THE PHOTINT INFORMATION PROCESSING WORLD STAGE 1 REPORT VOLUME VI APPENDIX H STAFF FOR THE COMMUNITY INFORMATION PROCESSING STUDY (
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP80B01139A000600050001-2
Release Decision:
RIPPUB
Original Classification:
T
Document Page Count:
39
Document Creation Date:
December 15, 2016
Document Release Date:
April 29, 2004
Sequence Number:
1
Case Number:
Publication Date:
December 1, 1963
Content Type:
REPORT
File:
Attachment | Size |
---|---|
CIA-RDP80B01139A000600050001-2.pdf | 1.91 MB |
Body:
NNINN \~~N
Approved For Release 2004L ZIQR : gg~0B01139A000600050001-2
1~~JJ11''
SCIPS I-VI H
Copy 23
s
45 Pages
THE PHOTINT INFORMATION
PROCESSING WORLD
STAGE I REPORT
VOLUME VI
:PPENDIX H
Staff for the Community Information Processing Study
(SCIPS)
Approved For Release 2004/d716$ : CTf-W 0B01139A000600050001-2
25X1
I-I GROUP 1 e
E..I.ded Irani nv~oma~ic 1
Approved For Release 204T67/0~EC1R--%P80B01139A000600050001-2
THE PHOTINT INFORMATION
PROCESSING WORLD
Staff for the Community Information Processing Study
(SCIPS)
Approved For Release 2004/07/08 : CIA-RDP80B01139A0006000500
TOP SECRET
25X1
Approved For Release 200'41Q7/O Ec - i P80B01139A000600050001-2
STAFF FOR THE COMMUNITY INFORMATION PROCESSING STUDY
(SCIPS)
LETTER OF TRANSMITTAL
SCIPS D-2/3/S
18 October 1963
MEMORANDUM FOR: Chairman, United States Intelligence Board
THROUGH : Chairman, Committee on Documentation, United
States Intelligence Board
SUBJECT : Transmittal of Stage I Report, Volume VI
REFERENCE . (a) CODIB-D-82/9 and USIB-D-39.7/1, 24 July
1961, (Terms of Reference)
(b) CODIB-D-82/16 and USIB-D-39.7/3, 23
February 1962, (Stage I Plan)
(c) USIB M-202, 28 February 1962
Transmitted herewith, in accordance with references and supple-
menting my memo of 17 October 1963, is Volume VI of the SCIPS Stage
I Report.
The National Photographic Interpretation Center is due a vote of
thanks for its fine work and effort in publishing this volume on behalf
of SCIPS/CODIB.
25X1
25X1
Approved For Release 2004/07/08 : CIA-RDP80B01139A0006000500011-2
TOP SECRET
Approved For Release 200 / 108:U P80B01139A000600050001-2
Page
1. Foreword .................................. 1
II. Introduction ... .............................. 1
III. Findings and Discussion ........................ 3
A. Picture of Present Information Processing ...... 3
1. Organizations, People, and Equipment .... 3
2. Items .......................... 5
3. Flows .......................... 7
4. Processes ...................... 14
5. Files ............. .............. 16
B. Information Processing Problems ............. 22
1. Indexing ........................ 22
a. Photography ............... 22
b. Information .......... ...... 22
2. Data Exchange . ................... 23
a. Photography ............... 23
b. Reports Resulting from Exploitation
of Photography ............. 24
3. Formatting ...... ..... ..... .. .... 24
a. Photography ............... 24
b. Reports ........ ........... 25
4. Film Storage ..... .. .............. 25
C. Discussion . ..... . ... ........... ....... . 25
1. Report Formatting and Data Exchange .... 26
2. The All-Source Problem .. .. ......... 30
3. Techniques ...................... 30
4. Information Processing Coordination ..... 31
D. The SCIPS Study ....... .................. 31
IV. Summary and Conclusions ......... ............. . 32
V. Recommendations ............................ 33
Attachment A: List of Equipment ...................... 34
25X1
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001I-2
TOP SECRET
Approved For Release 21 7/08':9-RDP80B01139A000600050001-2
Figure H-1.
Figure H-2.
Page
Project World - Organizations Surveyed ...... 2
Project World - All Organizations .......... 4
Figure H-3. Project World - .......... 8
Figure H-4. Project World - KEYHOLE Film Flow ....... 9
Figure H-5. Project World - BRASS KNOB Film Flow ..... 10
Figure H-6. Project World - TALENT Far East Film Flow .. 11
Figure H-7. Project World - Penetration Film Flow ....... 12
Figure H-8. Project World - OAK Reports Flow .......... .13
Figure H-9. Processes ............. . ....... ..... 15
Figure H-l0. Information Files ......... ............ 18
Figure H-11. Film Files ........ .......... ....... 19
25X1 . V i -
25X1
pproved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
TOP SECRET
Approved For Release 2004/9108S~TA'1 l P80B01139A000600050001-2
This volume, No VI of the SCIPS Stage I
Report, consists solely of Appendix H, The
Photint Information Processing World." Al-
though published separately for dissemination
control purposes, Appendix H is nominally an
integral part of Section III, "Findings and Dis-
cussion," contained in Volume II of the Stage I
Report. However, this appendix is an entity in
terms of findings, discussion, conclusions, and
recommendations inasmuch as survey and anal-
ysis were conducted on a limited world basis,
similar to Appendix F on the Foreign Publica-
tions World, and not on the all-worlds basis
used for other parts of the Stage I Report. Be-
cause of this, Appendix H does not include the
collateral materials portion of the PHOTINT
world, but only that which is within the T-KIl
System.
This study includes most of the large organ-
izations performing information processing
within the TALENT-KEYHOLE (T-KH) System.
The most important domestic organization which
was not surveyed was the USAF Aeronautical.
The T-KI-I System is probably ahead of most
of the intelligence community in the application
of up-to-date information processing (IP) meth-
ods. Among the causes of this can be listed:
a. The System is relatively new, dating
back only to 1956, and therefore has had
the advantage of new techniques in informa-
tion processing since its inception. There
were no great holdings of stored information
which had to be maintained in old ways or
converted to new.
Chart and Information Center. Other Scope
Notes on depth given in Volume I and II of the
Stage I Report are applicable to this Appendix.
The T-KII System was created to provide
intelligence exploitation of photography obtained
by methods which must be guarded by the
strictest possible security measures. It is,
therefore, a rather small group of organiza-
tions and people who are processing information
in a separate, identifiable, and somewhat inte-
grated system within the total community infor-
mation processing system as visualized in the
SCIPS Terms of Reference. Survey-wise, SCIPS
experienced fewer access problems than were
experienced in the SIGINT and Clandestine Ser-
vices areas.
Unfortunately, time and clearance problems
prevented the machining of this portion of the
SCIPS data base, and as a result, there is less
quantitative data included in this Appendix. The
analysis phase likewise suffered. Thus, al-
though the analysis of the T-KII world for the
Stage I report was necessarily more subjective
and individual, it is included as a basis for
further work.
b. Faced with the necessity of exploiting
large masses of photography, and of com-
bining and comparing it with intelligence
from other sources, the T-KH System was
forced to improve on information processes
by whatever means were possible.
c. Perhaps because the source and system
were new, supervisors and most employees
have encouraged and adjusted readily to new,
improved techniques or innovations.
d. The System consists of a small group
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001
TOP SECRET
25X1
25X1 Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 200 19 08S~Ti'~-. P80B01139A000600050001-2
of relatively close-knit organizations which
tend to share ideas and developments
readily.
e. Funds have been available to the Sys-
tem organizations to buy or lease computers
and other expensive information handling
equipment capable of improving processes.
Despite these advantages, however, prob-
lems have arisen. The System organizations
have tackled their problems and in many cases
are approaching a solution. In some other cases,
they have so far failed to properly identify
their problems.
