounts along with the proportional
81/'Memo,, DD/I to Project Review Corp
~ittee, 3:8 Apr 60, sub: M RD, C,
OF UUUAn_V_Z.31.Lp 3v dune 60v CIA/OCR MI ARD Test, C. (in O RCA
Folder
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 1 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
role that machine searches play in the overall OCR
service pictures and the present limitations on
staffs moneys and spaee. These considerations le8
to a decision not to adopt the MQA'D System as a
substitute for the Intellofax System..,,
So 8 years later (starting with the early investigations in
1952) after a tremendous amount of manpower and effort? upended s
_.G!
- iii
adequate printout method was actually a more serious matter.'2
25X1A5A1
ApproQ94I s 04/O~A' I~-R B ?tYt
5. NietF F J& We-W
/01 */,' Memo, Chairman DARE Committee, 30 Jan 64, sub: Status of DARE Proc-ram,
Special Assist t-to the AD/CR , /chairedt i'
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
2. July 1963 for the development of a printoA.t machine that
would enlarge the DARE image to approximately original document
dimensions and provide a positive print by anelectrostatic
1
process. In the meantime, ND's Photostat Expeditor was adapted
and used until the Xerox machine was available in November 1964.
ED's Equipment Services Staff also developed .a Viewer-Selector
for magnifying and selecting DARE cards,
Ftll realization of the potential of the DARE System
required the resolution of other problems, some of which in-
tolved interagency c ooperatio i o Problem one was the d.evelopr nt
of "meaningful't control numbers for incoming documents. (See
page 58 for seven-digit common number s?rstem4) The Defense
lirtelligence Agency (DIA) adopted a 10-digit meaningful number
a:
faith the Agency's establishment in Janyir~r 19530 Through the
efforts of a spacial OCR study group' the CIA reporting components
nit
began to assign a new 10-digit number on 1 July 196)4 State
remained the significant holdout as in the past, and DD
control
continued to essign. numbers to incoming State reports. (State
still. does not comply as of 1973.) Problem two., also venerable.$
concerned the standarization of quality and format of documents,
25X1A9A
See Intellofax ProcedureManual 1964 for complete list of 10-4igit
control numbers, 1 (Intell.ofax Procedure Manuals in intellofax
Historical Files 1-718G)
The effective date was set at I .July 196)4 in order to permit DD/Pts
25X1A2f,rge-scale machine o eration., rroject I to rise its programs
to accommodate the new 10-digit control number,
'~=F The history of -U-iis effort is well summarized in CODD3-D-78, which
cites 37 other COD-109 documents on the subjects ~_l (0017 3 D-78~ 7
r'!ar 61, sub: Common Format for State Departmebt =orei'gn Service Report-
an Related Problems, C. File : COLII3 1962 Job : 644a/5,
IlTemo,, Chairman DARE Committee to C/DD,ID, and Librar_$ 10 Oct 63.,
IQW
sub: Appointment of Study groups for DARE Project, Sc, File: DARE
Approved TI P " 95~.on Reno to
an su on ?*o uin e.rs nor orma ton R or t , S So
File : ~ / Ibid.
 Appr34'd sWaa2bb4td 123t c tt DP84-00951 R000300040002-5
or bibliographic purposes depended in large measure upon
the degree to which the source agencies would provide
for the inclusion of certain elements, such. as enclosure informa-
tion and expanded titles. A DARE Committee- study_...gr on,
3Lnforznation report formats took steps within the Agency to con.
on white paper inseq~d of the traditional ~`3t had been using.
