PHASE II FINAL REPORT VOLUME III IMPLEMENTATION PLAN FOR INITIAL SYSTEM
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP78-03952A000100030001-9
Release Decision:
RIPPUB
Original Classification:
S
Document Page Count:
75
Document Creation Date:
November 16, 2016
Document Release Date:
February 14, 2000
Sequence Number:
1
Case Number:
Publication Date:
March 1, 1965
Content Type:
REPORT
File:
Attachment | Size |
---|---|
![]() | 3.32 MB |
Body:
Approved For Release 200 il~ P78.03 0400030001-9,,
PROJECT
CHIVE
DIRECTORATE OF SCIENCE ANC ":CI
0010 0
an - ved r 1 L a 2. _ _... ~ O 00030001-9
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
WARNING
This material contains information affecting
the National Defense of the United States
within the meaning of the espionage laws,
Title 18, USC, Secs. 793 and 794, the trans-
mission or revelation of which in any manner
to an unauthorized person is prohibited by law.
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
~VQNFIDENTIAL..
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
Phase II Final Report
IMPLEMENTATION PLAN
FOR THE INITIAL SYSTEM
CHIVE/R-3-65
sAC _ ! -
E Df 0
.-_ gY
kiRie,
01
TYPE
GLASS
CRIG CLASS
CIS
REV
,LUST
NEXT REV gWa- AUTH !RR 1fl?2
CONx IDENTI
Approved For ease 2000/05/
RDP78-03952A000100030001-9
.. o
Approved For Release 2000/05/OGWA-RDP7?'4~ 100030001-9
Page
3.1 Summary Description of Phase III Product 1
3.1.1. Recommendations for Initial Increment 3
3.1.2. Organization Needed to Achieve Goals 14
3.1.3. The Operational Testing Concept 18
3.2 Implementation Tasks
3.2.1. Basic Schedule
3.2.2. Personnel Tasks
3.2.3. Logistics Tasks
3.2.4. Data Specification Tasks
3.2.5. Procedures Tasks
3.2.6. Programming Tasks
Approved For Release 2000/0 ZA-RDP78-03952A000100030001-9
CONFIDENTIAL:
Approved For Release 2000/05/0
E f 189030001-9
Page
3-1
Alternative Initial Organization Configuration
4
25X1A
3-2
Volume and Source of - Intellofax Inputs
for 1963
10
25X1A
3-3
- Annual Production
3-
d A
l V
l
f
E
ti
t
25X1A
4
ume o
ma
nnua
o
s
e
25X1A
3-5
- Intellofax Requests for 1964
3-6
Task Assignments (OCR)
39
3-7
Task Assignments (OCS)
40
Approved For Release 2000/05/A-RDF7?,p~130001-9
SSEE } `t i F < r
Approved For Release 2000/05/08 A-RDP 8 0395 A Op100030001-9
Page
3-1
PERT Chart for CHIVE System Development
36
3-2
Task Schedule
37
3-3
Manning
38
Approved For Release 2000/0 A-RDP78-03952A000100030001-9
Approved For Release 2000/05/Q RDP"91140101011630001-9
Chapter 3.1.
SUMMARY DESCRIPTION OF PHASE III PRODUCT
The basic goal of the Phase III effort is the creation
25X1A
of a well-established and organized EDP environment
and the construction of an initial system component
which shall be judged ready for operational testing.
This component will be an experimental branch devoted
to work on The operational test period
should provide sufficient information to enable Agency
management to make a decision on whether or not to
proceed to full system implementation.
The initial operational component, so far as is possible
within the limitations of initial programming and
available manpower, should:
a. Provide a base leading toward the ultimate
system objectives, i.e., broader coverage,
more specific indexing, single point service,
an all-source data base, capability to answer
more complex questions, etc.
In addition, it should:
b. Have a limited, well-defined data base.
c. Benefit more than one customer group.
d. Provide the seed from which the total system
can grow in an evolutionary, not revolutionary,
manner.
Approved For Release 2000/05 JA-RDP7-
SUMMARY DESCRIPTION
3.1.
"In
Lt.
w19.0001-9
4
Approved For Release 20 58 : Cl
fjj)J0100030001-9
e. Require a minimum of input and retrieval coor-
dination with the existing system.
f. Offer a reasonably controlled environment for
testing both man/machine relations as well as
management techniques.
g. Represent a wide range of the problems anticipated
in the final system.
h. Not have a disrupting effect on current service
provided by OCR during the build-up period.
Indexing and file structuring concepts will be
tested and refined and the operational simulation period,
during the latter months of Phase III, should provide
sufficient information to assist in making a final
determination on the operational readiness of the initial
increment.
The initial design will not stress sophisticated or
unusual techniques per se. Humans will continue to do those
things they are able to do best. While the ultimate goal
may be extensive automation of the data input/output
process, the initial steps must settle for less--a
partially automated system utilizing a computer to
mechanize all tasks which clearly need not be performed
by hand.
While creation of a well-established and organized
EDP environment may require a heavy initial investment
with no immediate payoff when compared to the data base
SUMMARY DESCRIPTION
3.1.
-2-
Approved For Release 200 : ClP 8 O ~A000100030001-9
Approved For Release 2000/OWU CIA-U0 MOIA00030001-9
- tT
to which the investment is applied, the long-range
objective of a broader data base, together with new
concepts in system information management, would
ultimately be satisfied with little additional investment
in EDP techniques.
3.1.1. RECOMMENDATIONS FOR INITIAL INCREMENT
A further description of the organization, manning
requirements, and tasks of the Test Branch is 25X1A
included in later sections of this volume. The discussion
which follows is included to set forth briefly the rationale
behind the decision to begin initial system activity
through the mechanism of the - element.
3.1.1.1. Geographic Approach
While CHIVE has argued, in earlier documentation, that
the central document/information reference system should
be organized along geographic lines, this view was not
automatically imposed on the initial system. Rather, a
number of possible implementation paths - by document,
by subject, by named object, and by area - were explored
before a final selection was made. These various alter-
native approaches are examined below.
SUMMARY DESCRIPTION
Recommendations
3.1.1.1.
25X1A
Approved For Release 2000/0 IA-RDP 0 c1130001-9
Approved For Release 2000 T &I@ PtgF- PAt000100030001-9
- -fq
3.1.1.2. Single Area Approach
In addressing the problem of selecting a point of
entry from which the full scale system could evolve,
the guidelines and objectives set forth at the beginning
of this volume were adopted.
The table below portrays in summary form the extent
to which various alternative initial system configurations
appeared to satisfy those guidelines and objectives.