Because of the unique advantages the System
enjoys, if it can solve its own problems and
develop within its own boundaries information
processes which are truly efficient, it might
serve as a model for the rest of the intelligence
community.
b. People
The total number of personnel in the organ-
25X1
a. Organizations
Although there has been some reorganiza-
tion and consolidation since the field survey was
conducted, the 13 organizational units studied
are shown in Figure H1 as well as the relative
size of those organizations in terms of number
of personnel. Also shown in Figure H1 is the
Navy Photo Interpretation Center which was not
surveyed, but was visited during the course of
the study. The subsequent consolidation of the
13 organizations into 10 is indicated on the chart.
There are some 30 other organizational ele-
ments which comprise the T-KH System. These
are shown in Figure 142. A complete study of
the system would require a look at the pro-
cessing in all these organizations. As in the
case with collateral organizations, the major
central elements were selected for study in
Stage I. Whereas it is estimated that Stage I
coverage represents only 307, of the original
scope for collateral organizations, it is esti-
mated that Stage I coverage of T-KH activities
represents perhaps 50-60% of the needed scope.
izations surveyed in The
allocation of these people by organization is in-
dicated in Figure HI by the size of circle. As
can be seen, the three largest elements are
the National Photographic Interpretation Center
(NPIC), the Army Map Service (AMS/DESPA),
and the Strategic Air Command (SAC). Following
the reorganization, the Defense Intelligence
Agency (DIA), including six of the serve ed
organizations, is among the largest I 25X1
persons. (Only part of the Army detachment of
NPIC went to DIA.)
As in the case of language translators (ana-
lysts) in the "Pubint" world (see Appendix F),
the photo interpreters (analysts) are included
in the information processing personnel picture.
c. Equipment
In addition to the equipment usually found
in the other information processing organiza-
tions, there is a large amount of special equip-
ment designed and used for photographic inter-
pretation. This special equipment ranges from
the simplest viewer to very sophisticated men-
suration equipment employing the best available
optics and electronics. (See Attachment A for a
25X1
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001I-2
TOP SECRET
25X1 Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 200497/08-: lA-RDP80B01139A000600050001-2
comprehensive list of equipment.) Even so,
virtually all the equipment other than photo
developing serves as an aid to a human photo
interpreter; no system of automatic pattern
recognition was found.
The exploitation equipment in use within the
System is fairly well standardized. This has
been achieved by close coordination among the
organizations in their purchases, testing, re-
search and development of equipment. Richard-
son projection viewers, light tables, and stereo
microscopes are common to all. Photogram-
metric instruments vary as to makes and models
but not too much as to capabilities. This is also
true of photo laboratory equipment. Although
much has been tried and great sums of money
have been spent, there is nothing startling in the
way of exploitation equipment presently in use
in the System. A breakthrough does appear
imminent, however, in the development of high-
quality film chips. One vast improvement in
exploitation techniques has occurred in the
general acceptance of film, rather than paper
prints, as a viewing medium. This has given rise
to the use of projection viewers and microscopes.
In contrast to the relatively standardized
development and application of exploitation
equipment is the variety of automatic data
processing applications in T-ICI-l System or-
ganizations. Three Air Force units surveyed
had somewhat parallel developments under AF
Project 438L, and each had an IBM 7090 com-
puter available. These computers are outside
T-KH areas and are "secured" on an ad hoc
basis when system material is being processed.
NPIC has three computers, a Univac 490, an
IBM 1401 and a Minicard. The Univac 490 sys-
tem is being developed as a highly sophisticated
tool for providing on-line accurate measure-
ments to photo interpreters from remote in-
quiry stations within the NPIC Building. The 490
is also being developed for other scientific
applications which it will handle simultaneously
with the PI-assistance computations. The 1401
is being used for information storage, report
writing, and administrative functions. Minicard
is being used for storage and retrieval of PI
reports. Three of the other organizations sur-
veyed have computers which are available to
them for limited use, outside of the secured area
but within their building. These are AMS
(Honeywell 800), CIA (IBM 1410 and now a 7090)
and NAV PIC (various). (An inventory of equip-
ment for organizations surveyed is contained as
Attachment A to this Appendix.)
The reporting process within the T-KH Sys-
tem has developed into a rather regular pattern.
a. Early Reporting
NPIC produces immediate reports, OAKs,
IPIRs (Immediate Photo Interpretation Reports),
etc., which are the result of quick scans of new
photography for highest priority targets.
NPIC later produces an MCI (Mission
Coverage Index) which is a summary of all
significant targets found on a new mission. The
MCI takes about one month to produce. The
SAC PRI (Photo Reconnaissance Index) covers
about the same ground as the NPIC MCI, but
presents its findings in a more codified and
tabulated form. It is produced much quicker
than the MCI. These three report series account
for most of the systematic early reporting of the
results of new missions.
b. Special Reporting
After the early exploitation, each organiza-
tion further exploits the photography according
to its own mission. Very briefly, this results
in the following types of products:
(1) NPIC answers requirements from USIB
agencies for specific reports on items of
25X1
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001 f
TOP SECRET
Approved For Release 20047f17/08~q'I1 RDP80B01139A000600050001-2
national priority interest. There are usual .y
formal reports anc receive rather wide
dissemination.
(2) Air Force units exploit the phorograpliv
primarily in three ways:
(a) By recording all targets, car
changes to existing targets, found on
the phorographr in the Aerospace In-
telligence File (All') and then by ob-
taining special listings, one of which
is the Target Data lnve rorv (TDI),
from this comouteri7ed file. Thi.;,
incidentally, is a most iriportant wty
in which information from T-KII mt.-
terial is reported out of the System.
("]'he TDI is SECRET.) he AIF is a
large machine file containingformatt?d
information on over IOO.(00 install. -
tions.
(b) By making up llnifile PI report,;.
These reports ),live photo-derived ir-
formation on targets in a high)v for-
matted style. Also, the information can
be readily converted to machine lar-
guage and can he stored- hi some cases
textual and graphical information are
added to the formatted information and
the result is disseminated throughout
the System as a report. Most traffic
in these reports, however, occurs be-
tween Air Force units, ;nd between
SAC and A FIC 'now DIA).
(c) By updating Air Target Material:;.
tributions to NPIC and DIA, is the mapping
program carried out by AMS. Under this
program the Soviet Union is being mapped
at a scale of 1:25O,OOOwith KElphotograph).
An Armv detachment at NPIC also turns
out Photo Intelligence Reports for Army
departmental interests. Chose are in
answer to specific and standing requi:-e-
ments from Army elements.
/4) Navy PIC produces quick-response :-e-
ports in answer to requests from Navy
components. It also updates the Air Target
Material Program by use of T-KI-I pho-
tography.
(5) The CIA DetRchment in NPIC produces
quick-response reports in answer to re-
quests from CIA components. These re-
sponses are usually directed to the re-
questor only.
Briefing
i3riefings and briefing aids probably play a
larger role in the promulgation of intelligence
within the T-KIi System than in any other por-
~=on of the IISIB Community. With each new
mission a series of large annotated photographic
ortefing boards are made on highest priority
zo snow the most significant items found on the
.lm. These hoards are usedtobriefthe highest
.nterested echelons of government. Vu-graph
.slides of these boardq are made and distributed
immediately throughput the System. These
.Aides are used by recipients to brief their chiefs.
I hus, very early after each new mission the most
lfflpurtant findings are presented directly to
those elements of government most capable of
utking action. As exploitation continues and new
items of importance are discovered, these too
are worked into briefing aids by NPIC or otter
system organizations. These often provide the
-first notice of new findings.
J. Other Items
The other principal items are: (1) the raw
film itself in both duplicate positive (DP) and
duplicate negative (Dl) form, (2) large prints
of selected frames produced locally for analysts,
X31 nearly all the items that are produced in the
other worlds, e.g., COMINT, HUMINT. These
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
TOP SECRET
Approved For Release 200 4/d1108$tTA' I P80B01139A000600050001-2
materials show up in one or more of the T-KI I
System organizations.
Insufficient analysis has been made of the
survey data to present an overall flow picture.
What would be particularly valuable would be the
flow of non-T-KH material to the T-KH organ-
izations. However, this flow is indirectly indi-
cated by the file holdings discussed in Section 5,
below.