Both 00 and DD/P adopted a commeon formate
The adoption of a USIB?vide coranon numbering system
had accentuated doubts about need for a Source Card File in the
Library. This File had been the subject of special study on a number
of occasions. In November 1962 the Library- proposed and the
Machine aid Document Divisions agree/ to A 190-day test to cbtermine
the Library's need for source cards, After the test period a four-man
Library Committee recommended that coircecards be continued for
all collateral documents received in OCR and be recognized as a record
of common concern to all OCR Divisions involved in the Intellofa-x
system as well as to USIB agenceesso~The three concerned Division Chiefs
(in O 1961!. > Gy 3
agreed with the Library Committee that the Source card File was the
on7,~r "author" file of intelligence reports in the community and.
should be repined until something better ~~t 10
ome alonge
/16 , 11I J) IV-,,
JODIB-D
CIA/OOB Information Repcrtsa OUO
lO % Memorandum of Agreement-CIA Idbrary,ND and DDs 5 Nov 62, Internal
lO:.
Use Only (in Library 1963-63 Job t 65.1a3t3,)
Memos L3brarY Source Card Committee to CIA Librarians 28 Feb 633 sub:
Source rd File. S. (in DARE Folder 68-1187/
f Memo' Chiefs DDslMs and Library to AD/CR9 13 Mar 63' sub: Progress
Report on DARE Equipment in OCR and Related Matters. C. (in DARE
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
to print itq,00-B reports
 page 92 := /, r
I ~2004/09/23 : CIA-RDP84-00951 R000300040002-5
_...r
At the request of the DARE Committee,in may Lybj a sLUUY
was made to determine whether a machinable source card file of uncut
IBM cards should be established and whether document retirement and
destruction information could be incorporated in the source card and
thus abolish the keeping of a "retiremer book" The report was
~r
essentially favorale to these propose In September 1964
IV-
the Chairrran# bare Committee suggested that the Library review the
earlier studies and determine whether the time was opportune to make
arr fundamental change in the operation of the Source Card File. The
Library concluded again nd -just as adamantly as a year before that
the discontinuation of the Source Card File Would have a criPPI-ing
effect on document services and result in additional manpower costs of
15 hours a dayo The Library was, however, in favor of uncut source
cards with the implementation of DAREo In the study it was stated
that s '''The Primary a dvantage the Library can expect to gain, if
the use of the DARE equipment for coded documents is implemented, would
be earlier delivery of source cards and simplification of search
for documents in the process.'*
In September 196J. the three Division Chiefs recommended the
implementation of DARE e several divisional reports set forth the
(feats DARE would have on their procedures, efficiency, and serviced
The chief advantages appeared to accrue at the input with some savings
in manpower and times', , as well as in some simplification of procedures?
The DARE machine tud out on1;}r the specific number of cards
25X1A9A J7 _ 'u161 "i'Iemo IJ to r. CIA. in --a Librarians, 27 Mgr 63, sub: Proposal forpUspeof
' P.211- n fir HART`. FAIiAr 68.-x,.8713)
i
ee F di-
Me4mo Staff Assto to CIA Librarian, 7 Oct 04s sub: Suvsuant
9 _ - ,'0.., _ rN f.1.- - --- -. above)
-- C C ttee to ADJCR 22 Sept 64, sub: Status of
Fdi f r "'WG /0,9CIA~RDrP84ULRO 100 )
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
for the Intellofax file, thus eliminating the wasteful operation
resulting in the destruction of 3 million IBM cards per year.
I
(PM, which was -- -= - > responsible for reproducing Intellofc
had printed a standard number of carder,
cards from the multilith mats
regardless of need.) The Chief/MD pointed out that the addition of a new
file to the already in existencef ?Ait?and '~Bn Tntellofax files added
complications to the processing cycled Procedures for handling damaged
cards would not be as satisfactory as for the non-DARE Intellofax cards
because there was noa one-to-one printing capability that could handle
the reduced images. The anticipated affects of DARE upon procedures and
services in the Library revolved mainly upon the usepf two different
formats and equipment. The Library still had soma reservations about the
two-track system and the negative affect it might have on the customer.0
Anything less than a thorough, sympathetic,and painstaking effort at
the point of customer-contact could well make the difference between
success and failure.