Table 3-1
Alternative Initial Organization Configurations
Alternatives Meets Objectives Violates Objectives
1. All documents, a, c, e, g b, d, f, h
all subjects,
all countries
2. Single document b, f, h
type,
all subjects,
all countries
3. All documents, g
single subject,
all countries
a, c, d, e, g
a, b, c, d, e, f,, h
4. All documents, d, f, g, h a, b, c, e
single named object,
all countries
5. All documents a, b, c, ** d, e
all subjects, f, g, h
single country
* Depends on document category chosen.
**Depends on country chosen.
SUMMARY DESCRIPTION
Recommendations
3.1.1.2.
Approved For Release 2000/D5/08 : CIA-C1'0030001-9
QCRET .. .
U'. T
Approved For Release 2000/05/0 `1A-RDP78-03952A0 100030001-9
Alternative 1 is obviously unrealistic since it
literally requires implementation of the complete system
right at the very start. This means we would be confronted
by every type of documentary problem, all indexer support
tools would have to have been developed in advance, all
plans worked out for supplanting every component activity
of the existing system, all personnel reassigned and
retrained, and so forth. More specifically, such an
approach would violate the objective of a limited, well-
defined data base; it would require a revolutionary
development program, it would fail to provide a suit-
able environment for testing purposes; and the effects
of failure might well be calamitous.
Alternative 2, while it would offer the limited data
base and controlled environment we desire, is unattractive
because it would not incorporate some of our ultimate
system objectives (e.g., all-source document handling),
and it would serve only those customers interested in this
one source. In essence, the SR entry point, for example,
merely defers (but does not solve) the problem of how the
system is to evolve. Once Comint materials were under con-
trol, we would be faced with the task of incrementing
the system by source, which would be most difficult since
SUMMARY DESCRIPTION
Recommendations
3.1.1.2.
Approved For Release 2000/054I.4k-RDP78-03952AO00i..QG030O01-9
Approved For Release 2000/(g! CIA-RDP78-03952A000100030001-9
25X1 B
there is little to choose between the extremes of a
single document series (e.g., State airgrams) and all
collateral documents. Moreover, if we chose to evolve
by document categories, the problems of coordinating with
the existing system would be immense. Finally, the
system would be criticized for having failed to address
a sufficient variety of document problems which could
be anticipated in the full-scale system and, therefore,,
having proved nothing about its long-run potential.
The notion represented by Alternative 3 is that
the system would process all documents reporting infor-
mation on all foreign countries, but that the subject
matter of these documents would have to pertain to a
single, though broad, subject area of intelligence interest,
e.g., economic affairs, military activities, or other.
Within the confines of this topical restriction, any
kind of data might be stored and retrieved, including
installations, locations,
activities, and events.
If we began with such a system, it would presumably
serve only one major component of the Agency, such as
OCI or OSI. Its most important disadvantage, however,
would be that the document base involved would inevitably
SUMMARY DESCRIPTION
Recommendations
3.1.1,2.
Approved For Release 2000/25 .. CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/0%~F~A-RDP78-03952A000100030001-9
25X1 B
overlap with the existing system with consequent
duplication of activity. This we can ill afford in view
of the already tight manpower situation. Any request
on the new system, with the exception of those asking
only for the most recent information, would also have
to be levied on the old since the established files
could not be readily broken apart. In summary, the
complexity of operating a system of this kind concurrently
with the old would be great, its data base would be
larger than we would want to handle in the initial system,
and the consequences of failure would be serious.
Alternative 4 would give us a variety of documentary
inputs and the full spectrum of area problems. Only the
type of data to be processed and retrieved (e.g., 25X1B
- organizational, or other) would be restricted.
Unfortunately, like Alternative 3, it would be impossible
to separate the documents on this basis., with the result
that both systems would receive essentially the same
inputs. Its other serious disadvantages would be that
the initial system could not test its integrated indexing
concept, the data base would be virtually that of the final
system, and it would require close coordination of request
handling with the established OCR systems.
SUMMARY DESCRIPTION
Recommendations
3.1.1,2.
Approved For Release 2000/05/0?-RDP78-03952A000100030001-9
Approved For Release 2000/68WT CIA-RDP78-03952A000100030001-9
Alternative 5 seems to offer the most advantages and
fewest disadvantages of the options listed and was the
one selected. Using this approach, a single country
would be selected as the entry point but the entire
range of documentary receipts on that country, and all
types of subject matter would be handled. This would
permit us to identify an initial geographical component
of our total organization and test it in a real envir-
onment. The data base could be reasonably well defined,
and if the country selected was of sufficiently wide-
spread intelligence interest a good cross section of
customers would be involved. The evolutionary path would
be clear cut, involving the addition of new geographic
area responsibilities to the new system. During the
period of evolution, both the new and existing systems
would have a clear understanding of their storage and
retrieval obligations, and the need for inter-system
communication and coordination would be minimized except
in the case of "international" requests.
Certain problems stemming from linguistic variables
and other processing problems peculiar to individual
countries would obviously not be addressed in this kind
of an initial system. However, none of the alternatives
SUMMARY DESCRIPTION
Recommendations
3.1.1 2.
Approved For Release 2000/%WRRETCIA-RDP78-03952A000100030001-9
Approved For Release 2000/051RCIA-RDP78-03952A000100030001-9
proposed except that which implements the complete
system all at once would incorporate all the problems
which would be reflected in the final system.
3.1.1.3. Country Selection
For initial implementation purposes it was felt
that the country selected should have the following
characteristics:
- A manageably low document volume, probably
somewhere between 50,000 and 100,000 items
per year.
- Considerable intelligence significance to
generate more than average consumer interest.
25X1A
- Documentation which gives a sampling of major
topic areas--i.e., political, scientific and
technical, military, economic.
- Available personnel familiar with the geographic
area and its documentation.
The USSR was eliminated from consideration immediately
since it would be dangerous to begin the operation with the
most strategic area of all, and because the volume of
receipts on the USSR would be extremely high. Most of
the West European, Near Eastern, Latin American and
African countries are either too weak in documentation
on major topics of general system interest, lack reporting
SU't IARY DLSCRIP'TION
Recommendations
3.1.1.3.
Approved For Release 2000/05ulk
kRA-RDP78-03952A000100030001-9
25X1A
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
Next 4 Page(s) In Document Exempt
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/0"FA-RDP78-03952A000100030001-9
25X1A
25X1A
needed skills identified. The test branch would be brought
to operational strength around the tenth month of Phase III*
and will move into an experimental indexing and training
phase. (See timetable in Section 3.2.1.1.) Full system
testing and operational simulation would begin in month 19
and the branch would operate thereafter in parallel
with current OCR operations until a decision could be
made on the readiness of the initial component to assume
operational responsibilities.