Flow within the T-KH community itself is
pretty well confined to the photographic film and
the items indicated in Section 2, preceding. For
non-T-KH material, i.e., collateral, each of the
T-KH elements indicated in Figure H2 depends
upon its parent department for support. Thus,
as shown in Figure III-n (20) in volume II of this
report, the flow of collateral items to NPIC is
There are ten principal recipients of the raw
take, and virtually no discrimination is made in
the initial dissemination. In fact, dissemination
is really made before content is known (other
than by predicted coverage calculations) and with
little regard for content. The furnishing of
duplicate positives (DP) is for photo interpreta-
tion and scanning purposes, whereas duplicate
negatives (DN) are used to make additional
copies and selected prints.
Figure H8 shows the standard distribution
of the OAK reports. The flow pattern is typical
almost entirely from CIA, and in particular from
the Document Division of OCR. This Appendix
shows the flow of T-KH film and the OAK re-
ports. Figures 113, H4, H5, H6, and H7 show the
flow of photographic film items of the various
series, respectively:
photography; KFYIJULE; BRASS 77077757TH
photography is now out of the T-KH classifi-
cation); Far East TALENT film; and Penetra-
tion film. As can be seen in these five figures,
the philosophy is one of broad dissemination of
the raw take in multiple copies. This is another
manifestation of the policy of "free exchange" of
information, but there is no evidence of re-
sponsibility for information support. The volume
of dissemination of each of these series is, of
course, very variable, but for the period under
survey the approximate annual volumes were
as follows:
for other early photo intelligence reports pro-
duced by NPIC. As can be seen here, the Special
Activities Office in DIA has become a major
focal point for dissemination to military depart-
ments which consume some 48 of the 60 copies
disseminated. This is in contrast to the simul-
taneous direct distribution of raw data from the
processing site.
As shown in the charts, movement is quite
free and voluminous between organizations in the
community. Film goes to all organizations with
exploitation capabilities. Generally speaking,
Approved For Release 2004/07/08 : CIA-RDP80B01139A0006000500
TOP SECRET
25X1
25X1
25X1
25X1 Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Next 5 Page(s) In Document Exempt
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 2000/4Ec11Et-14DP801301139A000600050001-2
25X1
reports go to all organizations. The main er -
ception occurs in the case of reports made for
internal departmental consumption. "These se':-
dom get disseminated, nor is their existence
generally known to other System organizations.
The requirements for the initial readout of
photographic missions performed by NPIC corre
from the Committee on Overhead Reconnais-
sance (COMOR). This tJSIB committee estat-
lishes priorities for targets according to their
hearing on national intelligence. The immedia:e
reports (OAKs, IPIRs, etc.) report cn the highest
priority COMOR targets.
Requirements for detailed reporting by
yIPIC go through a rather compli~_ated cycle*.
The initiator sends his requirement through his
ow.i agency's requirement channels: to the Sec-
retariat of NPIC's Advisory Committee (Ad-
COM). The Advisory Committee consists of
members from all interested USIB agencieE..
The AdCOM's Secretariat consists of members
from DIA, CIA, Armv, and NSA stationed tt
NPIC. It decides whether a reoui--ement is of
national or departmental interest. If national,
it is sent to the AdCom for a priority recom-
mendation. Then it can be accepted for work by
NPIC. If the Secretariat decides it is only of
departmental interest, it will be sent to one of
the departments to be answered. Departments
with a photo interpretation capability acceptre-
qui rements from their own components and wore
on them without reference to the AdCOM or r -)
NPIC unless they consider them of national
interest.
b. Processes by Organizations
Eleven kinds of processing that occur in thii
13 organizations surveyed (plus N.W PIC) are
shown in Figure 119. This figure is not stat~s-
tically based on the number of people or amo,int
of floor space. devoted to each process. Rather,
Figure H9 is E, graphic presentation, organizat -on
by organization, of the proportion of that organ-
ization's mission or total effort devoted to each
'rocess as "judged" by the surveyor. It is be-
eved that the relative proportions are suf-
ciently valid to give a picture of the primary
unctions of each of the organizations. Thus,
1MS'DESPA's main effort is to produce maps
and charts from T-KIT raw material. The CIA
detachment at NPIC is primarily doing photo
interpretation work, whereas SAO/DIA's func-
tion is primarily coordination and dissemination
control. The relative size of the circles in
Figure H9. show where most of each of the major
processes is performed. The relative degree of
centralization of those processes is as follows:
1) Photo interpretation - primarily by
NPIC detachments and SAC.
t2! Mapping and charting - AMS/DESPA
ACEC not surveyed)
3`r Text.tal reference services - all
(4) Phot 3 reference service - all
(5; Automatic data processing - NPIC,
SAC, FTD, AFIC
It can also be seen by the variety and even dis-
tribution of the processes that there are organ-
izations which are relatively autonomous or are
complete centers in the sense of performing 311
or nearly al:. processes: NPIC, SAC, AFIC,
and NAV PIC.
Storage and Retrieval
Storage and retrieval of information within
rile System varies from completely manual in
the case of AAIA, to highly automated as in the
cise of Al- Ur ifileand acomputer-assist system
in NP!C. The Unifile Photo Interpretation Report
System (UPIR), now centered in DIA, employs a
standard reporting form from the various Air
- 14 -
pproved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
TOP SECRET
25X1 Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 2O 14f0'7/O?~cr 1 DP80B01139A000600050001-2
Force photo exploitation units. File t1PIR sys-
tem employs two formats to report information
derived by photo analysis on each installatior.
The UPIR-1 provides for filing anc controlling
data such as: target identification; mission ani
frame identification: evaluation of giality of th>
photo coverage; specific, derivable installatioi
characteristics and conditions: and it free field
for pertinent remarks. The UPIR-2 provides
for narrative discussion of the target, interpre-
cations, and other pertinent informa :ion which i s
not susceptible to or appropriate fcr rigid for-
matting.
Upon completion, the UPIR- I s ready for
key punch without additional intermediate pro-
cessing. It is then entered into the tJPIR
ma;;netic tape file and is available or machine
processing and production. The UPIR-2 is
stored for retrieval in document ?orm in the
:MINIICARD system.
The system in NP1C is centerec around the
Target Brief File. Reports from all sources,
not just photography, are read and analyzed to
contribute to this file, but only evaluated ex.-
tracts or abstracts are introduced into the file.
The important features here are that analysis
occurs before entry into the machined data base
and is done manually by the people constructing
the data base from all sources includingphotog
raphy. Information is accepted or reected on the
basis of its use in adding to or correcting the
data base. The information is entered into the
data base in narrative form as contrasted with
Unifile which is highly structured and rigidly
formatted: however, both computer files are
target oriented.
The NPIC Target Brief File was developed
as a tool for photo interpreters. When ram-
information from a new mission is obtained, tht
pertinent portions of the Target Br of File are
printed out by the computer, roughly In the order
that the photo interpreter finds them as he scant:
the photography. Ile compares the new photog-
raphy of a target with the target brief statement
of what is already known about it and has to iden-
tify and record only the variations. New photo-
graphic coverage data, new targets, and changes
to the previous targets discovered on the photog-
raphy are added to the target brief in the ma-
chined data base, so that it contains references
to all previous coverage. The system does not
provide ad hcc query' service nor is there the
pertinance of retrieval that a formatted file
would provide. However, the narrative descri-i-
tion has its appeal and unique utility, and tie
system does .serve its original purpose of as-
sisting read-out of new photography, and, it is
believed, it could serve other broader purposes
also.
In the T-Kli System there are two types of
files, the reconnaissance film files and the noa-
film files which will be called information files.
Tie informat on files are about as varied as
those discussed in the main body of this report.
'1"ieir unit records consist of cards, documents,
tape records, etc. The photographic files, how-
ever, are quite uniform throughout the System.
