Approved For Release 200251RDP81-00261 R0003001Y$0,-,
4? ~~;Ai u G I S 7 Y
FIL
ASPIN -- A SUMMARY
FOREWORD
Project ASPIN was established to take a broad look at
the employment of ADP support to the production of intelligence.
It was stimulated to a great extent by recommendations made by
the President's Foreign Intelligence Advisory Board (PFIAB).
The DD/I, the DD/S$T and the DD/S had representation
on the ASPIN Staff. The project was assigned to the DD/I and
was subordinated to the DD/I Plans Staff for administration
and guidance.
The basic objective of ASPIN was to develop a broad concep-
tual design for ADP support to intelligence production.
PRECIS
The general condition of ADP to support intelligence
production in the Agency is excellent. The number, variety,
size, scope and utility of the applications studied compared
very favorably with the experiences of industry, the academic
community, and other government agencies.
As in o'--her organizations, the Agency has had its most
serious difficulties in trying to implement large systems. The
primary interactive system of the Agency also leaves something
to be desired at this writing.
l cso~'JP
E cF ei iraF ..:troafz
Approved For Release 2002/05/01 : CIA-RDP81-00261 R00030013~Alk~f`1,";` "?~,~s
Approved For Release 2002/05/01., IA-'16
P81-002618000300130001-8
With increased receipts of information and the capability
to manipulate it via computerized files, there has been a
tendency to canalize information flow within the Agency. This
phenomenon will probably cause us to reconsider our present
structure on information centers within the Agency.
The need for a large generalized data management system
has been recognized. While the development of such a system
is difficult and fraught with pitfalls, the Agency should pursue
this objective as a matter of some priority.
The most negative impression gained during the study is
that ADP activities within the Agency are poorly managed, i.e.,
not enough central planning, central coordination, and central
control of our ADP resources. To enhance management and control,
more complete and articulate documentation of computer project
proposals is required. Additionally, a tighter review of such
proposals shculd be undertaken, users of ADP should be charged
for the services received and a special advisor to the Executive
Director-Comptroller should be appointed to keep management
abreast of ADP developments.
On the whole we believe that the Agency has broadly en-
couraged the use of ADP techniques in its analytical and its
collection role as well as in administrative activities.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002to/v''~'C&=RDP81-00261 R000300130001-8
Computer Applications in Support of Intelligence Production
Applications designed to support production can be charac-
terized in three ways -- Calculation, Information Storage and
Retrieval, and Data Reduction. Certain modeling and simulation
applications and other special applications tend to be a meld
of the three basic characteristics. Most applications developed
within the Agency are discrete and are designed to accomplish
modest tasks which could not be undertaken without the power
of a computer. Production analysts also have been exposed to
three relatively large systems, they are:
CHIVE -- Designed to be an ultimate information system
which would extract relevant information from all incoming
documents, capture these data, store it and retrieve it
on demand. These objectives exceeded the resources the
Agency was prepared to commit to index, abstract and process
the records involved. CHIVE, while a technical success
of sorts, was reduced in scope and resulted ultimately
in a system termed AEGIS, which presently serves analysts
calling on the Central Reference Service for research
assistance.
COINS -- This community-wide system has been of virtually
no use whatsoever to production analysts. The system is
currently under review by Agency management to determine
whether and how we should continue with the program. Its
basic failings have been the lack of attention to the files
incorporated into the system and the complex technical
nature cf the system.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/06/01,: CIA-RDP81-00261R000300130001-8
OCS Interactive System -- This is a large scale effort
to provide a general, on-line, time-sharing service for
the Agency. It is used to a considerable extent in the
development of computer programs for customer offices,
as a generalized storage and retrieval system and as a
very powerful computational asset. It has been beset with
difficulties but continued development to increase its
versatility and reliability is definitely warranted.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/0 01 F It-FDP81-00261 R000300130001-8
Problems in Computer Applications
The principle problem in implementing a computer applica-
tion is the delay incurred between the conception of the appli-
cation and its delivery as an operating system. These delays
stem from the following:
a. Difficulty in defining the objectives and goals
of the application.
b. Lack of appreciation of the difficulty and complexity
of preparing data inputs for the computer.
c. Difficulty in communications between substantive
analysts and computer analysts.
d. Lack of a central point of project control, i.e.,
substantive office resource control and OCS resource control.
Until an analyst (or any customer) has worked his way
through to - and begun to support -- a sustained operating
computer application, he can seldom be convinced of the com-
plexity of data preparation, data formatting, and data testing.
All users attest to this.