To assist in the implementation of,the test
component, and to provide logistical assistance and
operator advice to OCS, the CHIVE Support Staff (CSS)
should be enlarged during month 1 to a five-man team
consisting of four substantive analysts (probably
representing and one OCR support
programmer who should have a knowledge of MD or SR
maintained EAM authority files. Two additional OCR
support programmers should join the staff at a later
time. Personnel from GR and FDD could be committed on
an ad hoc basis to work on inclusion of graphic indexes
and on the problem of obtaining select machine-readable
inputs from FDD and JPRS.
*Note: Phase III milestones are given in months numbered
serially from receipt of approval to proceed--month 1.
SUMMARY DESCRIPTION
Organization
3.1.2.
25X1 B
-15-
Approved For Release 2000/05/0 r"' A
,-RDP78-03952A000100030001-9
fic,
Approved For Release 2000/(6RETCIA-RDP78-03952A000100030001-9
25X1A
25X1A
Both the CSS and could be
slotted against OCR's T/O and would be under OCR management.
The CSS, however, should be based in OCS/Development Division
where work on the major system design problems and on develop-
ment programming is assigned.
3.1.2.1. OCR Responsibilities
OCR should be responsible for guidance of the design
activity, for establishing system boundaries and objectives
and for review and acceptance of design decisions, The
DD/I CHIVE Project Officer should be the responsible
official for monitoring these functions. In addition,
OCR should assume responsibility for providing the per-
sonnel of both the CSS and Test Branch, and for
obtaining clearances for these people. Provision of
space for the - Branch as well as space in a secured
area in which to establish the Document Delivery System,
should be OCR's responsibilities. Personnel for the
Document Delivery System could presumably be recruited
from the present Library and SR document personnel
and from MD's film personnel. Contracting and purchase
of the Document System should also fall to OCR. Filming
and key punch services related to implementation of the
document system would also be OCR responsibilities.
SUMMARY DESCRIPTION
Organization
3.1.2.1.
25X1A
Approved For Release 2000/0S76 -- CIA-RDP78-03952A000100030001-9
SECOET
Approved For Release 2000/05/0?E RDP78-03952A000100030001-9
3.1.2.2. OCS Responsibilities
Full design responsibility should rest with OCS.
This includes system refinement and the execution and
management of computer program design and development
and system testing. In addition, OCS should be responsible
for space to house the CHIVE design team, the CSS, the
computer and attendant hardware. Contracting and costs
of the computer equipment and optical character reader
intended for system input should fall to OCS, as should
contracting, housing, and management of the design
and programming contract personnel. Contracts directly
concerned with CHIVE should be coordinated with OCR. OCS
should provide key punch and computer support to the
system as it develops on a demand basis for requests
and on a scheduled basis for input and maintenance.
3.1.2.3. Contractor Responsibilities
Under the guidance of OCS/DD, the contract personnel
should be responsible for participating in and advising
on all programming activities. This would include the
development programming required for the initial system,
and the programs required for extant OCR vocabulary
control files designated for conversion to CHIVE formats.
SUMMARY DESCRIPTION
Organization
3.1.2.30
Approved For Release 2000/05/0%EIRDP78-03952A000100030001-9
Approved For Release 2000/0968ETCIA-RDP78-03952A000100030001-9
The contractor would be expected to make recommendations
on program design, content and structuring, to propose
economical and feasible EAM file conversion routes, and to
continually bring to the system an awareness of new
programming or EDP developments which might benefit the
design activity. In addition, contract personnel should
participate in system review, and take part actively
in examination and evaluation of document delivery systems,
character readers and remote I/O devices, including
making recommendations on selection of equipment for
the system.
3.1.3. THE OPERATIONAL TESTING CONCEPT
The conclusion of the Phase III effort should
have modeled an initial CHIVE component which will be
ready for testing. The first testing stage, beginning
in month 19, will be a simulation of actual opera-
tional experience--a shakedown, in a sense, of the final
configuration derived from the experimentation of the
preceding months. Some further experimentation, adjust-
ment of programs, and refinement of procedures will
carry over into the initial testing stage. As far as
possible, however, the initial component will act like a
fully operational OCR component, i.e., it will receive,
SUMMARY DESCRIPTION
Testing Concept
3.1.3.
-18-
Approved For Release 2000/%W8&TCIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/OO~& A-RDP78-03952A000100030001-9
analyze, and index documents, forward index records to
the computer system, build files, and respond to queries.
The major difference between the Branch and a
"live" OCR component will be that it will not assume
actual operational responsibility to respond to queries,
since these will be borrowed, or captured at the time
of receipt, from the various OCR registers. It will
also use the old OCR files maintained by the registers.
The initial component must also prepare itself,
to contribute to such OCR summary files or scheduled
products as the
tribution only),
In addition, the initial system must specify procedures
25X1A
25X1 B
for preparation of finished intelligence reports and the
meeting of production deadlines for the scheduled products
cited above.
This testing will take place in parallel with
existing OCR operations for a span of six months or
longer, or until system and Agency management are satisfied
that the initial component is ready to accept opera-
tional responsibility. At that time operations 25X1A
in the OCR divisions would be closed out and their per-
25X1B sonnel and =and FIB-type working files transferred to
SUMMARY DESCRIPTION
Testing Concept
3.1.3.
Approved For Release 2000/05/0% 9.RDP78-03952A000100030001-9
Approved For Release 2000/(TCIA-RDP78-03952AO00100030001-9
25X1A the new component. The manpower balance of the
initial increment would not be increased, however, since
the seasoned personnel would begin work on incrementation
of the next geographic area as soon as the new personnel
were trained.
Early experimentation, beginning after month 10,
might involve an.attempt to cope with the entire-
document flow on a "no-backlog" basis. (It is understood
that backlogging might occur in an operation mode, but,
as no query would be going on during indexing, an
attempt to avoid backlogs should be made.) Plans would
call for capturing a one month take of - documents
from all sources. Following the initial training period
in concepts and procedure-s, the staff would be required
to index these documents in one working month. If they
were unable to do this, the indexing would be cut back
and the identical documents would be indexed again within
the identical time constraints. This process would be
repeated until a match was obtained between document
receipts and time available for indexing. At this point
a second month's document take would be introduced and the
entire process repeated to verify the successful indexing
procedure.
SUMMARY DESCRIPTION
Testing Concept
3.1.3.
25X1A
25X1A
Approved For Release 2000/g t CIA-RDP78-03952AO001 00030001-9
Approved For Release 2000/05/OSECtU-RDP78-03952A000100030001-9
25X1A
25X1 B
During this iterative indexing period, a basic set
of query problems would be developed and passed against
the indexed data following each iteration. Presumably
a trade-off point would be reached between reduction of
indexing and intelligibility of query response which
would aid in the determination of the final indexing
plan. Indexing levels for each iteration would be worked
out in close cooperation with CSS, who would recommend
exclusion standards and reflect requirements of their
various OCR components, and with OCS/DD where index
changes affected EDP requirements and specifications.
