BRIEFING ON CRS 'AEGIS' SUBJECT SEARCH SYSTEM

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78B05703A000300010030-9
• Release Decision: RIPPUB
• Original Classification: C
• Document Page Count: 5
• Document Creation Date: December 28, 2016
• Document Release Date: December 31, 2003
• Sequence Number: 30
• Publication Date: March 13, 1970
• Content Type: MF

Approved For Release 2004/02/11 : CIA-RDP78BO5703A000300010030-9
PSG/AID-90/70
13 March 1970
Vf-
MEMtDR UM FOR: Executive Director, NPIC
SUBJECT: Briefing on CRS "AEGIS" Subject Search System
1. In reporting on the CRS public relations effort in support of
their AEGIS effort I must make clear at the start that this is not a
consolidation of the reactions of the NPIC personnel present at-We brief-
ing. This is my own analysis and is, I am certain, at variance with the
reactions of many of the NPIC personnel present at the briefing.
2. The AEGIS presentation was in two parts. One part was an ex-
position of the AEGIS subject search system itself. The other part was
a quite unrelated demonstration of the E ::]Rapid Search Machine (RSM). 25X1
The AEGIS system has replaced the several diverse indexing systems (Intell-
ofax, Special Register, etc.) formerly operated on a functionally-structured
basis by CRS with a single "shallow" indexing system that is operated by
areally-structured indexers and retrieval specialists. This approach
recognizes that most of the great volume of documents indexed by CRS will
never be recalled. Thus the way to minimize the overall system cost is
to reduce the effort spent on indexing the input documents, and concentrate
instead on the formulation of strategies for searching the indices to
answer requests against CRS and also on reviewing the citations produced
by such searches in an attempt to eliminate irrelevant citations from the
responses furnished by CRS to its customers. The system is working quite
well, according to individuals from Director down to Division Chief level
in CRS and also according to our "independent" consultant (whose report
CRS is distributing to all who pause long env one thrust in
their hands). I suppose that I must menti at the indices
are searched on an IHM 360/50. I will als ssert that the
information in the preceding sentence has nos an any assessment
of the impact (real, potential, or imagined) of the AEGIS system on this
Center. What is of importance to the Center are the answers to the follow-
ing questions:
(1) Can the information services now available through Central
Reference Services be efficiently used to provide more support to
the NPIC image interpretation process than is presently the case?
Declass Review by NGA 1 Excluded ROfrom UP automatic,
Approved For Release 2004/02/ Iuajniraemg one
L?u",USsificatfan
 VUI1I IULIV I it1L
Approved For Release 2004/02/11 : CIA-RDP78B05703A000300010030-9
SUBJECT: Briefing on CRS "AEGIS" Subject Search System
(2) Is there any lesson to be learned by this Center from the
CRS experience with shallow indexing?
t the first of these questions can only be answered by the
hoto interpreters and the various individuals, of whatever
title, who channel "collateral" information to the photo interpreters.
If the answer to the question is "yes" the action to achieve such in-
creased support must be undertaken by those who serve as suppliers of
information to the P.1.'s.
3. The second question posed above is one that centers around the
cost-effectiveness of the exhaustive indexing that has been done in the
past for documents committed to the Minicard system and which will appar-
ently be carried forward with only slight modification and extension for
the EPDF. It must be recognized that as soon as one starts searching
for information contained in documents (or their micro-images) by examin-
ing surrogates rather than by reading through the total set of documents,
one is going to have a certain number of misses. In CRS' experience the
reduction in number of misses that would result from in-depth indexing
as compared to their shallow indexing approach would not be worth the
cost. Obviously our refere services are a rather different kettle of
fish than the man it may well be that the question can be
answered with categorical " " But as one observing from the side-
,---
4. The RSM pitch was presented (and received) in a vein of "Gee
Whit, isn't this wonderful" with all the discussion centering around the
machine and being couched in anthropomorphic terms. The real question
is, of course, whether or not the ability to search for occurrences of
words (I am well aware that complex Boolean search functions are possible)
in a corpus of text is sufficient (or perhaps even more than sufficient)
to obviate the need for indexing the documents or sub-elements of that
corpus. It is easy to generate interesting and amusing examples of how
one can find all sorts of items from a file of running text but these
examples shed little light on the basic question. Only when given an
answer to that question is it appropriate to start evaluating the cost-
effectiveness of the various ways such a capability might be achieved.
The RSM is but one of these ways and I would think it most unwise for
this Center to jump in this direction simply from an assumption of
effectiveness for the method.
25X1
Chief, T crmicai ng a ,AID/PSG
Approved For Release 2004/ 1,:iJ 2
IPf1A 5703A000300010030-9
 jj I I IIJLIII I I11L
Approved For Release 2004/02/11 : CIA-RDP78BO5703A000300010030-9
SUBJECT: Briefing on CRS "AEGIS" Subject Search System
Distribution:
Orig & 1 - Addressee
1 - NPIC/PSG
1 - NPIC/PSG/AID
1 - NPIC/PSG/AID/TPS
Approved For Release 2004/02/ R 05703A000300010030-9
., Illl
 Approved For Release 2004/02/11 : CIA-RDP78B05703A000300010030-9
MEMORANDUM FOR: Al
6/0 OLO f -
7' fs Pvc'/,e,
-7 0
IUI runm 10. 101
WHICH MAY BE USED.
~ ~OL~ .00 L
(
DATE)
Approved For Release 2004/02/11 : CIA-RDP78B05703A000300010030-9
 CENTER- UTING SLIP
pROwted For Release 2004/02/11: CIA-RDP78B05 % 4 00300010030-9
I Technical Planning Staff, AID
13 March 1970
TO
INITIALS
DATE
REMARKS
DIRECTOR
DEP/DIRECTOR
/~ < 3 Pi I
EXEC/DIRECTOR
.
,j
SPECIAL ASST
ASST TO DIR
Y
ASSTTO DEP/DIR
~1//11'C4yyr
CH/PPBS
DEP CH/PPBS
A EGA ma'y` )At
EO/PPBS
p~ if
"1I
~ /~Yy,OD
CH/IEG
-
DEP CH/IEG
EO/IEG
CH/PSG
DEP CH/PSG
EO/PSG
/AID
I
CH/TSSG
DEP CH/TSSG
EO/TSSG
CH/SSD/TSSG
PERSONNEL
LOGISTICS
TRAINING
RECORDS MGT
SECURITY
FINANCE
DIR/IAS/DDI
CH/DIAXX-4
CH/DIAAP-9
CH/SPAD
25X1