The communications problem between customers and pro-
grammers is a serious deterrent to effective ADP success. In
one effort to overcome this problem APSIN has recommended the
assignment of applications programmers to production components
to provide better understanding of their mutual problems.
Considerable repetition or overlap in program development
has been observed among the several data processing centers in
the Agency.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002 O 1cj :1,4-RDP81-00261 R000300130001-8
File maintenance has been identified as an area posing
serious problems in data processing within the Agency.
Extreme caution must be exercised in making files available
for "public" use within the Agency. While ASPIN endorses the
creation of general or joint files, it wished to underline that
their development and operation is a complex matter which must
be carefully controlled if they are not to misinform rather than
inform the user.
The problem of ADP security is treated exhaustively and
well in a large number of publications. The ARPA report Security
Controls for Computers and the USIB Computer Security Subcom-
mittee document Identification of Computer Security Problem
Areas IBSEC -. CSS - R -2, present the best treatment of this
subject.
Approved For Release 2002/05/01 CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/05/0-' :-oIA'F 81-00261 R000300130001-8
General Conclusions and Recommendations
A procedure be developed to integrate new collections
systems with data forwarding systems and ADP systems through
the use of standard formats.
That one or more general data management systems be
acquired from commercial vendors for use in the Agency.
That the Agency retain its present capability for development
of unique systems.
That mechanisms be developed to enhance communications
between customers and computer oriented persons.
That the personnel resources required to explore, develop
and test new techniques and concepts be MADE available by the
Agency.
That the present interactive, time-sharing system be
expanded to provide for remote batch processing and remote
job entry tasks.
That all interactive applications be homed on one computer
rather than employing dedicated computers.
The COINS experiment should be evaluated at the earliest
moment to provide guidance for Agency participation in community
efforts.
That the IPB concentrate on and define standards for ADP,
particularly for applications serving more than one component.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 200215a/Os1 l 'CIS-RDP81-00261 R000300130001-8
Office of Computer Service Activities
The Office of Computer Services developed from an amal-
gamation of the business and administrative data processing acti-
vities of the Agency and the rapidly growing scientific com-
putational activities which evolved from the creation of the
DD/SET.
The Office of Computer Services: 1) operates a general
computer processing center for the Agency and 2) provides com-
puter programming and computer applications design as a service
of common concern.
Customers tend to be pleased with the service they receive
from OCS. Everyone would like his work done more quickly but
few customers thought their processing requests received less
than satisfactory attention.
The rapid and dynamic technical changes in the entire com-
puter industry has engendered uncertainty on the part of OCS
personnel and customers as to what to expect in the near term.
As system changes take place they usually have a direct impact
on individual applications with an attendant reworking of programs
that does not contribute to the utility of the application.
We have recommended that a mechanism for communicating plans
for major system changes be instituted.
It is the ASPIN opinion that problem solution is and must
be separate form computer operations. This feeling contributes
to our recommendation that applications programmers be assigned
to customer offices for tours of duty.
Approved For Release 2002/05101 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/0S/0-l"'61' q -RDP81-00261 R000300130001-8
A complete set of procedures should be published and
maintained for the use of customers.
A strong computer graphics capability should be developed
by OCS in concert with NPIC/AID and ORD/An.
OCS is encouraged to acquire and test a proprietary general
data management system; this action should be coordinated with
user offices and with other major computer centers within the
Agency.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/05/01.: CiA RDP81-00261 R000300130001-8
A Central Reference System
Considerable study and analysis of a central reference
system within the Agency was undertaken by ASPIN. We arrived
at a conclusion that the reference facility of the future
should be one which would route a request for information to
the appropriate substantive source rather than a reference
service being a sole source or repository of information used
by and generated by the Agency. This judgement is based, primarily,
on the increased specialization which has resulted in information
tending to reside with those who have the ability to interpret it
and relate it to other activities.
A central reference system should provide also for the
rapid and accurate dissemination of intelligence information
and for the control of this information.
ASPIN has recommended:
That CRS be the point of contact for requests for
intelligence information from outside or within the Agency.
Only those data which are accessioned or generated by
the reverence center will be provided in response to direct
requests. All other data sought will be requested from
other information centers which have the resources to
respond.
Current work on the automated dissemination system
should be maintained.
Ultimately all SI, State, Defense and Agency positive
intelligence should be disseminated via a machine-aided
system.
Approved For Release 2002/05/01 CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/091 `y =F DP81-002618000300130001-8
Substantive analysts should be encouraged to influence
the indexing efforts of CRS.
A standardized document reference number scheme should
be developed -- initially for internal Agency use, ultimately
for community-wide use.
A standard scheme of document indexing by a reference
center should be established by high level management
decision.