-Branch may assist in the development of such
requirements and standards.
CHIVE's present organizational thinking postulates
a group of geographic divisions staffed with area/topic-
specialized information/query analysts who would mark
elements of information in documents for later indexing
by a large, separate staff of indexing personnel in a
"Content Indexing Group" (see Vol. 5). Phase III should
permit experimentation with this, and other, possible
personnel configurations, but will not exclude the
possibility of limiting its ultimate approach to document
control, saving the installation "analysis"
problem for a later time.
SU11MARY DESCRIPTION
Testing Concepts
3,1.3.
-21-
Approved For Release 2000/05/00&i--RDP78-03952A000100030001-9
Approved For Release 2000 8T: CIA-RDP78-03952A000100030001-9
25X1 B
and organizational specialists from -
and FIB could be brought into the test prior to the start
of the general training period. These people would already
have been given ISC training. They would go through
the general CHIVE training program, but with particular
emphasis on the query and file building aspects. Once
experimental indexing was begun, however, their partici-
pation would be directed toward specification of named
object content for document indexing and on preparation
25X1 B
25X18 of special and installation information and
summary files.
By the time the indexing/retrieval balance was
tentatively established, the CTB would have reached full
operational strength. The document flow would be started
and routine indexing commence.
3.1.3.1. Phasing in of New Areas
Since the planned system would increment by geographic
25X1A
SUMMARY DESCRIPTI
Testing Concepts
3.13.1.
Approved For Release 2000MUT: CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/0#cUA-RDP78-03952A000100030001-9
should have begun by mid-1968, with full integration of
the Far East completed by the end of 1969. At this point,
assuming that system concepts are working out as planned,
the major priority area, the USSR, could be attacked.
Eastern Europe might be the third point of incrementation,
followed by the Near East and Africa, Western Europe, and
Latin America.
As new geographic areas are added, a repetition of
many of the implementation tasks performed for the initial
component will be required, e.g., vocabulary control file
building (although not on the scale required for the
initial component), document and data selection, file
identification, and training.
Since CHIVE's recommended organizational configura-
tion assumes geographically oriented information/query
analysts marking elements of information in documents
for later coding by a large, possibly separate, indexing
staff, physical separation of the expanding indexing and
analytical staffs and their files may become a system
consideration as soon as a planned Far Eastern increment
is added to the initial component.
Here, the idea of OCR adopting a "pre-CHIVE"
configuration which will ease this transition becomes
SUMMARY DESCRIPTION
Testing Concepts
301.31.
-23-
Approved For Release 2000/05/0?-RDP78-03952A000100030001-9
Approved For Release 2000/c68ETCIA-RDP78-03952A000100030001-9
25X1A
attractive. Following one possible line that OCR reorganiza-
tion might take, the CHIVE indexer staff, as it expands, could
become a component of an OCR "document processing center";
the information analyst staff could be based in an
information center--a component formed by merging present
OCR information functions under a single manager. As the
system grows, the personnel from the older document and
information components could move into the new activities
often without physical or managerial relocation. Trained
by co-workers, they would simply move from one processing
mode to another.
3.1.3.2. File Co-location
No co-location of - files with the initial
increment is planned for the operational test stage.
25X16 Access to major document, - and EAM files will
become a major problem only with the beginning of real
25X1A operational activity by the _ component. OCR document
and microimage files could be placed under a single manage-
ment, and every effort made to centralize their storage
and maintenance. Failing this, repeated file relocation
will be required as new areas are added to the developing
system, and a large amount of analyst activity will be
required to tap all of the older, unrelocatable files
SUMMARY DESCRIPTION
Testing Concepts
3.1.3x2?
-24-
Approved For Release 200012SLOR T CIA-RDP78-03952A000100030001-9
Eci~E
Approved For Release 2000/05/e c't A-RDP78-03952A000100030001-9
which may be involved in query activity. This matter is
elaborated somewhat in Section 3.2.3.1.
3.1.3.3. Exclusions
As indicated in Volume IV and in previous CHIVE
documentation, a number of OCR functions are excluded
from initial system attention. These include dissemin-
ation, acquisition, book cataloging and circulation,
translation services, EAM support to other offices of
the Agency, and the activities of OCR's liaison and
administrative staffs.
Each of these functions presents a separate set
of problems which will, in all likelihood, require
separate solutions. They have been excluded from initial
consideration because they are not central to the initial
implementation plan. While OCS resources are available
to OCR to attack any of these problems, concentration on
these functions by the CHIVE Design Staff at this time
could not but impede the course of initial system develop-
ment.
One area which is of concern to system development
will be explored during Phase III. This is the matter
of translation services. CHIVE is in close contact with
EDP developments on this front through the ALP project.
SUMMARY DESCRIPTION
Testing Concepts
3.1.3.30
-25-
Approved For Release 2000/05/%tRf -RDP78-03952A000100030001-9
Approved For Release 2000/(1MRETCIA-RDP78-03952A000100030001-9
However, the problem of obtaining selected translated
materials as machine-readable system inputs is of signifi-
cance as it impacts on the manpower requirements of the
developing system, and a plan for obtaining such inputs
should be ready for implementation prior to month 19. Some
further discussion of how this activity might be carried
out is available in Volume V.
Dissemination would probably be the next area to be
addressed, and would involve the development of an auto-
matic dissemination system, based on machine-stored user
interest profiles which would be kept constantly under
update and review. Automatic document dissemination
could probably be linked to the. inital header data index-
ing activity so that centralized dispatch of documents to
CHIVE components, to user analysts, and to the central
document store could be effected in one set of actions.,
3.1.3.4. Inadequacies
The initial system, while it will model the total
system objectives, will suffer from certain constraints
imposed by development costs and rates and because of the
existence of the old system. A heavy reliance on old
document and EAM files can be anticipated through the
first several years of operation, and this will require
SUMMARY DESCRIPTION
Testing Concepts
3.1.3.4.
Approved For Release 2000/90,
CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/UCRM-RDP78-03952A000100030001-9
considerable manual labor in preparing products for system
customers. The variety of output products involved in
searches of the inherited or old files and new CHIVE files
will also prevent the production of a truly integrated
system response to many requests. Also, a variety of
query routes will have to be a part of the new system in
order to provide access to older EAM-based files. Specialists
in old file maintenance and query will have to be involved
in query activities on many occasions.
Limitations in the EDP environment of the new system
will also create certain hardships. Since a general
purpose program for query and file structuring will not be
available by the time of initial implementation, files
will be structured on the basis of requirements known
to exist in the various OCR registers. When unanticipated
types of system outputs are required, special programs will
have to be created by the information analyst specifying
his requirements to programmers or "EDP file analysts."
The need for special programming actions, however, need not
mean that special input processing will be required.