I:l ~+
A Quality Control Committee, consisting of one
the Library, ND and DD, was assigned the task of
member each from -~'
0
and bringing.,,
preparing
together all of the procedural instructions and forms
for theme $
implementation of DAR, the establishment and administration of
quality controls at'selected points in the system, the development of
vidual briefing materials, and the conduct of a dry run during the
week of 26 October 1964.
Z ?te,y~ rC4 v}k~~ % ;tG. 7i~w? r
The DARE System became fully operational for both N9{?.*D and
indexed portions of the Intellofax System
on 2 November 19611, another
important November date in the history of, Intellofax. (November 1960--'7 -P
l.Tt~ demo, Chairman DARE Committee to 1 1 12 July 61i, sub:
"""".DARE ality Control. C. (in DARE Folder 71
Approved or Release 2004/09/23 : CIA-RDP84-00951 R000300 002-5 25X1A9A
 p a ge 8t,
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
25X1A9A
Within 90 days of the full implementation,
wqs able to report that performance e xceeded expectations
"both in econ and efficiency of operations and in customer
11 ~
acceptance." The lapsed time from mail d to Aperture Card and Source
Card Files was dramatically reduced from 2 or 3 weeks to 7 or 8 days
or less and vast being further reduced. N9 items reached the
file in 3 days or lesso Significant savings in personnel were
effected, especially in with the dissolution og
the Typing Section. (Two typists transferred to the Iexperiment;5X1 A2G
three to MD., one to BRs and one to the IPIa) The merger of NO and
index cards in the Source Card File was accomplished and this
venerable file finally became mechanized because uncut (long) DARE
cards were used as source cards, An Intellofax requester~,as provided a
large greener booklet of document-size prints of the DARE cards
for documents indexed after 2 November and a small green booklet
of cut sheets for all references to documents indexed prior to DARE,
(See Figures and 2)
A review oftthe performance of the DART System in September
1965 after about 10 months of operation indicated that the gains in
eeonanr and efficiency reported on in March had been maintained and
improved upon. Customer acceptance of the system and satisfaction with
1.1.E
it continued The DARE Committee-was dissolvedo DARE continued to occupy a
maj or role in OCR ' s document processing until the change to a computer-
based system in 19670
25X1 A,914 j 2 "/Memos 0 to Acting AD/CR 10 Mar 65, sub : DARE System Performances
S. Pir~ DARE Fold 0
l.'t pond faPMeo&t 30= O:SDARE System
P Q ( DAREFId . 68 .9/
o~ce
r
25X1A2G * II was an OCR/office of Computer Services (GCS) program for
developing a computer-'eased central reference facilityo It began in FY
e
.
0 0 In o eL l 7
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
c,~ Egiipment Develo gents (exce tp DAB)
Along with the changes in input and systems design in the
1960's came improvements in the equipment used in the various
facets of the system. MD continued its efforts to find. technological
innovations that seemnC to hold promise fora better systen. Sor"e
proved successful; others did not.
In 1962 MD replaced the Tntellofax Tape, which had beena _-_--
folded tape since its inception' in 19150, with a cut-sheet booklets
The requester's name and address appeared at this bottom of each
citation. This e as to encourage the requester to submit the ci-
ta.tiol7. when requesting a con r of the document and thus simplify the
Library search. It also expedited the tae preparation because the
1
number of processing steps was reduced,
An T3T.I type 108 card selector that operated at 1,000 cards
per minute and two IBM 088 collators that operated at speeds
of up to 13000 copies per minute were installed to replace slower
machines. The card input portion of the Tntellofax System was
prograrrmed in 19614 for an 1131.1 VhOl corrrputer, and this part of the
operation was performed in OCS, producing significant savings in
manpower and faster input. Of particular significance was the extent
to which the computer T---s used to generate the contents of the
-Files and to "exploda11 (create) the necessary number of Intellofaxt
1?.