P.?most universally these files consist of rolls
of film, either positives or negatives, which are
stored in cans. The unit records for these files
are the indivicual frames (or exposures). Coi-
=idering all types of film, there are roughly 100
unit records (frames) per roll of film.
b. I o:al Files and File Size
A total of 95 information files were identi-
fied within the 13 T-KH System organizations
studied during this survey. This compares with
926 files identified within 43 organizations sur-
veyed outside the T-KH System. In number of
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
TOP SECRET
Approved For Release 2001/07/085E&R- DP80B01139A000600050001-2
files, then, the T-KI-I files represent about one
tenth of the files surveyed. In terms of items
filed, however, the T-KII information files show
only 2,600,000 items as compared with over 220
million for all organizations, or roughly 1/100th
of the total, which reflects the smaller average
file size within the System. This is as expected
because of the relative age of files. Figure H10
is a graphical presentation of the size and organ-
izational distribution of information files in the
surveyed organizations.
Twelve film files were studied during the
survey. Six of these contained roll film and six
contained cut film or prints. The six roll film
files contained 73,000 rolls of film or approxi-
mately 7,300,000 unit records (frames). The
table in the Flows Section, above, shows that
there are, however, an average of some 25 copies
issued, of which about 15 go to the organizations
surveyed in Stage I. The six cut film files con-
tained 100,400 unit records. Figure H11 is a
graphical presentation of the six roll film files.
Combining information and film files gives a
total of 117 files and approximately 10 million
unit records or roughly 5% of all unit records
surveyed during Stage I.
c. Age and Growth of Files
All files in the T-KII System are relatively
young. The oldest date back only to 1956. Some
of the larger files within the System were less
than one year old at the time of the survey. In
these files the annual growth rate is more
significant than the present size of the file (see
Figure 1110). The average annual growth rate of
the T-KH information files is 60% of the present
size of the file. This compares with an annual
growth rate of 157% for files outside the T-KII
System. This great difference can be explained
in part, of course, by the fact that an addition to
a small file will show as a much larger percent-
age gain than the same addition to a large file.
The annual growth rate of the roll film files
presents an even more alarming picture. There
were 33,000 rolls of film accessioned by these
six files during the last 12-month period re-
corded by the survey. This represents approxi-
mately 3,300,000 exposures, or 47% of the total
present holdings. Although the percentage figure
is smaller than that of the information file
growth, the pinch caused by the film file growth
is much more severe. This is because the ab-
solute growth is far greater, and the bulk repre-
sented by a frame of film is far greater, than that
of the average unit record of the information files
which include such minuscule units as tape re-
cords and minicards. The storage capacity of the
six roll film files studied is already seriously
taxed, and the prospect of doubling every two
years is not comforting. It should be pointed out,
however, that the 33,000 rolls of annual growth
and the 73,000 rolls of total holdings represent
much film duplication. Each of the six files
surveyed is largely duplicative of the others, and
within a given file there is further duplication,
both between passes and missions and multiple
copies of the same reel. An estimate of the
original negatives within these files is 5,000
rolls or about 500,000 frames. All the rest are
duplicates of these. Similarly, the annual growth
of original negatives would be about 2,400 rolls
or 240,000 frames. The annual growth of all
original negatives in the T-KH System, both in
those organizations studied and those not.studied,
is projected to be 4,660 rolls or about 540,000
frames.
d. Purge Criteria
Most (79 files) of the information files sur-
veyed showed no purging at all. Of the files that
were purged, the most common criteria were
supercession (20 files) and correction (15 files).
Only one, the NPIC Target Brief File, used
evaluation as its purge criterion.
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001
25X1
TOP SECRET
25X1 Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Next 1 Page(s) In Document Exempt
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
Approved For Release 2007/#'C1KRDP80B01139A000600050001-2
There is almost no purging of tha film files.
In cases where the film is originally lower than
the T-KH security level or has been subsequently
dow graded to levels below the T - Kit, it may be
filed and serviced by organizations outside the
System. It would seem that at least some dup-
licate files of this film could be purged to make
room for new stacks of I - KH film. bit evidently
it is not being done.
e. Standard Format Files
The 12 reconnaissance film files can be
considered as having internally consistent for-
mat-3. Of the 95 information files, 33 of them
have only one or a few different formats. These
were in only four of the 13 surveyed organiza-
tions. As was the case with her files, about
one third of the files in the T-Kit System have
internally consistent formats.
f. Manual versus Machine Files
About one third of the inforrration files
stuc led were machine maintained. wt ich is about
the same ratio as for files outside the System.
All of the 12 film files are manual, althougF
several studies are now in progress investi-
gating ways of automating film storage. It should
be noted here that minicard is classified as ar
information file although it has it capability of
storing film in reduced format. Within the T-
KH System, the only application of rT inicards is
in the field of document storage despite some
experimentation with its possibilities for filtr
storage.
g. Types of Files
The breakdown of the files by type showec
69 master files, 34 derivative files, 3 duplicate
files, and one suspense file. Almost 30% of the
files in the T-KH world were derivative files
whereas only 10% of the files outside theSystetr
were derivative files. One reason for the higher
proportion of derivative files is that most of the
film files are classed as derivative even thougt
the films therein are duplicated from the original
film contained in a master film file. No two of
these files necessarily contain exactly the same
film and therefore were not classed as duplicate
files for survey purposes. From anon-technical
standpoint, it should be borne in mind that these
derivative files are highly duplicative in content.
It. Content Categories
In contrast with non T-KH System files,
there are few abstract !extract or data files in
the T-KH organizations studied. In terms of
number of files, nearly half are document files
and half are index files. In terms of the number
of unit records, over 807, (8 million) of all tle
unit records are whole documents in one form or
another: and, of course, most of these are tt.e
reconnaissance film files in which each frame
is considered a document. There were about
one and one half million index records in tie
files surveyed. In summary, the files picture
in the I'-KH organizations is one of massive
documents files with a lot of relatively small
index files.
i. Purpose of Files
As was the case in other files, the majority
of files in the T-KH System were for regular
service. However, there was a higher proportion
of special project files in the System than there
was outside. A breakdown of the other pur-
poses snows:
{1) Special files (ad hoc and standing re-
quests) 13
12) Document accounting, security control,
inventory control, requirements cor-
trot and dissemination control 12
(3) Ease of update, process support, prob-
lem solving and hold for processing 5
(4) Retirement, emergency use and ar-
chival 2
Frequency of search and duplica:e
service 2
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
TOP SECRET
Approved For Release 2001/D`!'/08S~'WZRDP80B01139A000600050001-2
Intended Utilization
Of the files studied, 44 were for the use of
all USIB agencies, 52 serviced only the parent
department, and 11 were for the sole use of the
holding component. This pattern is quite similar
to that for non-System files with a little higher
department utilization factor but an even lower
proportion of files intended for the use of all
agencies.
k. Security Classification of Files
As might be expected, the great majority of
the files (96 files) were classified "Top Secret,
Handle Via T-KH System Only". Only 11 of the
files were classified "Secret." If the 96 files
were added to the "P" category in Figure 30 in
Volume II of this report, it would show that this
category of file still ranks third in size but is
a major category.
The comments made in the body of this re-
port on this factor apply to the files within the
T-KH System in even greater degree. Almost
95% of the files surveyed had either "all" dis-
semination controls on them or the dissemination
controls had not been determined. The only
category other than "all" or "undetermined"
that occurred was "No Foreign Dissem" on 9
of the files. The conclusion again is that security
classification and dissemination controls on
items are not being used as a criteria. in file
structure or file access points.
Of the surveyed T-KI-I System files, 21 pro-
vided for direct access to the files by consum-
ers, whereas 43 provided for indirect access and
43 provided for no customer access at all. The
striking comparison here with files outside the
System is the relatively high ratio of no con-
sumer access, and yet 20% of the files are used
directly versus 117, in non-System files.
n. Physical Form of Files
Whereas punched cards were the most popu-
lar form for files outside the T-KH System, six
files within the System, containing some 7.4
million unit records (frames) or approximately
three fourths of all unit records in the 117 files
surveyed, are in reel form. Reel form in non-
System files ranked about 12th. Excluding the
film files, the file forms for the information files
ranked in popularity as follows:
Minicards: 12 files - 1,130,000 unit records
Punched cards:
records
16 files - 573,000 unit
Maps, charts:
records
11 files - 426,000 unit
5" x 8" cards:
records
15 files - 178,000 unit
Doc, vol., page:
19 files - 150,000 unit
records
Photo, frames, negatives: 6 files - 110,000
unit records
3" x 5" cards: 8 files - 76,000 unit records
Magnetic tape: 9 files - 59,000 unit records
Folder: 5 files - 25,000 unit records
Adding these to the overall community files
picture shown in Figure III v 36 and 37 in the
body of this report does not notably affect the
total except that the "reel" form moves from
12th position to 6th in number of unit records.