3.1.3.5. Initial EDP System Products
The initial EDP system will provide a capability to
produce an entire document index record--header data plus
SUMMARY DESCRIPTION
Testing Concepts
3.1.3.5.
-27-
Approved For Release 2000/05/%~1 tf -RDP78-03952A000100030001-9
Approved For Release 2009 1pf : CIA-RDP78-03952A000100030001-9
information indexed to the document--or to select combina-
tions of data from the index record, suppressing unwanted
indexing and arranging the required information in
meaningful combinations at the direction of the system
user. The initial system will provide machine assistance
in the production of extensive, updated, and non-redundant
formatted summary files of information on people, install-
ations and commodities or topics. One great power in the
EDP system is its ability to collect and arrange independent
data elements from many index records to aid the analyst
in producing non-redundant synthesized reports which can
be stored, updated or printed at the operator's command.
The system will also provide a limited capacity for
remote visual review of index records, will output
printed indexes, listings and summary reports, and will
retrieve documents and make them available in either
hard copy or microform.
3.1.3.6. Problems
The path of initial implementation, extending
through the Phase IV live operational tests, is beset
with a number of major problems. It is possible at
this time to suggest tentative solutions to many of
these, but, in most cases, definitive answers can be
SUMMARY DESCRIPT 1OI:
Testing Concepts
3.1.3.6.
Approved For Release 2000T( 108 : CIA-RDP78-03952A000100030001-9
SECRET
Approved For Release 2000/05/OQR$TA-RDP78-03952A000100030001-9
derived only from testing and experimentation. Not the
least of these problems is how best to organize the central
reference activities to make maximum use of the advantages
which EDP support can provide. While CHIVE has suggested
a number of organizational configurations which appear
to permit meeting system objectives as CHIVE sees them,
primary responsibility for decisions in this area must
rest with OCR. As yet no organization configuration has
been hit upon which solves all system problems. OCR
must judge CHIVE proposals in the light of its own
experience and with a view toward providing continuity for
major OCR functions. Hopefully, OCR will limit the
number of organizational options by adopting a "pre-
CHIVE" configuration which does not disrupt major func-
tions and yet permits experimentation without inhibiting
the development of the sound elements of the design and
without drastic and repeated overhauling of major com-
ponent outlines.
Another critical problem area is refinement of the
indexing procedures which, are as yet, excessively com-
plex and difficult to apply (see Volume V). An
extensive training period is required to master the index-
ing technique, and even after this period the indexer
SUMMARY DESCRIPTION
Testing Concepts
3.1.3.6.
Approved For Release 2000/05/OV BETA-RDP78-03952A000100030001-9
Approved For Release 2000/eTCIA-RDP78-03952A000100030001-9
on occasion finds himself in doubt about how various
segments of the technique are to be applied to differing
problems. Reduction of the indexing complexity will be
a major Phase III activity.
Closely tied to indexing problems is the matter of
selectivity both of documents and of information within
documents. It is clear at this point that available
manpower cannot cope with the depth of indexing and the
range of document categories anticipated in early stages
of the design. The requirements of the various registers,
however, are such that in an integrated system no simple
exclusions by document category can be made without adverse
effects on one or another of OCR's requrements. No major
attack on this problem as yet has been undertaken, but it
must be carefully studied during Phase III in close
cooperation with the OCR divisions.
Still another major problem area is how the initial
CHIVE component is to interact with the the OCR divisions
as yet untouched by CHIVE. This involves not only such
problems as duplication of indexing with geographic areas
not yet included within CHIVE, or access to unconverted
files, but the involvement of CHIVE in such scheduled
outputs of OCR as the IPI and publications.
SUMMARY DESCRIPTION
Testing Concepts
3.1.3.6,
25X1 B
Approved For Release 2000/05188: CIA-RDP78-03952A000100030001-9
SECRET
Approved For Release 2000/05/0 c8?A-RDP78-03952A000100030001-9
System accessibility to uncleared requestors must
also be examined during the Phase III activity. This
matter will require careful study prior to any major
investment in information file building. Safeguards
to unauthorized disclosure of information must be con-
structed without placing limitations on the system's
ability to provide service to customers who are not SI-
cleared or who require information below the SI level.
The capacity of the system to go as high as SI/T/KH
or higher in answer to a query, but not to require the
customer to accept data at this classification level
whether he needs it or not, is a basic implementation
objective.
SUMMARY DESCRIPTION
Testing Conecpts
3.1.3.6.
-31-
Approved For Release 2000/05/0?E$-RDP78-03952A000100030001-9
Approved For Release 2000/TCIA-RDP78-03952AO00100030001-9
Chapter 3.2.
3.2.1. BASIC SCHEDULE
Given below is a timetable for passing major
implementation milestones during the Phase III effort:
3.2.1.1. Major Tasks (See chart following p. 357)
Date Complete
Month 1 Phase II work and Phase III plan review completed
EDP equipment ordered (IBM System/360 Model 60)
Secure specifications of operating System/360
(software)
Begin
25X1A
Enlarged CSS on board
- Test Branch nucleus formed
Begin building indexer aids
Begin specification of information and summary
files
Begin definition of input/output procedures
Program design underway
Review of indexing procedures begins
Month 2 Complete
Document Image System selected and ordered
Analysis of OCR/CHIVE indexing experiment
completed
Specifications for page reader completed; page
reader selected and ordered
Begin
Study of Map Library and Graphics Register
inputs begins
IMPLEMENTATION TASKS
Basic Schedule
3.2el~lo
-32-
Approved For Release 2000/05/08 : CIA-RDP78-03952AO00100030001-9
SECRET
Approved For Release 2000/05/iic1t A-RDP78-03952A000100030001-9
25X1A
Month 3 Complete
Forms design completed
Begin
Begin outlining dissemination procedures
Month 4 Complete
- Test Branch T/O identified
EDP equipment (IBM 360/Mod 30) on site
Review of indexing procedures completed
Program design completed
Begin
Development programming begins
Month 5 Begin
Program unit testing begins
Month 6 Complete
Support file design completed
Month 7 Complete
Detailed design procedures for Document Image
System completed
Month 8 Complete
Document site selected and prepared
Month 9 Complete
Document Image System installed
Job sheets written for
Support files ready
Indexing procedures established
System test plan completed
25X1A
Beg. in
System test plan development startec.a
IMPLEMENTATION TASKS
Basic Schedule
3.2.1.1.