source, and aperture cards, In 1966 OCR programers added a subject
code validity check to the Iritellofax computer prograaxr,y this eliminated
apnrox irrntclj-, one hour of work on Tntellofax in ECM equipment.
no -
Improving Document Retrieval System, 11eetin 28 Mar 62. FOUO
rile: Machine Division 1962-63 Job: 65-11.3
1 OR Annual. Report 17Z 1961+ Filer OCR Annual eports 19`59--Qt
Approved F~Q44el6?sh~4w/09/23: C14DP84-00951 R000300040D02-5
 Approved For Release 2004109/23 : CIA-RDP84-00951 R000300040002-5
The Photostat Z4?-d--'.t-Ors used for copying doctnrnts :21iu
aperture cards had been in continual operation since 19
extended their maximum efficiency by 1960. The maintenance cost
per unit was e stinated at S428, that, making a total of Y'2i568 for
the six units. MD experts had begun in 1959 to investigate replace-
ments for these machines, which were becoming difficult to main..
taro. The requirements were for a nachine -Uhat could handle hard
copy., aperture cards, Acytifilm cards, and reel microfilm. and TMID PSD personnel came to the conclusion aft :-:r much investigation that
electrostatic printing had progressed to the poinu where it could truly
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
considered the system of the future'(speeds up to 10,000 characters per
second had been demonstrated). A proposal therefore went forward to
in March 1960 and was approved in Tune
25X1 AO/ A for a contract with
If or the delivery and installation
118/
of a Videograb Reproduction System, using Videograph Facsimile equipment.
after much testing and numerous consultations 0
More than a year late S1'- Videograph had nrt turned out
acceptable cony from either hard copy or microfilm and the contract was
11.'~/t
therefore terminated -,in November 19610 "P) personnel designed and
tested modifications to the Photostat Expeditors so they could continua
to be used (on pre-DARE material).
Since the advent of full-scale microfilming of most
documents in 195)4, OCR had been using a 16mm (= -&) aperture card for
document storage. MD discovered in the early 1960ts that most other
organizations in the country had changed to a 35mm (Mil. D) aperture card
or microfiche. All research aid development work leading to improvements
in the microimage field was concentrating on 35mm. OCR management
determined that it would require the expenditure of considerable CIA
funds for research and deve1th pent to update and make its own 16mm
aperture card more efficient. Meanwhile commercial companied were
developing improved or new mthods and machines for handling the 35mnt
aperture cards One such developmnt that NID investigated
was the Fiimsort 2000 camera manufactured by Minnesota Mining and
Manufacturing Comppn (3M)e This machine offered the advantage of
producing automatically an aperture Hard for any document tp to eight pages.
Filmsort's one -step operation would permit OCR to speed up the processing
The 'Mil': abbreviation refers to military specifications
Memo tctino AD/CR to Project Review Committee, 30 March 60., sub: Video-
grap e eproauction System. OUO (in Chrono 60 64-341/ 9)
STAT
M rgLy
,"r Q8 &#AIBQ
,Fftpg35~QORW08tjOggg-f Videograph
/ --' Contract C. Internal Use On3y0 (in Chrono 1961 64-3414)
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
of the documents so that in lieu of the 3-4 day del.-6r in getting
aperture cards into the files OCR would be able to reduce that delay to
one day at the most and possibly only 3-4 hoursa~
OCR managers worked closely with 0 managers in the latter' s 25X1 A2
plans for 43n improved t. document delivery system (DDS)o Ar r changea_in....-
DD/I
OCRs immediate system, although not initially electronically coupled
with the computer, had to have the fl.e ibility in design that would permit
such coupling in the future*
recommended that OCR go to either 25X1 A2
35mm aperture card or microfiche, but microfiche was not susceptible
to machine sorting and filing1whereas aperture cards were* , ..OCR t =~
opted for the latteP, - DD/I approved the request for an expenditure of
X30 OOOon 7 June 196! to purchas eve Filmsort 2000 cameras and
three companion Quadrant printerso
As of January 1965 the Aperture Card File and the machines that
reproduced copies were organizationall# under different managements
within OCR the former was controlled by the Document Section of the
Library and the latter was under the jurisdiction of MD. When an
aperture card was pulled from the file f or reproduction, a pink charge
card was prepared and put in its place. Four courier trips a day w ere
required to move the aperture cards from the Document Section to rD and
back againo With the arrival-of the Quadrant_ printers, small. machines re-
quiring.- only feet- of space and easy to operate (in contrast to the
bulky ancl bard to operate Photostat peditors) , the, AD/CR approved:
1 90
Mean.o,, Acting AD/CR to DD/I, i June 65, sub: Document Delivery 33rstem
3. File : Chrono 65 Job: ! r
25X1 C4C
The term DDS was first encountered inlIl #ure. It described the
plans for a large-scaleedocument storage..system withcapbility for
A The fi3Md2 as*e2" g i ~'ro tfA&RI t Model 222 D ve Printers
were installed in August 1965- Three technical engineers from IM spent
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
25X1A9A
the colocation of the files and the printers in the Document
Section.
