COINS
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP79M00096A000300060022-1
Release Decision:
RIPPUB
Original Classification:
C
Document Page Count:
8
Document Creation Date:
December 16, 2016
Document Release Date:
June 13, 2005
Sequence Number:
22
Case Number:
Publication Date:
May 5, 1972
Content Type:
REPORT
File:
Attachment | Size |
---|---|
![]() | 371.44 KB |
Body:
Approved Fofelease 200e1010/ '1EIA,R6*9M00098AOO0300060022-1
5 May 1972
When the Information Handling Committee, USIB, was
established in April 1968, the largest and most noticeable
s
project it inherited from CODIB was the Community On-line
Intelligence System (COINS).
COINS was the result of the recommendation (among
others) made by the President's Foreign Intelligence Ad-
visory Board in 1965 that the intelligence community might
improve its "backward" information handling techniques by
extending to the Washington intelligence community the TIPS
on-line system then being developed at NSA. (Recommendation
2, PFIAB memorandum of 15 June 1965, as Attachment A to Ex.-
Director Memo of 18 May 1966.) For security reasons NSA re-
jected having any TIPS terminals outside their Ft. Meade
Headquarters. The idea of the COINS system was modified`lpy
degrees in 1965-66 into a network of computers at CIA, NSA,
DIA, and NPIC (with State as a terminal on the DIA system)
joined by a store-and-forward switch with each agency making
available to the others some of its own data files. During
the period when the utility of such a system was being tested
(COINS I), the store and forward switch would be at DIA.
However, during COINS II, the system's operational period,
Approved For Release 2005/06/23 : CIA-RDP79M00096AO00300060022-1
CONFIDENTIAL
0
C O N F I D E N T I A L
Approved FdCelease 2005/06/23 : CIA-RDP79M000?000300060022-1
the system switch would be permanently fixed at CIA (CODIB,
COINS Implementation Plan, 25 May 1966, Secret.) CIA had
agreed on COINS II in 1966 (Planning Memorandum from the Ex-
ecutive Director-Comptroller to the Deputy Directors, 18
May 1966, SECRET). By 1968 the idea of COINS II had faded,
however, and in the 70's41 term, COINS II, came to mean
something quite different. (NSA, COINS 146, COINS Master Plan,
2 December 68, Secret.)
During the 1965-68 period COINS had been a CODIB pro-
ject, with the COINS Committee a subcommittee of CODIB. There
had been considerable slippage in the program, difficulties
in gluing the disparate hardware together, and some friction
among the various people associated with the experiment.
Consequently, there were discussions on the management of
COINS about the time that the formation of the IHC was being
considered. The resulting change came out as a memorandum
for Director of NSA from the DCI (undated, but issued about
31 May 1968) appointing NSA as the Executive Agent for COINS,
a Project Manager, and subsystem managers. It left the IHC
with only two responsibilities concerning COINS:
1. "The Chairman, IHC, will provide your (Dir. NSA's)
project manager guidance on the substantive intelligence re-
quirements of the system, including a list of primary files
to be included in the system for the purpose of evaluation,"
and;
Approved For Release 2005/06/2.3:,VIA-RDP79M00096A000300060022-1
C 0 N F I D E N T I A L
C O N F I D E N T I A L
Approved For Rele 2005/06/23 : CIA-RDP79M00096A08W00060022-1
2. ''A COINS Evaluation Panel will be established by
the IHC. This Panel will be working very closely with your
project manager and it is requested that he be directed to
give them every assistance in their evaluation work. Their
principal effort will be during the period January to July
1969 with a report to the IHC as soon as practical thereafter."
The July 1969 date for the completion of the COINS Ex-
periment evaluation seems ironic at the time of this writing
two and a half years later. NSA has requested a delay in the
evaluation period several times, until the date for such an
evaluation is an uncertain period in the future. There were
a number of reasons for delays. There were constant equipment
changes as the agencies upgraded their systems-; with resulting
complications in interface, software, etc. There was a very
lukewarm response by various agencies' analysts to a system
that, through the lack of standards, was very difficult for
a non-ADP person to query, and required the use of numerous
look-ups. The COINS System Manager made many valiant efforts
to develop users' groups in the agencies, but it was an uphill
struggle to overcome,analysts' inertia with the files and
the type of system provided. (COINS Semiannual Review, 1
Jan-20 June 70, COINS/329-70, p. 25, Secret.) The basic COINS
paradox was this: not enough analysts found enough to
interest them in the COINS files to justify the effort in
Approved For Release 2005/06/23 CIA-RDP79M.00096A000300060022-1
C 0 N F I D E N T I A:L
C; U N F I !) t N 1 i A L
Approved For Rel6 a 2005/06/23 : CIA-RDP79M00096ACS@600060022-1
25X1
querying such a system, while the agencies did not want to
spend the considerable amounts necessary to build machine
files more interesting to the Analysts just for a community
experiment. If there had been a larger number of machine
files already available when the experiment began, or if the
system technically had been easier for analysts to query
sucessfully, this problem might have been circumvented. But
they were not, and several years later it was still a frus-
trating system for the non-technical to query. (COINS Semi-
annual Review, 1 Jan-30 June 71, COINS-277-71, p. 7-8, Secret.)
Even the two responsibilities left to IHC, files and
evaluation, were not easy to carry. In its last year CODIB
had set up--a Panel on Files for use in COINS, under
then Deputy Director of Central Reference, CIA.