From a present growth-rate standpoint the file
forms rank:
Film reels: 1,850,000 frames per year
Minicards: 1,050,000 cards per year
Punched cards: 456,000 cards per year
3" x 5" cards: 312,000 cards per year
5" x 8" cards: 159,000 cards per year
Documents: 84,000 does per year
Magnetic tape: 64,000 records per year
Maps and charts: 60,000 per year
Photos, prints: 13,000 per year
Folders: 3,000 per year
Approved For Release 2004/07/08 : CIA-RDP80BOl139A000600050001
25X1
TOP SECRET
Approved For Release 20U9?7/08` RDP80B01139A000600050001-2
o. File Order
Nearly two thirds of the files ar d 90;Yo of the
unit records are filed in serial number order.
This contrasts with only 3% of the snit record:t
in files outside the System beine filed first b"j
serial number. Again, the reconnaissance film
in reel form dictates the filing order for the bull:
of the records. Subject and area ranked next in
popularity as first filing order, with other
criteria relatively insignificant. Without the
limitations of the reel as the phvsical form,
subject and area might he the most popular
filing order.
As with the other worlds in the main body
of -his report (Volume It, Section III B1, thit;
section considers the present systett in the T.
KH photographic intelligence (PHOTINT) world
from the view point of particular nfcrmatioti
processing problem areas as spec fled in the
Stage I Terms of Reference. Altlsough then
are aspects of these problem area:; which are
unique to the photo intelligence world, the bask
cha-acteristics of the present PIIi )TINT IF
system are similar to the systems in the
SIGINT or PURINT worlds.
It is useful to consider two astxoects of in.
dexing in the PHOTINT world:
a. Photography
There are two requirements for indexing;
photography: one, to indicate the area of the
earth's surface covered by the photography; the
other, to record the objects which a,e covered
The first of these requirements is usually met
by a map overlay or corner coordinates of in-
dividual frames. For instance, AC[C providett
the corner coordinates for each frame of pho-
tography and a quality evaluation of some 15
:;eginents of each frame. AFIC (now DIA) puts
this information into a machine file, the Qualita-
tive Mission Coverage Index (QMCI), out of
which the photo coverage for any desired area on
geographic-coordinates basis can be obtained.
`- ne machine can also produce maps showing
photo Coverage by quality criteria. Aside from
borne duplication and lack of standard scales for
map overlays in use for these indexes in t.ne
various agencies, there is not much complaint
about these graphical indexes. Even so, tie
esiahlishment of some standard indexing tec'i-
riques for use throughout the System would re-
sult in apprec able benefits. There is much less
agreement as to the method of indexing intelli-
gence items appearing on the photography. The
higher priority intelligence items are indexed as
a result of the immediate reporting (OAK, IPI3,
MC!, PRI), at least as far as the header informa-
tion is concerned. liowever, there are a tre-
mendous number of intelligence items on each
exposure which are never indexed because of low
priority and lack of m#npowcr. This is a prob-
lem which has received very little attention no
far by the T-KH System organizations. A good
tool for portraying, publishing, or using these
items is the T-KH maps prepared by AMS (see
Item 9, AU 410)*. These cannot properly be
called indexes, however, and theproduction time
required is cosiderable.
he indexing of information within the T-KH
System is only partially effective. The indexing
of photo interpretation reports has received
more emphasis than that of collateral informa-
tion. quite naturally, but even these are not
thoroughly indexed. One of the more thorough
jobs of indexing photo intelligence reports is that
being done by NPIC in their Minicard System.
*Reference is to tnpublishod SCIPS surrey form.
pproved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
TOP SECRET
Approved For Release 2004/91085'CW DP80B01139A000600050001-2
The depth of indexing for Minicard is consider-
able. The difficulty of the Minicard index is in
the Minicard equipment. It is not a quick task to
make a search and retrieval from Minicard re-
gardless of the excellence of the indexing. At
the time of the SCIPS survey there was over a
year's backlog in indexing for Minicard, but it
is understood that a considerable effort has since
been made to eliminate the backlog. If current,
this system would index all photo intelligence
reports from all sources available to NPIC.
The UPIRs include many installations that
are not identified but look significant or sus-
picious. Out of the UPIR file at AFIC (now
DIA) lists of these unidentified installations can
be obtained and sent to analysts for further
study and correlation with collateral data.
AFIC (now DIA) maintains an index to PI
reports by target (BE number) and can list all
PI reports available on each BE-numbered in-
stallation.
On the collateral side, the largest effort:
within the T-KH System to make information
available is being made by AAIA (now part of
DIA) with their area cards and report summary
cards (Items 2 & 9, AU 430). These cards
represent extracts of information mostly from
COMINT. The information contained within the
System is extensive, but the depth of indexing
is not great. These are entirely manual files.
The most complete combination found within
the T-KH System of photo-intelligence and
collateral information was in the Target Brief
File of NPIC (Item G5, PU). As was true for
NPIC's MCI files, however, these Target Briefs
are not indexed in depth. Since they are in
machine language, the prospect of indexing them
in depth can be entertained.
A common indexing problem observed in all
of the surveyed T-KH organizations that had
automated data processing was the absence of
quick specific response capabilities from the
computers. This is not really so much an in-
dexing problem as a limitation of the machine
system. However, since the purpose of indexing
is to provide access to information in files, the
indexing system and the machine system should
be complementary. In no case where we surveyed
an automated system could they provide ad hoc
counter-service on research requests. The most
usual way of answering specific requests is
either to refer the requestor to a tabulation
previously and periodically made from the file or
to run a new tabulation for him. It would be his
job to find his answer from the tabulation. The
AIDS System (438L) had developed a method for
making direct inquires of the data base stored
in the computer (IBM 7090). The method con-
sisted of writing a short program to take care of
each request. This, of course, took quite a long
time. The method was not operational when the
survey was made, but the developers thought that
the time could be reduced to two to three hours
per request. This, of course, is still far too long
if we are to expect effective information support
service from computers.
It is useful to consider the data exchange
problem in terms of the photography itself and
the intelligence reports resulting from exploita-
tion of the photography.
a. Photography
The method of exchanging photography with-
in the T-KII System is essentially to make a copy
of everything for everybody (see Figures H3
through H7). Considering the cost of film and the
costs of developing, copying, transporting and
storing, this is an expensive way to exchange
data. Considering further that only a small
amount of any one roll -- or even any one ex-
posure -- contains any priority intelligence,
this is an inefficient way of managing either a
Approved For Release 2004/07/08 : CIA-RDP80B01139A0006000500
TOP SECRET
25X1
Approved For Release 20 197/l :c -T DP80B01139A000600050001-2
dissemination or a storage and retrieval pro?
gram. Photographic missions are now so long
and so frequent that most of the storage facilitie.a
wit tin the T-KH System are fast reaching ca -
pacity. The exploitation capability within tir3
cormunity has long since fallen behind the raw./
inp.it rate. Hach mission receives a more-or-
less thorough scan for highest priority targets,
but each does not receive exploitation in depth.
As a result, there is sa much latent informatio~i
on each roll of film that no holdins*_ organizatioi
wants to get rid of as much as a single roll cf
photography. So this method of data exchange,
i.e., giving entire copies of all missions to all
interested parties, could well end ip at last i i
crowding everyone out of their buildings and still
not produce the desired significant ;gins in ex-
ploitation. Half way between photogr iphy and re-
ports are briefing boards. I hesa are larg,!
annotated photos which serve as reports. Iher.!
is not a complete exchange of these graphics,
since many organizations prepare item for their
own use and give little attention to their possibl.!
advantage to the community.
b. Reports Resulting from the Fx3loitation cf
Photography
In some respects the data excbarge of formal
reports within the -F-Klt System is like the ex-
chaange of photography, that is, sent everyone a
copy of everything (for example, see Figure
lI8). In each of the organizations surveyed, the
libraries had one thing in common, they each had
copies of all of the formal PI reports turned of t
by the other organizations within the F- KFiSvs-
tern. T here is some exchange of machine
language data within the System but, so far, not
very much. Because of the close-knit nature of
the T-K Ii System, the exchange of machine Ian-
guatge data between organizations :ould be ex-
panded relatively easily. There is much internal
photo exploitation done which is never published.