During 1966 technical discussions continued with 3 M for
improvements in the 2000 camera process,)which was not meeting
density variationtspecifications. 3M_ was most anxious
to satisfy CIA becai'se many potential customers were awaiting the
outcome of OCR's deliberations, During the period of 3Mts work on
the camera, a team of MD experts with assistance from Prtfft Ag
artd`"S-e-rv :ce ? on undertook further testing and countless
meetings with government and cominercial experts to define more
precisely CCRts technical requirements, The team moved ahead
with exploring the possibilities of using step ahd repeat cameras
to backstop the 2000 operation, to provide an optimum mans for capturing
document images in the marginal categories,., and to provide equipment
more suitable for microfilming bound volumes.
In June 1966
!fAl
who as Executive Assistant to the
AD had been charged with the problem of compatibility with 125X1 A2G
in the document delivery system, announced that 3 days of testing the
2000 cameras that had supposedly been upgraded by 3M revealed
the same bizarre density variations from card to card and w ithin the frame
of a card experienced in previous tests. He recommended that OCR initiate
procurement of modified step and repeat cameras that would lead OCR
into the Mil-D-8-up format (that is, eight pages of information in a
Document file activity in 1965.wa0)181,621 documents with 895,288 pages
filmed; 12,500 aperture cards pulled each month; 30,000 pages of microfilm
produced on Quadrant printers each month. M?mo, tp 25X1 A9A
D/CR, ]J July 66 sub: Document Delivery S s -Final Recommendation,
A L vacli- ,,.n c. S. File: 'Chrono 1966 Slob :
`~tpp A~~ r' eli Q~4 ,Z~s l 0 0 L1~2 a document
had to be 1n d out and hot, d
ra
h
p g
p sr neous ,
e step and repeat
camera was able to move and record pages of information in a grid pattern.
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
123/
25X1A9A
35mm size frame)o Six months later I Iconcluded that
the 3M cameras were not suitable for the type of application
needed and were inadequate in terms of quality control, supply
costs, and manpower requirements.
The five 2000 cameras were placed elsewhere and OCR ordered
four National Cash Register
1
0 Model SR-ID step a rid
repeat 36imn cameras at a cost of $20,9)46 each. They passed
acceptance tests in November 1967 and were put into operation in
February 1968, w ith the final demise of the Intellofax System.
maw .n ouse improvements as necessary.
In the meantirrnes ND contixiued'to use .IrSmm aperture card _sy
74
Retrieval of these aperture cards is,as of 19?3gstill handled
by the old equipment because conversion to 35mm is not feasible.
25X1A9A
' no,s 0 to D/CR,, 11 Jan 67s sub: Recommendation-Document
Delivery System. So (in Chrono 67 69-5920E)
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
d. Usage (See also Appendix C)
Intellofax program utilization as determined from
requests in FY 1966 was divided as follows among custamer
offices:
Non-CIA
DD/I 24.1i percent 44.8%
DD/P 20-7
DD/S&T 7.2.