-33-
Approved For Release 2000/05/0?5&-RDP78-03952A000100030001-9
Approved For Release 2000/OAF CIA-RDP78-03952A000100030001-9
25X1A
Month 10 Complete
Dissemination procedures completed
M brought to full operational strength
Begin
Indexer training begins
Montb1 11 Complete
Initial information files fully defined
Basic operating System/360 available (software)
Page reader delivered and checked out
Begin
Start experimental indexing
Month 12 Complete
360/Mod 60 equipment installed
Month 13 Complete
System operating procedures completed
Month 14 Complete
System test development completed
Month 15 Completed
Indexing tests completed
Program unit testing completed
Graphics and Map Library inputs ready
Begin
Definition of interaction with OCR (projects,,
inherited files)
Month 16 Complete
Secure full operating System/360 (software)
IMPLEMENTATION TASKS
Basic Schedule
3.2.1.1.
-34-
Approved For Release 2000/5&~T CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/& A-RDP78-03952A000100030001-9
Begin
Final program checkout begins
Document dissemination started
Graphic and map inputs begin
Month 18 Begin
Detail procedures for customer access to
Branch
Customer training begins
Month 19 Begin
Operational testing begins
3.2.1.2. Manning
The suggested manning requirements to accomplish
Phase III tasks are shown on the charts included below.
Manning requirements for an operational - com-
ponent would be substantially those of the - Test
Branch. In addition, four people (GS-4/6) will be required
to operate the Document Delivery System. Only the man-
power required for the design and implementation activities
is given here. Support manpower requirements, such as
key punch and computer operator manning, are summarized
in the Appendix 2A to Volume II of this report.
It must be recognized that manning levels, as assigned
to various tasks during initial planning, must be some-
what arbitrary and are subject to change as problem
areas are explored or as more realistic estimates of the
IMPLEMENTATION TASKS
Basic Schedule
3.2.1.2.
Approved For Release 2000/05/0gj -RDP78-03952A000100030001-9
25X1A
25X1A
25X1A
Approved For Release 2000/05/08: CIIA-RDP78-03952A000100030001-9
Figure 3-1 PERT Chart for CHIVE SYSTEM Development
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
C
C
C
System
System
System
T
est
Test
Test
De
sign
Plan
Development
C
C
Design
System
7
of Support
Operating
ie s
Procedures
C
Analysis of
Indexing
Experiment
wal of
N
.III
Ian
Support
File
S
Building
Selection and
Procurement
C
of the Document
C
S
Delivery & System
Indexing
Detailed Design
Procedures
Procedures for
Established Experimental
S
Document Delivery
Indexing
System Operation
S
Systen
C
C
Test
Forms
Document Delivery
Indexer
Design
System Site
Training
Preparation
C
Installation
C
of Document
Delivery System
Specifications
of
Page Reader
C
Secure
C C
Basic Operating
System/360
Secure
Specifications Program
of Operating Design
System 1360
C
C
Program
Program
Unit Test
Unit
on an I
BM
Installation
Testing
System/360 on
of
Site
Page Reader
C
S
C
Secure
Full Operating
I RM System/ 360
Installation
at IBM
System/360
Procurement
System136%
Model 60
S E C R E T
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
Figure 3-2
Task Key
1. Program Design
2. Development Progranuuing
3. EDP System Test
4. Support Programming
5. Indexing Development
6. Data Gathering
7. System Procedures
8. File Building
9. Exper. Indexing and
Training
10. System Integration
11. Operational Management
12. Design Management
2CS
2s_
_
CS-
; I
-3
Ci"_a
lt=
-_
? -C
-
.~--
_-
OCS
li_
-
Bbl-
i_c
-
S-?
_
___
__=
S_
Lt=9
39=a=
OCS
2i_
S-
11
14-1;
C T13
-2
-I
-i
S
1; I
BM-1
S
; I
EM-2
-
7
CS-
I
1
? CS
-4?
CT
24
311
_
2t-
--
CTB
i--
---
--
--
--
--
--
--
--
--
---
---
--
---
-
---
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
25X1A
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
Next 2 Page(s) In Document Exempt
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/OOFCIA-RDP78-03952A000100030001-9
amount of manpower required by certain tasks are reached.
It has been assumed that skilled OCR personnel can be
borrowed by CHIVE for short periods to assist in such
activities as file building. As a result, manning require-
ments, as listed for some major tasks, may prove to be
conservative.
3.2.2. PERSONNEL TASKS
In the following discussion it is assumed that Phase
III task initiation and completion will adhere to the
implementation time-table (or schedule of milestones)
listed in the preceding section of this report. Tasks
tend to break out into two closely related paths:
1) management tasks - those related to T/0 problems,
training, clearances, site location and preparation,
procurement, dissemination, and relocation of people and
their files; 2) substantive tasks - those involving
preparation of indexer aids, refinement of index procedures,
specification of CHIVE information and summary file formats,
EAM file conversion, and data selection.
Management tasks relating to larger aspects of
system implementation - equipment procurement, site
preparation, data gathering, and OCR liaison and support
IMPLEMENTATION TASKS
Personnel Tasks
3.2.2.
-41-
Approved For Release 2000/05/0?Ea[A-RDP78-03952A000100030001-9
Approved For Release 2000: CIA-RDP78-03952A000100030001-9
should be the responsibility of thp- enlarged CSS working with
25X1A OCS:and the - Test Branch management.
3.2.2.1. Substantive Personnel Tasks
The first implementation step could be to organize the
enlarged CSS. At the same time, the senior personnel who had
taken part in the Phase II Indexing Experiment, together with
a number of support personnel, and at least one EDP systems
analyst to advise on EDP possibilities, could be organized
25X1A into the nucleus of the future - Test Branch. The
25X1A personnel destined for the Test Branch could be
25X1A
organized on a task basis to work with CSS and the Design
Staff on the problems of preparing or improving the
basic indexer aids (organization dictionary, location
dictionary, etc.) needed for indexing in an operational
component.
Once the work on these basic tasks is well underway,
the indexer group--those responsible,for example, for
ISC revision and creation of new area rules--would devote
their main attention to working with OCS and CSS on
establishing and simplifying the indexing procedures,
setting up suggested indexing levels by document series,
and describing indexing levels for data.
Presumably these tasks will be experimental and will be
IMPLEMENTATION TASKS
Personnel Tasks
3.2.2.1.
Approved For Release 2000/lYvM/?18 : CIA-RDP78-03952A000100030001-9
SECRET
Approved For Release 2000/05/O0F A-RDP78-03952A000100030001-9
under revision and modification throughout the Phase III
activity. Experimental indexing should begin as soon
as the basic tools are constructed, the revised indexing
procedures defined, and interim EDP support made available.
25X1A Branch, CSS, and OCS design personnel should
coordinate work on other major system procedures tasks
such as forms design, specification of summary and infor-
mation file input/output formats, maintenance and query
routines, customer access problems, and training.
3.2.2.2. Management Tasks
25X1A Management tasks for the outset of the Branch
activity would involve, first of all, the setting of
individual task schedules and the monitoring of
progress. Continual liaison with CSS and OCR must be
maintained on both substantive and administrative
matters. In the latter area the early identification
of OCR personnel to be detailed to file building is
of prime importance, and it is assumed that persons
with special skills or file knowledge may be borrowed
from time to time from OCR to aid in the creation of
indexer aids or summary files.