The CRS should create an index, or indices to other
organized collections of information in the Agency.
The present system of document storage and retrieval
should be maintained; the speed of delivery of documentary
material should be improved.
An extensive interactive capability with regard to
the CRS intelligence document index should be developed
and tested as quickly as feasible.
Approved For Release 2002/05/01 : CIA=RDP81-00261 R000300130001-8
Approved For Release 2002/05/0=1-:-GIDP81-00261 R000300130001-8
Research and Development in Information Processing
We observed a number of problems associated with R&D in
ADP matters as conducted by the Agency, the more significant
are:
The physical separation of ORD/An does little to con-
tribute to user -- developer communications.
Most RFD has been conducted by external contractors
with ORD/An serving as a broker and furnishing funds. This
has led to undue influence on development efforts by non-
Agency elements and insufficient influence by the customer
by virtue of not controlling the funding of projects.
The RFD efforts of ORD/An have not been sufficiently
esoteric and sophisticated to establish a clear dichotomy
between the programs pursued by ORD and those of OCS.
R&D has had a tendency to be oriented toward the novel
in terms of hardware rather than tailoring projects to
work on OCS machinery.
ASPIN has recommended that the DD/SFT review the division
of effort between ORD and OCS in the area of ADP with the end
in view of:
a. Moving problem definition and computer application
design and development to OCS from ORD.
b. Transferring standard ADP equipment from ORD to
OCS to provide OCS with the facility to conduct appropriate
RFD.
c. Transferring other equipment to anticipated develop-
ment programs or declare it surplus.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/05/01 CIA-13DP81-00261 R000300130001-8
d. Conduct a review of existing ORD contracts via the
IPS and prospective users to determine which should be
continued and under whose cognizance.
e. Subsequent ADP equipment or software test and ana-
lysis be conducted by OCS except where the items are in
direct support of a special processing center such as NPIC.
f. OCS issuing a current awareness publication to
announce new activities, products and developments of
general interest to those engaged in ADP.
ASPIN recommends that the DD/S$T and the IPB reject the
proposal of the IHC which would provide for a community wide
R&D center.
Finally we recommend that R&D projects in the realm of
ADP be submitted to the same scrutiny and review as for other
significant projects within the Agency.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002%O5101 ~, "-CiA-RDP81-00261 R000300130001-8
Organizational and Management Elements of Automatic Data
Processing
This part of the ASPIN Report should be read in its entirety
by senior officials of the Support Directorate. Its conclusions
and recommendations are far-reaching and have both direct and
indirect impact on the Directorate. The conclusions and recommen-
dations must be read within the context of the total discussion
of Part VI of the report and with the paper prepared by the
25X1A contract consultant firm, The appendices
of this part of the report should be of particular interest
to the Office of Logistics and all elements of the Support
Directorate engaged in ADP activities.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/66/01-;*, 1 LA-DP81-00261 R000300130001-8
Office Studies
It is difficult indeed to reduce the fund of statistical
information and observations made in the individual office
studies to a simple summary. The material contained in these
studies covers a wide range of specificity and subject matter
and the conclusions and reocmmendations reached have been incor-
porated in the general Parts of the Report.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/05/01 6G1A-I4QP$1-00261 R000300130001-8
25X1A Working Papers
During the course of the ASPIN Study certain problem areas
were identified that appeared amenable to solution by an outside
contractor. A contract was let with
conduct further study of these problems and Annex II of the ASPIN
Report contains a set of 6 working papers produced by the con-
tractor. A summary of these working papers is presented below.
TEXT PROCESSING SYSTEMS
CIA is developing several mechanized systems with text
processing capabilities for current awareness and/or retrospective
retrieval. Results to date indicate that marked qualitative
improvements in the dissemination process will be realized and
that dramatic improvements in retrospective retrieval will be
experienced. Consolidation of development effort, preparation of
an Agency-wide text processing plan, and elimination of some of
the current systems appears a highly desirable goal.
There is a good deal of duplication of effort in text
processing within the Agency. The thrust of the contractor's
recommendation is that these efforts should be consolidated and the
total talent presently engaged could develop a viable system of
considerable quality.
COMPUTER GRAPHICS
Computer graphics appears as yet to have found little appli-
cation within the production components of the Agency. Work has
been done and is proceeding in three general areas which could be
the basis for more extensive use of computer graphics: Automated
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/05/:' RgP81-00261 R000300130001-8
Map Making, Assistance to Photo Interpretation, and Assistance
in Visual Cognition of Large Data Bases.