DD/S and 2.9
other
Total CIA 55.2 percent
Requests from oD customers totalled 1037 in
equests from DoD were limited as
1966 and 812 in 1967.
of March 1967 because of OCR budget cuts. All requests
had to be validated through tie DIA Library and coordinated
1281
with the OCR Coordinator. See Chapter (Document Systems
Group for further details of service to DoD.
Support to the National Intelligence Survey (NIS)
program was heavy throughout the 20 years of Intellofax
operation. Standard runs were pre-coded for specific chapters
of the NIS. In FY 1966 OCR received 441 NIS requests (97 CIA
and 34+4 non-CIA) providing 156,828 references of which 113,959
were printed onto Intellofax tapes.
12 / Intellofax Requests Levied by D epartment of Defense
Consumers During 1966 and 1967.5 File: OCR Surveys 1967
Job: Intellofax Historical Files in ISG.
12/Memo, DD/I to Director, DIA, 31 Mar 67, sub: Reduction in
Intelligence Support Services. S. File: Chrono 1967
Appid,Wd kiie 2 074 g9/20t.iCg4 @84-W95c1 20@@300a40@Q3 ions,
27 July 67, sub: Central Reference Support for DoD Components.
S. File: Ibid
 III Ut2J uitun.-
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
25X1A2G
4+. Finale
No one doubted that the Intellofax System was
a high cost operation. Intellofax questions made up only
four ,percent of the total number of search questions
put to OCR. During most of the Intellofax history,
30-50 ,people were directly associated with th
necessary indexing operations. Another 50-60 operated
the IBM equipment and conducted auxiliary operations, such
as microfilming and DARE, exclusively in support of Intellofax.
Faced with T/O and budget cuts on the one hand
and the prospect of expensive
on the other, the D/CR
looked at the Intellofax System with a critical eye during 196+-67.
25X1A2G Should there be more in-depth indexing as
was planning or
should there be shallow indexing as an economy measure
Whichever way OCR went, the Intellofax System as it had been
operating for 20 years was doomed. Even if greater numbers of per-
sonnel were used to provide greater indexing depth, the
system with EAM equipment could not cope-,:with the resulting
25X1 A9A flow of index information.
25X1A9A
25X1A9A
25X1A9A
25X1A9A
12J/
in his 1966 study of OCR
41HIC
and the 1967 User Study Group recommended shallow indexing for
most information. The User Study Group indicated that users
, Plans and Technology Officer, OCI, chaired
a DD/I Study Group.
** established a User Study Group of Agency representatives
1 k, 1 Dec 66, sub: Choosing the OCR File System. S.
File: older job: 68-487//
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
 Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5
requested in-depth indexing only for militar -related
12'7/
subjects in critical areas of the world.
Therefore, in the CRS reorganization. of September
1967, the Intellofax System gave way to a relatively
inexpensive computer-assisted indexing and retrieval system
through which CRS codld get minimal control over that portion
of the document flow that had to be controlled at all. The
ISC was replaced by a greatly modified version of the
25X1 C4C
fubject Intelligence Code, which had been a combination
of the ISC and the SR coding scheme.
In spite of the many criticisms levied against
it, ranging from too many references retrieved to too
few, the Intellofax System was unique. It was the only
system in the intelligence community that provided machine
retrieval of all information reports issued by USIB member
agencies. It finally bowed to the needs of the all-source
improved input and retrieval computer capability.
1 #~/ Memo, D/CR to DD/I, 11+ Apr 67, sub: Re-examination of
Gam- OCR's Role. Attachment A. Summary of User Requirements.
S. File: Chrono 1967 Job: 69-592:
Approved For Release 2004/09/23 : CIA-RDP84-00951 R000300040002-5