Once the initial increment of knowledgeable
personnel is found and engaged in system tasks,
IMPLEMENTATION TASKS
Personnel Tasks
3.2.2.2.
25X1A
-43-
Approved For Release 2000/05/0BEcS -RDP78-03952A000100030001-9
Approved For Release 2000Td : CIA-RDP78-03952A000100030001-9
Branch management can proceed to attack the problem of
the numbers and skills of the people to be involved
in the indexing activity, identify when their partici-
pation is to begin, arrange for clearances, and begin
preparation of draft job sheets describing the duties
to be performed and skills required for the personnel'
25X1A of an operational - component. The proposed T/O
would be set during this period. Other matters requiring
detailed definition are the security procedures for the
operational testing phase, the establishment of dissem-
ination patterns, and the coordination of initial system
activities with the OCR registers, particularly in areas
where CHIVE activities and those of an OCR division
overlap.
Management tasks devolving on the CSS will include
advising on selection, procurement, site preparation,
and manpower problems relating to the Document Delivery
System, study of the problems of co-location of the
OCR document files, securing OCR support for data
collection and other design activities, and coordination
with OCS on EAM and EDP support requirements relating
25X1A to ~ activities.
IMPLEMENTATION TASKS
Personnel Tasks
3.2.2.2.
Approved For Release 2008 : CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/O?E: U-RDP78-03952A000100030001-9
25X1A 3.2.2.3. -Organization
The manning of the CHIVE Support Staff has been
described above. Manpower for the - Test Branch
can only be tentatively suggested at present since a
final determination will depend upon whether the test
component is established as an independent unit or is
attached, for administrative and logistical support,
to one of the extant OCR divisions. In either event,
25X1A
clericals would be required for header data indexing,
document control, filing and routine administrative
tasks. Specification of the exact manpower distribution
can be made only after final analysis of the results
of the OCR/CHIVE Indexing Experiment (see Volume V).
Since the primary aim of the - Branch is to
launch a parallel program which, after a period of
experimentation, could become the first operational
25X1A
25X1A
IMPLEMENTATION TASKS
Personnel Tasks
3.2.2.3.
-45-
Approved For Release 2000/05/0%EC FRDP78-03952A000100030001-9
Approved For Release 2O994.dO8 : CIA-RDP78-03952AO00100030001-9
25X1A
CHIVE increment, test indexing should being at the
earliest feasible date. It is assumed that the full
complement of -members would be on board by
month 10 so that ample time may be allowed for
training.
3.2.2.4. Operator and User Training
Operator training, in its early stages, would repeat
the training pattern established for the Phase II
Indexing Experiment: three weeks of training on the
revised ISC book, an OTR course
if required, and an intensive training period in CHIVE
indexing concepts and techniques. Rotational ISC
training of OCR -analysts selected for later
participation could take place well in advance of
25X1A their assignment to _ At the conclusion of this
training they could return to their home components
while awaiting assignment to the - component.
Training in a larger, practical sense would continue
throughout the early experimental indexing stages
25X1A Of
Branch activities. Training in the CHIVE
query language would be started as soon as indexing
personnel were well grounded in indexing procedures.
25X1A
25X1A
25X1A
IMPLEMENTATION TASKS
Personnel Tasks
3.2.2.4.
-46-
Approved For Release 2Ogg4 f 8 : CIA-RDP78-03952AO001 OOO3OOO1-9
Approved For Release 2000/05/CifCMA-RDP78-03952A000100030001-9
A full scale program of communication with potential
system users would be undertaken during the weeks just
prior to assuming operational responsibility. A
special series of briefings could be given potential
users, and a short brochure describing the Branch,
its data base, files, and functions, could be prepared
and distributed widely. In the final operational check-
out period, during the last months of testing, many of
the ultimate system users could be involved in testing
and evaluating the system through query of the reference
and information files.
3.2.3. LOGISTICS TASKS
Logistical problems involving the computer center
are the responsibility of the operations Division and
Administrative Staff of OCS and are not detailed in
this report except as noted in the milestone chart
given in Section 3.2.1.1.
25X1A
3.2.3.1. Space
As indicated above, logistical tasks related to
the - Test Branch and the Document Delivery System
would be the responsibility of OCR. The present 2G24 site
of the OCR/CHIVE Index Experiment provides a logical
IMPLEMENTATION TASKS
Logistics Tasks
3.2.3.1.
Approved For Release 2000/05/0%EtD1L RDP78-03952A000100030001-9
25X1A
Approved For Release 2006$O$ : CIA-RDP78-03952A000100030001-9
25X1A
point from which to begin Branch operations.
Additional secure work space must be located and
prepared by month 10, however, in order to house
the full _ complement and its files, and to provide
some requestor work space. If possible, final space
selection for this component should take into
consideration its potential to expand to a full Far
25X1A
Eastern division and its proximity to a central document
store. Total space.required by the initial operational
component will be approximately 2000 sq. feet - 1670 sq. ft.
for personnel and approximately 330 sq. ft. for files.*
Initially, the Document Delivery System should require
no more than 675 sq. ft. of space in a secure area.
Assuming co-location of the present Library Circulation
Branch (Documents) and the SR Document Library with the
Document.Delivery System, a total secured area of about
6,825 sq. ft. will be required.
The problem of the accessibility of the various
OCR document stores during the development of an all-source
25X1A system will arise as soon as Branch begins operational
*File space is required for indexer aids, some reference
works, and documents in various stages of processing. Final
file space requirements will depend, in part, on the operational
status of the microimage system.
IMPLEMENTATION TASKS
Logistics Tasks
3.2.3.1.
Approved For Release 20Q I9 9.8 : CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/09E%A RDP78-03952A000100030001-9
life. In preparation for this event, centralization of
the largest stores of documents and EAM files in OCR
could be considered as a way to provide a common document
repository out of which both the developing CHIVE system
and the areas not yet affected by CHIVE could work. If
the main library area on the first floor could be
revamped for this purpose,the SR documemt files, the
library hard copy and aperture card files and the new
document delivery system could all be located together.
The area on corridor 1H1, where the Librarian's office
is located, could also be studied as a possible site
for this common file. This area is also adjacent to the
25X16 = file rooms. Placing the first floor "H" corridor area
behind the SI barrier would permit free movement of system
personnel through this file space, and also allow access
to the nearby FIB files. Relocation of the large files
would permit central access to all major files by operators
of the old and new systems. It would also prevent repeated
relocation of file segments as new geographic areas are
added to the CHIVE initial component. Maintenance
and retrieval from the central file store could be
placed under a single management.