Computer graphics software development and operational costs
are high. Efforts must be made to prevent the proliferation of
uncoordinated development in this area which is bound to follow
an increased interest in graphics applications.
ORGANIZATIONAL AND MANAGEMENT ASPECTS
The present highly decentralized management of the ADP has
numerous virtues, including the encouragement of local initiative,
freedom from an ADP bureaucracy, and the ability to satisfy well
defined,short-term ADP needs. But it also suffers from a lack of
longer-range vision, insufficient planning, inadequate control,
inability to profit from experience and the absence of a central
ADP focus.
Recommendations were made to:
a. strengthen the existing central review of hardware
acquisition.
b. consolidate headquarters computer operations into
no more than two units, and initiate a charge-back system
for computer utilization;
c. establish a career service for ADP personnel;
d. retain local initiative in applications development
but assign technical/functional leadership responsibility
for broad areas of ADP development;
Approved For Release 2002105/01 CIA7RDP81-00261 R000300130001-8
Approved For Release 2002/0501 ; CIARDP81-00261 R000300130001-8
e. in general, continue local planning for ADP support,
but strengthen central review.
f. establish a full-time post of ADP advisor to the
Executive Director.
This entire paper should be reviewed by senior officials
of the Support Directorate.
TEXT STREAM PROCESSING: DRAFTING, EDITING AND PRINTING
Text stream processing, in the eyes of the contractor,
undoubtedly can find wide applicability in the Agency, provided
the system designs can be kept fairly simple and that development
efforts be useful to several activities.
The contractor recommends:
a. A small task force be assembled to obtain more
detailed information of this subject within the Agency to
uncover the characteristics, requirements and priorities
of those activities which could benefit from text stream
processing and the composing process.
should be encouraged to examine how plans
for automating the
can be accelerated, and to
make use of text stream processing at an earlier date than
presently proposed. It is felt that the methods developed
could have applicability to other activi-
ties uncovered in item a above.
Approved For Release 2002f~5/01 : OA-RDP81-00261 R000300130001-8
Approved For Release 2002/Ctl3/01= 'C1A-RDP81-00261 R000300130001-8
MONITORING, FILE PROCESSING AND COMPUTATIONAL SUPPORT TO MISSILE
AND SPACE RELATED SYSTEMS
This paper treats ADP support systems for FMSAC, OSI and
OEL and covers both scientific and management systems employed.
Two aspects of the paper, signal processing and machine-
aided dissemination would be of particular interest to the Office
of Communications. The paper is general in nature, leaning more
to the need for restructuring of files rather than to sophisti-
cated scientific applications.
Approved For Release 2002/05/01 CIA-RDP81-00261 R000300130001-8
Approved For Release 2002105101 :CIA-RDP81-00261R000300130001-8
STRUCTURED FILE SYSTEMS
There is a very large number of seemingly uncoordinated and
overlapping development efforts within the Agency addressed to
coordinate/hierarchie files. This, despite the similarity of
requirements, file structures, and of computer processed employed,
despite the burden that will be placed on users if they have to
learn different languages, and despite the scarcity of in-house
talent available. Such consideration argue for a radical reduction
in the number of development efforts and the consolidation of the
remainder.
The Agency should be able to buy, rather than make, its next
generation of general purpose file handling systems.
The low access rate to most of these files, economics of scale
in providing backup processors and direct access storage, operational
complexities, and the difficulties of implementing and maintaining
a bug-free operating system, argue for concentrating on-line
services in the smallest possible number of computer centers.
Separate development efforts for SANCA and the DD/P's STAR
now seem inevitable; but every effort should be made to assure that
they can be used in a time-sharing environment with other on-line
systems.
For the major Agency files (AEGIS, MISTAC, et al), a formal
development plan for on-line processing, with projections of
goals, costs and milestones over at least a five year period, should
be put together in a standard fashion.
Approved For Release 2002/05/01 : CIA-RQP81-00261 R000300130001-8
Approved For Release 2002/Q~/Q~'~; ~jC~lA-8
DP81-00261 R000300130001-8
The Agency should decide, as a matter of policy, that on-
line file-based services will be provided out of the fewest
possible computer centers at headquarters.
Approved For Release 2002/05/01 CIA-RDP81-00261 R000300130001-8
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
14EMOf AWN FOR:
/4'2'/V~ A D\
I.\?,-(\-
wy-
O GN 54 11)I WHICH RELACES
4AY FORM
I AU
BE USED10I
MEMO AND FOR: 11 -
(DATE)
.
\AA
FORM 101 RELACES FORM 10.101
BE USED.
Approved For Release 2002/05/01 : CIA-RDP81-00261 R000300130001-8
(DATE)