Many reports are produced for use within organi-
nations (notah;y FTD, CIA and Navy), but are not
disseminated further. The fact that these studies
have been made never comes to the attention of
other System organizations. This results in
local duplication of exploitation efforts. Like tie
biographic information problem discussed in
Volume 11, photographic exploitation is often
local-use oriented; but most organizations are
short on exploitation resources, and the local
resolutions of information ambiguities should be
made available to all.
,. Photography
Le format of photography may at first seem
to be more of a technical problem than one of
information processing. However, it leads to
:a: rious information handling problems. Some
ai these problems are:
t V Different sizes of film make the manu-
Iacturing of viewing equipment very diffi-
cult. Very complex lens and projection ar-
rangements are required if one viewer is
designed to handle all film; too many n-a-
chines are required if a separate viewer
is built for each type film.
Some of the film sizes in use in the
l-KH Systems are: 70 by 70 millimeter,
70 by 750 millimeter, 9 by 18 inches, 5
inch strip, etc. (See unpublished survey
forms: ilk 48-51, PU.)
i2) Because of the difficulties above, good
reliab1L- all-purpose machines are not vet
available for viewing and interpreting pho-
tography. Photo interpreters are forced to
usetoja equipment (usually light tables and
measurirg tubes to scan and interpret pho-
tography), thus limiting their productivity.
l3) 1 he practice of keeping film intact in
its original roll creates another problem,
again traceable to format. This is -he
Rpproved For Release 2004/07/08 : CIA-RDP80B0l139A000600050001-2
TOP SECRET
Approved For Release 2004/U7108~'Cl' 8l P80B01139A000600050001-2
difficulty a researcher has in locating his
object on film. On a single exposure of
KH film, hundreds of square miles of the
earth's surface are imaged. Most of the
area is extraneous to the researcher's
needs, yet, because the area of interest
has not been isolated for him, he must pick
it out from all the rest. This is often a
difficult procedure.
b. Report
One serious problem regarding the format
of reports is the tendency to present in tabular
form information which is stored in machine
language.
The Mission Coverage Index produced by
NPIC is an example of this, as is SAC's Photo
Reconnaissance Index. The purpose of both of
these reports is to notify the intelligence com-
munity of the targets covered by a new mission
and to give brief information on their current
status as interpreted from the photography. Both
reports, however, end up serving something less
than their purpose. The SAC report is pub-
lished in highly formatted style with highly coded
language. The coding and length (average about
100 pages) combine to make the report difficult
to read. The NPIC report presents its informa-
tion for each target in clear English, but the
length of this report (average about 300 pages)
and the repetition (the same items are reported
in each consecutive report) make it almost
equally uninteresting. Furthermore, the NPIC
report requires an extensive manpower effort
to produce, and it takes so long in preparation
that the value of the information is seriously
reduced by the time it reaches its customers.
Both reports bear the word "index" in their
title, but both serve as an index on only limited
factors. The stored information on the tapes
from which each are produced are far better
indexes in that they can be searched for any
element in the header information. The NPIC
report is indexed only by the header information
that goes with each target description. This
means that most of the substantive content cannot
be searched automatically.
Some problems closely associated with the
foregoing are: the lack of optimum means of
entering information into the automated data
base; the difficulty of evaluating the informa-
tion against that already in the data base before
entering it finally; and the failure to provide
for disseminating the information rapidly in
response to both general and specific inquiries
after it is part of the data base.
The film storage problem is the result of
high rate of input and no prospects for consoli-
dation. The current trend in film file growth is
alarming. Another problem of storing photog-
raphy is the varying sizes of film. Most roll
film is packaged on reels which are then put in
cans. The cans most commonly in use now
vary from 3 to 10 inches in height and from 5 to
9 inches in diameter, with about six sizes be-
tween these extremes. This causes great diffi-
culty in the orderly physical arrangement of file
space. Storing film on rolls and in cans is a
very inefficient method, yet it is virtually the
only method now in use. Cumbersome serial
number search, usually manual, is required. The
T-KH System has not yet developed anything to
solve this problem, but several organizations
have large scale experiments in progress. There
is not, however, a concerted effort directed
toward this problem, which is, after all, a
mutual one.
Because a. minimum of analytical effort has
been applied to the survey information on the T-
25X1
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001F2
TOP SECRET
Approved For Release 2Q 07/0'S--9,9 DP80B01139A000600050001-2
KH organizations, the following discussion must
be considered no more than explorato:-v, however
positively it may be stated. It is hoped that the
discussion will stimulate further stuiv alongthe
avenues identified.
a. Reporting
The very early content reporting, per-
formed by NPIC, is effective in presenting high
priority information from each new mission in
the relatively rapid form of OAK, IPIR and
similar reports. Detailed photo interpretation
reports in answer to specific requests are
generally of high quality. The formal detailed
reports produced by NPIC, however, take a con-
siderable amount of time and effort in their
preparation. This limits both the number of
such reports, and the timeliness of tie informa.
tior. Another difficult problem is the systematic
reporting of targets covered by each mission
This is the area presently being served by the
Mission Coverage Index prepared by NPIC and
the Photo Reconnaissance Index prepared by
SAC. These reports are difficult and time con-
suming to prepare, are lengthy and hard to read,
and are not the-best reference tools.
The present practice within the T-KH Sys-
tem of reproducing copies of the film of entire
missions for all organizations with an exploita-
tion capability is very expensive in initial ;re-
sources and continuing maintenance. An alterna-
tive to this practice might be to reproduce and
disseminate only selected portions of the film.
The selection would be on the basis of those
portions of the film needed for the particular
exploitation interests of a given organization.
There is a lively exchange of fcranal docu-
ments among the System organizations. Each
organization has a complete file cf the other
organizations' formal reports. On the othe -
hand, several organizations have extremely
valuable files of abstracted and evaluated in-
formation which are not shared at all or at best,
are shared very sparingly. Among the latter we
can no-e: the ARIA: file or area cards and
summarv cares; Air Force and DIA tape files of
Unifile PI reports (UP1R); Air Force tape files
c.f Photo Reconnaissance Index Reports; FTD's
files of "In house" PI reports; NPIC tape files
of -Target Briefs and Mission Coverage Report
items. Ways should ;be found to combine the
efforts that go into constructing these local files
and others within the system, in order tocreate
a common product which all could use to equal
advantage. This product would be evaluated,
abstracted /extracted information.
(I) Suggested Standard Report Format
It seems feasible that a basic report format
common to all elements of the T-KH System
could be developed. The purpose of this report
would be a single method of reporting, referenc-
i ig, storing, retrieving, and exchanging informa-
ton of objects (targets) imaged by overhead
reconnaissance. I- or ease of reference in this
discussion, st:ch a report format will be called
the Standard '[larger Description (STD). The STD
would have the following features:
(a) A separate STD would be written for any
one target. The body of the STD would be
short -- not to exceed four or five pages,
and normally to be of one page or less --
and would be written in plain English. The
header of the STD would be an encoded and
formatted index with extracted information
from the body of the STD, probably very
much the same as the present UPIR header.
(b) It would contain a summary of the
latest and best information known of a tar-
get, including both photographic and col-
lateral information.
(c) Each STD would be prepared for and
stored on magnetic tape or other mecia
Approved For Release 2004/07/08 : CIA-RDP80B0l139A000600050001-2
TOP SECRET
Approved For Release 2004/9108S'CTXC-I DP80B01139A000600050001-2
compatible with high speed automatic data
processing equipment.