Twenty-one files, whose chief virtue was that they were
already available in machine language,;were offered to COINS,
most of them by DIA. That Agency was preparing its own
DIAOLS (DIA On-Line System), and since the COINS switch was
at DIA, DIAOLS files could concurrently run on COINS. The
Marengo Panel, however, felt that quality was more important
than quantity, selected ten files they felt would make a
good corpus for a test of the feasibility and utility of
COINS, and left a reasonable criteria for selecting and ap-
proving future files. The ten files reflected a bias toward
25X1
Approved For Release 2005/06/23 : tlA-RDP79M00096A000300060022-1
C O N F I D E N T I A L
C 0 N F I D E N T I A L
Approved Fo lease 2005/06/23 : CIA-RDP79M00096 000300060022-1
in response to the PFIAB recommendation,
and were chosen because they were of interest to analysts
in more than one agency--but always within the limitation
that they were already available in machine form. (Report
of CODIB Panel on Files for Use in COINS, CODIB-D-113/4.9,
8 March 1968, Secret.)
However, though recommendations
were approved by CODIB as one of its last acts, and not dis-
approved by IHC, there was still a school of thought that
everything available should be put on COINS to attract the
widest possible use. Though not sanctioned by IHC, nearly
thrity other available files did appear at one time or the
other, with files divided into a compromise of "PRIMARY FILES"
(the ten recommended by
which would be
used in the evaluation of the system) and "SECONDARY FILES"
(which would be used to attract users who had no interest in
the PRIMARY FILES). (COINS 146, Master Plan, I, 3 Dec. 68,
p. 16, Secret.) It was obvious to most people that the
addition of the secondary files would undermine the Marengo
Panel's purpose in selecting a quality test corpus, because
would be almost impossible to separate out "primary" queries
from "secondary" queries when they were automatically logged
at the switch. The clumsy separation of the files into primary
and secondary (with two different procedures for adding and
deleting the sets of files) disappeared over the years. In
Approved For Release 2005/06/23 5 CtA-RDP79M00096A000300060022-1
X1
I
2$X1
C 0 N F I DE N T I A L
C O N F I DENT I A L
Approved Fir Release 2005/06/23 : CIA-RDP79M0W6A000300060022-1
early 1972, with the system still unevaluated, only four of
the original tent Ichoices still remained on COINS-
to NSA's computer,from their original agencies.
The Evaluation has been even more difficult to achieve.
The original implementation plan (May 1966) had invisioned a
technical evaluation--simply to see that it could be done--
about six months after the system became operational, or some
time in 1967, with an evaluation of the system's usefulness to
the analyst, later and after more thought based on experience.
This timetable could not be met, but there has been technical
data on the performance of the system, based
at terminals and an automatic log,1(p.
on manual logs
e.g , COINS Semiannual
Review, 1 Jan-30 June 70, COINS 329-70, Secret)[ at the switch,
ever since the system switch went operational.' This infor-
mation has been collected and evaluated by a COINS Test and
Evaluation Panel, which has met since February 1968. The
Panel drew up a Test Plan (IHC-D-113.4/11, 31 May 1968) which
had a data gathering.schedule ending 30 June 1969, with
conclusions and decisions by September 1969. After NSA had
requested and received delays in the evaluation timetable the
whole question remained dormant for several years.
Approved For Release 2005/06/23 6CIA-RDP79M00096A000300060022-1
C. 0 N F I D E N T I A L
C 0 N F I D E N T I 'A L
Approved ForIease 2005/06/23 : CIA-RDP79M00000300060022-1
However, the question of evaluation came to life again
in late 1971. In the early days of the IHC, CIA had been
anxious to get on with the evaluation of COINS, to see if its
usefulness matched its costs. But at the Forty-Second IHC
meeting (15 December 1971) the CIA member expressed some
misgivings on the subject of evaluating COINS. He argued
that there would be a successor system to COINS, and this
system might be quite different from the present COINS, utting
nN %T 14
money into contractor support, testing, etc./r
ay not be cost-
effective. At the same meeting, the COINS Project Manager,
in briefing the IHC on COINS progress, mentioned that the
problem of testing the utility of COINS had been troublesome.
The Project Management Office had never been in a position to
do a utility evaluation of COINS, so they had contracted with
Dr: Ed Adams of IBM Research Labs to look at the problem and
develop utility indices, by which the system could later be
measured.
Dr. Adams briefed the IHC on these indices at the Forty-
Third meeting, 19 January 1972. In his introduction of Dr.
Adams, the COINS Project Manager stated that he thought
was still too soon to evaluate COINS, but not too soon to
develop evaluation techniques. He wished to know how hard it
would be to collect the data. If it looked good and when the
system had stabilized, they would have meaningful figures.
But in the meantime he had no measure of system utility. The
- 7 -
Approved For ~eI a POIS/0&2h3 : fIf -TPA79jfl00096A000300060022-1
Approved r Release 2005/06/23: CIA-RDP79MON96A000300060022-1
Chairman, IHC, felt at that time that Dr. Adams had not yet
progressed far enough to address the larger question--convincing
the budget people of the utility of the system. While COINS
could be compared with a manual system in, for instance, a
crisis situation, there still remained the question whether this
was the correct system to choose.
Howvering behind the question of a COINS evaluation had
always been a fear that has not been dissipated by any reports,
programs, or future COINS plans. That fear is that the system
may never be truly evaluated but will evolve into some follow-
on system that must be accepted as an accomplished fact--without
anyone ever truly knowing what the community's real requirements
for a networl are, nor if COINS is the best answer for these
requirements.
Approved For Release 2003/0'6/23 : CIA-RDP79M00096AO00300060022-1
C 0 N F I D E N T IA L