(d) A new STD would be prepared when-
ever a new target is identified. Revisions
would be made whenever a significant change
to. the target has been identified.
(e) New or revised STDs would be dis-
seminated with the same speed as presently
given by OAKs to highest priority targets
and only slightly less promptly for lower
priority targets.
(2) Associated Reference System:
The magnetic tape record of the STDs would
serve as the substance of a uniform file of all
select current information on all targets. This
would form the basic information file within
the T-KII System and its format would be com-
mon to all organizational elements.
b. Storage and Retrieval
The storage and retrieval of film in reel
form is expensive and inefficient. The alterna-
tive here, too, would be for all secondary files
to store only selected portions of the film and to
store it in a more manageable form than in reels.
Secondary files refer to all files except the one
for the original negative. That file, the primary
file, should retain the original negative and one
duplicate positive set in complete reel form. A
physical security backup file would also, of
course, have to be maintained.
The tremendous volume of stored film
throughout the community could be reduced by
the following practices:
(1) Produce and store only two complete
sets of film, the original negative and a
duplicate positive. These would be main-
tained by a central facility (e.g. NPIC).
(2) Produce film chips for all subsequent
uses of the photography. This would include
most interpretation and research uses. The
initial selection of chips would be done
centrally at the time of the early reporting
of a mission. This selection would be on
the basis of requirements which all Sys-
tem organizations would establish with the
central facility. The selection might be
done by central facility personnel or by
on-site personnel from the using organiza-
tions. The chips would be distributed by
the central facility as soon as possible
after the mission. The facility would have
to provide rapid service to System organiza-
tions on later orders for chips; and it would
have to provide precise photogrammetric
measurements. The facility would have to
make such measurements as a service
whenever use of the original negative is
demanded. The central facility also would
have to service requests for reproduction
in formats varying from chip size. These
may include full frames and portions of
rolls. Standard chip specifications would
have to be determined.
(3) The retrieval system for film chips
should be part of or interlocked with the
system used for retrieving target descrip-
tion information referred to above. Perhaps
a general purpose computer would be used
in obtaining an address for pertinent chips
at the same time it searches and retrieves
on target descriptions. A search capability
on a variety of criteria (subject, mission
number, data, coordinator, etc.) would be
required.
The idea that film chips might replace
roll film as the basic medium both for
storing significant photographs and for pho-
to interpretation work is being evidenced
in many places within the System. Several
organizations working independently seem
to be arriving at nearly the same answer:
NAV PIC, SAC and FTD have arrived at
the same size chip (70 by 100 mm) in their
25X1
Approved For Release 20 T~P0SECA-R P80B0l139A0006000500011-2
Approved For Release 2007/W'DIARDP80B01139A000600050001-2
25X1
systems which they call Ml1'KA', SCRAM,
and FOTECH respectively. This is not a
coincidence since their chip format is es-
sentially the same as the one developed
earlier for the SAMOS program under sub-
system 1. Of the three, VI D's f O`i'ECH
relies on a general purpose computer
(113M 7090) to index the information on the
chips and to locate the right c 'ips by ad-
dresses. SCRAM is still experimenting
with combination computer/stor ege devices
which are cumbersome in matching uppho-
tographic and textual information on the
same target. SCRAM has developed good
viewing equipment in prototype models.
Much of the F(TEC11 equipment has yet
to be produced. On paper the who a F TECH
process including scanning, printing, en-
larging, viewing and final interpretation
equipment as well as storage and re-
trieval facilities has considers )le appeal.
Navy's h4ITRAN was developed more for
the storage of textual information than for
photo interpretation, and at the time of the
survey the photographic apniications had
not been worked out.
At the same time that the development
of the 70 by 100 nim format i:a going on,
other film chip experimenting is beingdone
by other organizations. NPIC has started
using cropped film positives from KH ex-
posures to accompany its target briefs.
NPIC:: analysts have been collecting film
positive stereograms for years for use as
interpretation aids and occas onaliv for
briefings. No systematic stori ge and re-
trieval method has been worked out for
these stereograms. NPIC also cried Mini-
card as a means of reduced storage for
aerial photography, but so fai it is un-
satisfactory.
It might be said that the vu-graphs and
film slides that are used throughout the
System for briefing purposes also are re-
lated to this film-storage problem, since
annotated film chips could replace present
projection slides and vu-graphs.
The process of initially selecting film
chips from new missions for all System
organizations should be worked into the
early reporting procedures (OAKs, IPIRs,
etc.).
A very significant by-product of a film chip
storage system would be to provide pictorirl
tie-in with the basic target description suggested
in (a) above. It could provide for a textual
description of a target to be linked with stereo-
pair film chips of the same target in a single
computer search.
c. Fxchange of STDs
Speculating further on a revised system,
S Li reports might be prepared and distributed
routinely by NPIC. in accordance with its mission
to provide early exploitation of photography.
When System organizations other than NPK;
originate or revise SI'1)s, they, too, would dis-
seminate hard copy reports. The central facility
would acquire and maintain a central tape and
chip file of all STDs. In addition to the dis-
semination of hard copy reports by originators,
the central facility would provide the other desig-
nated System organizations with copies of the
pertinent portions of the tape and chip files.
Transmission of the tapes and chips would be
by the fastest means available. At first, this
will probably mean sending them by air trans-
port. Eventually, however, direct remote input -
output facilities will probably be available whic'i
will allow for immediate automatic updating of
all local files within the System.
The STD might replace the textual portions
of most of the early-type PI reports such as
Approved For Release 2004/07/08 : CIA-RDP80B0l139A000600050001-2
. TOP SECRET
Approved For Release 200L II`T'/088U P80B01139A000600050001-2
OAKs, IPIRs, Sitsums, and perhaps the MCIs,
PRIs and synoptic target reports such as Unifile
PI reports. It should be designed to replace as
many of these as practicable. The STD reference
system might also supplant present integrated
photographic and collateral reference systems
such as the AAIA/DIA Area and Summary
Cards, and the NPIC Target Briefs. They could
serve both as the basic reporting medium and
the basic reference file for the entire System.
As such, they would be current at all times be-
cause they would be compared, re-evaluated and,
if necessary, revised whenever new information
is received and exploited by any participating
component.
d. Conversion of Present Holdings to STD
Format
All System organizations with current files
of target information of priority areas throughout
the world (priority areas to be established by
COMOR) would convert present file contents to
STD formats. To avoid duplication this might
be carried out as follows :
(1) NPIC would convert its Target Brief
File to the STD format (creating the narra-
tive body and some header information in-
cluding index terms). Hopefully, the for-
matted header information could be ob-
tained by machine from the AIF or UPIR
files in DIA.
(2) NPIC would make the STDs available
to all System organizations.
(3) Each organization would check the
STDs against its own holdings, and would
then prepare and disseminate new STDs for
targets not included by NPIC and revise
the STDs for which they may have additional
significant information.
A system might be worked out by the central
facility whereby the various participating organ-
izations might be assigned geographic coverage
areas to begin the conversion process, each or-
ganization taking only one area at a time and not
starting one that is being worked concurrently
by another organization. In this way there would
be controlled overlap of effort during the con-
version process.
e. System Development
The capability of establishing such a system
of accepted STDs already exists in part at sep-
arate points within the System. The DIA/Air
Force files have wide coverage of targets and
extracted, formatted information but are not all-
source. NPIC's Target Brief File has many of
the needed ingredients but not formatted in-
formation nor indexing in depth nor wide enough
coverage of targets. With present Electronic
Data Processing (EDP) techniques and hardware
it is necessary to extract, format, and usually
encode information if it is to be manipulatable.
The day appears far off when we can auto-
matically process information in its natural
running textual form without extracting/for-
matting it. There are, however, numerous
techniques both in operation and under develop-
ment to automatically abstract and index machine
language information in varying degrees of
acceptability. One such effort is within the
T-KH System itself (FTD). Such a capability
might be applied to the NPIC Target Brief, which
is in machine readable form, and result in an
adequate basic file of indexed significant infor-
mation on targets. The file would have the
following features:
(1) Evaluated information only
(2) Index in depth
(3) Information in plain language, but in
short waste-free sentences
(4) A standard tape format enabling the
search and the exchange of target briefs
by means of tape exchange, teletype or
other remote input devices.
25X1
Approved For Release 2004/07/08: CIA-RDP80B01139A000600050001
TOP SECRET
Approved For Release 20U907/1V;l> - DP80B01139A000600050001-2
25X1
There are requirements for extracted/for-
matted information files (UNIFILF /AiE versus
narrative unformatted information indexed i i
depth, and yet the utility of the narrative is al-
most universally acknowledged. This either the
two should be combined into a common com-
munity file (either physically or access-wise),
or the indexed narrative might becone the com-
munity standard, with formatted information
files established at local option.
Other desirable characteristics of the svs -
tern would be:
(1) The stored (i.e. tape, disc, drum) re -
cord for each STD would include references
to all reports and all photographsc materials
that were used in preparing or updating it.
These references need not be p: inted out on
the STD reports but would be available fo:-
research.
(2) The unformatted portions of the STna
would be indexed in depth tprotably by ke%,
words and classifications).
(3) Computers would be geared to provide
on-line counter-service for ad h 3c research
requests on the STD files.
(4) Reconnaissance film clips could N!
prepared to accompany each STD. The
chips would be revised as often as is neces-
sary to reflect latest changer; or better
coverage. The chips for each S'rl)wouldbc?
referenced with the STD numbe', and would
be distributed with the STD whet requested.
Chips would be stereo-pairs wheneverpos-
sible.
One of the most trying problems facing the
T-Kll System is how to release information
without compromising security to analysts in
other areas of the intelligence community. P.
partial answer has been found in sanitization and
downgrading procedures, but these are soma-
thing less than satisfactory. Other possible
solutions that invite exploration are:
(a) The creation of an all-source target
reference file svfitem. All such research
requests and requirements would be searc`t-
ed through this file regardless of the clear-
ances of the originator. If an answer was
classified too highly for the originator's
clearance, then it could be sanitized. This
would necessitate, of course, the establish-
ment of to on-the-spot sanitization capa-
bility, which is where cross-systemcorre-
lat ion is needed. The basic file proposed
in Section 1, above, would lend itself, of
course, to the all-source file suggested
here.
(h) The sanitizing and downgrading of the
AMS T-KI1 maps to SECRET. There is al-
ready a development along this line which
would replace the photo mosaic on the maps
with an air brushed replica, but this is an
expensive approach. To be effective in
carrying much information to analysts out-
side the System, the maps should be left
as much intact as possible.
a. Electronic Data Processing
I he effect iveness of all automated reference
files observed during the survey was limited by
one factor, the inability to provide quick answe:-s
to specific inquiries. Evidently, the day has not
yet arrived when computers can provide counter-
service in the community. Because of this, the
computers are flooding the System with volu-
minous listings of dreary information. From
these, the researcher. is left to find his answer
and to make any further correlations. Another
problem which plagues the computer operations
is the slowness of getting information into ma-
Approved For Release 2004/07/08 : CIA-RDP80B01139A000600050001-2
TOP SECRET
Approved For Release 20OT9PIO . 6A EP80BO1139A000600050001-2
chine language. This usually involves key-
punching; and whereas computers may perform
in microseconds, key punching is measured in
tens of seconds per card.
A specific-inquiry technique is under de-
velopment at SAC and presumably at other or-
ganizations. This should be watched with great
interest. A quick inquiry /response capability
is the sine qua non of the basic reference file
suggested in Section 1, above.
Three of the surveyed organizations are
putting great effort into the computation of the
exact track of satellites involved in KII photog-
raphy. These tracks are then converted into
precise locations and/or rectification factors
for each exposure. Before the film is received
in the Washington area, hasty computations are
produced by AFIC (now DIA) and SAC, based on
data received from tracking stations. More pre-
cise computations are made by NPIC and by
AFIC after the photography is received. All
three of these organizations are seeking to im-
prove their methods so that precise information
can be produced at the earliest possible time to
aid initial exploitation. Consideration should be
given to assigning the task entirely to one or-
ganization or to dividing it among the three.
The T-KH System has benefitted from the
work COMOR performs in coordinating the in-
terests of all the organizations by assigning
priorities to targets. There is a great need for
coordination in the area of information pro-
cessing. There is much yet to be accomplished
in the development of optimum information sys-
tems, and it is still very doubtful whether any
one component is going to come up with the
answers.
There is need for a group or mechanism to
coordinate IP system development toward an
optimum information usage program within the
T-KH system. The goal of this coordination
effort should be to make the processing of
T-KH photography as excellent as the photog-
raphy itself and, with the desired specificity of
content control, to make it available to all who
need it within appropriate but varying time
factors.
The SCIPS effort in the T-KH System was
hampered considerably by the security problem
and by inadequate staffing with properly cleared
personnel. A great majority of the field survey
and analysis of the data had to be accomplished
by one staff member, and he was recalled by his
parent organization long before the work was
finished. Two of the surveyed organizations
furnished personnel to make the survey within
their own organizations. These local personnel
were given a short training course bySCIPSbe-
fore beginning the survey. One visit by a SCIPS
staff member was made about midway in the sur-
vey. The results achieved by this technique were
not as satisfactory as those achieved by the
SCIPS staff surveyors. However, in these two
cases it was a way to get some information
which otherwise would not have been obtained.
Again because of the security problems, it was
not practical to convert the collected data to the
machined data base, and consequently there was
no machine assistance in cataloging the data for
analysis.
However, the surveyed organizations were
more receptive of such cleared SCIPS personnel
as were available than were organizations out-
side of the T-KII System. There was no case in
which access to the organization or to informa-
tion was denied. However, considerable data
Approved For Release 2004/07/08 : CIA-RDP80B01139A0006000500
TOP SECRET
25X1
Approved For Release 2O 04t(17/1EcPe DP80B01139A000600050001-2
were not available simolv because the organiza-
tions did not keep records of the tvpe desired ty
SCIPS.
As is pointed out in Section III - D of tl-e
main report, the greatest amount of data result-
ing from the SCIPS effort is in the file of survey
forms and exhibits, and not in this St we I repor'.
That is particularly true here because there are
no catalogs of data available as there are in sup-
port of the analyses contained in the hain body of
the report. In this case, however, the forms have
not been carefully reviewed for a,!curacv and
completeness and care must be exercised in
their use. It would he well for any follow-on
effort to review and update the information col-
lected, and to get it into the machine data blase
so that it can be made available in the same
format as the other data.
The reader should refer to Sections I-C .nd
Ili - D of the main report for further discuss::.on
of the SCIPS study effort. The discussion therein
applies equally to this appendix. One notable
aspect of the Study effort in the T-Kl 1 organi2a-
tions was the frequently expressed desire of one
component tc have SCIPS-type information on
counterpart components in other departments.
In general, the findings within the T-KH Sys-
tem paralleled those of the other community
organizations that were surveyed. l'orthisrea-
sor the summary and conclusions o1 the body cf
the report are almost equally valid for this
appendix. The most pronounced elements of the
T-KH System are treated here:
A. The T-KH System is small in terms cf
organizations involved and ishighly Specialized.
B. It has certain problems ?ha: are unique.
These problems include:
1. Film Storage and Handling: Some
108,000 rolls of film are being distributed
to the System each year. and this has beei
increasing. It is important that methods lay
devised for reducing the number of copies
without denying needed photography to anf
consumer. Improved methods of storage
and retrieval are paramount in the solution
of this problem.
2. Integration of Information from Photo-
graphic and Other Sources: The sensitivit"
of the photo-collect ion systems and the spe
cial skills required to "read" information
from photographs, as contrasted with textual
information, makes It difficult adequately to
integrate information from all sources i-sto
the intelligence picture. A great portion of
the analysts in the intelligence community
are precluded from utilizing fully theT-??;iiar in sire nra capability to 4FIC (see item 6d, above). Exact figures
.t~?n? not ihluinev in survon.
liar in ,ire oral capability to DESPA (st'i' item 13d, below). Exact
rr?`? a,-r,' not f,btAainetl in curl iy.
i'vlr