IMPLEMENTATION TASKS
Logistics Tasks
3.2.3.1.
-49-
Approved For Release 2000/05/0ltE(M RDP78-03952A000100030001-9
Approved For Release 206HILU8 : CIA-RDP78-03952A000100030001-9
EAM files and equipment might also be located in
one central area, possibly under a single management.
There would thus be a single point of address for EAM
support for non-CHIVE files or from inherited files,
and a single address point for document support whether
by system users or operators.
Area background files, special project manual
files, and smaller working files of interest to
particular groups of information analysts or particular
geographic areas would be housed with the information
analysts as system implementation progresses. CSS has
a full inventory of _ files and file sizes (see
Volume V). An exact determination of total space
requirements can easily be made as soon as the size
and configuration of the initial system is agreed upon.
3.2.3.2. Site Preparation
The site of the Document Delivery System must be
located and prepared by month 8 so that equipment
can be delivered and checked out. If centralization of
OCR document stores is undertaken, considerable reconstruc-
tion and preparation will be required in the area chosen
to house this facility. The same is true for EAM files
and machinery. CSS has obtained a plan of OCR floor
IMPLEMENTATION TASKS
Logistics Tasks
3.2.3.2.
25X1 B
Approved For Release 20821(?8 : CIA-RDP78-03952A000100030001-9
SECRET
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
space, including secure areas, and an inventory of
equipment and shelf and cabinet files, and could advise
and assist in site planning and preparation.
3.2.3.3. Auxiliary Equipment and Supplies
New equipment required for initial implementation
includes the Document Delivery System (basic card files,
cameras, printers, etc.) itself, and the optical page
reader designed for system inputs. Specially equipped
typewriters to prepare transcript sheets for acceptance
by the page reading devices will probably not be required.
The IBM "Elite" type appears to be the most common Agency
typewriter font, and the basic font for the page reader
and printers should probably be selected on the basis of
the availability of this font. Final selection of a type
font, however, and a possible subsequent requirement for
additional typewriters, will both be contingent on the
specifications established for the optical recognition
equipment. Other auxiliary equipment might include two
or more remote input/query stations (keyboard or CRT)
or, possibly, a long-distance Xerography link.
Whenever possible, equipment now owned or rented by
OCR or OCS will be used to support the initial system.
Examples of this would include Xerox 914 machines now
stationed in OCR, and OCR-owned film processors and
IMPLEMENTATION TASKS
Logistics Tasks
3.2.3.3.
Approved For Release 2000/05/OEEECEMA-RDP78-03952A000100030001-9
Approved For Release 200SIUM : CIA-RDP78-03952A000100030001-9
keypunch equipment which can be utilized to support the
Document Delivery System. A number of these are cited
in the Appendix 2.A. in Volume II. The complete
inventory of OCR equipment completed during the past
year will be used to determine whether needed equipment
is available or must be obtained.
The 360/Mod 60 computer system and attendant
hardware is an OCS requirement, with or without CHIVE,
and may therefore be considered available Agency
equipment. Initially, OCR will share this computer with
other users. However, the CHIVE machine-processing
load should probably be reviewed after the first two
years of operation to determine the best ultimate course
to follow in managing the computer element of the system.
A number of remote communication terminals (IBM
1050 keyboard/printer, 1015 CRT Inquiry Display Terminal,
and 2250 large CRT with light gun) are included in the
system/360 equipment package being considered by OCS for
installation in the Headquarters Building Computer
Center. These terminals will be available for experimental
purposes, and it is planned that one will be used to link
25X1 A the
with the Computer Center. While these terminals
will provide only a limited capability, as initially
IMPLEMENTATION TASKS
Logistics Tasks
3.2.3.3.
-52-
Approved For Release 20g2/TRT?8 : CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/WTA-RDP78-03952A000100030001-9
programmed, they can permit remote display of index
records or remote examination of query response while a
search is going on. Initial experimentation may help
in determining whether or not a total system requirement
for such remote display devices is likely to develop.
In a total system remote inquiry/output devices could
be used to link the geographic divisions with the Computer
Center or to connect either of these areas with such
remote requestor locations such as 00, NPIC, or OCI
Operations Center. Both ends of the remote terminal
will be in a secure area; security of the cable will
be explored as soon as a selection from the various
available remote devices has been made.
3.2.3.4. Move Schedules
Preparation of move schedules, beyond those needed
for relocation of the expanded M by Month 9 or 10, and
those required to prepare space for the Document
Delivery System, must await a decision on when the
initial system is ready to assume operational
responsibility. During a period of parallel operation,
colocation of any OCR files with a CHIVE component would
25X1A work a hardship on current OCR operations.
Colocation of plant and town
IMPLEMENTATION TASKS
Logistics Tasks
3.2.3.4.
25X1A
-53-
Approved For Release 2000/05/0?Et -RDP78-03952A000100030001-9
Approved For Release 2000/05M ElA-RDP78-03952A000100030001-9
25X1A
folders, or large document collections has not been
included in the basic task schedule. However, movement
of some of these files to a centralized location has been
recommended earlier in this volume. Designation of the
files to be relocated, a final estimate of their floor--
space requirements, identification of their containers,
preparation of new containers as needed, labeling of
files and their placement in the work area is all
comprised in this task. CSS should play an advisory
role in this task so that procedures for CHIVE access
to relocated files may be defined well in advance of
operational activity by the
Delivery of the Document Delivery System has been
scheduled tentatively for month 8. Moving of
files and clearing of floorspace to accommodate this
system and its personnel should be completed by late
October. Movement of files destined to be housed with
the Document Delivery System need not wait upon
installation and checkout of that system, but could be
scheduled for relocation as soon as the determined
25X1A site was cleared of other occupants and secured for
IMPLEMENTATION TASKS
Logistics Tasks
3.2.34.
-54-
Approved For Release 2000/@ t CIA-RDP78-03952A000100030001-9
Approved For Release 2000/05/O0CMX-RDP78-03952A000100030001-9
3.2.4. DATA SPECIFICATION TASKS
This group of tasks involves the selection of the
documents and document series to be indexed in the
initial system, the identification and preparation
of information and support files, the conversion of
OCR EAM files required by the system, and establishment
of selection criteria for the initial system indexers.
3.2.4.1. Document Selection
The document base and volume for
were determined prior to the Phase II CHIVE Indexing
Experiment, and a description of this is given in
Volume V of this report. Of the document categories
known to be destined for inclusion in the initial
system, those received through established Agency
distribution channels; i.e., the basic USIB reporting,
have been called "repository" documents. These will
automatically receive header indexing. A number of
document categories which are of interest to only
certain information analysts
and journal clippings, etc.) have been called "non-
repository" documents. These can be excluded from
25X1A
STATSPEC
25X1 B
IMPLEMENTATION TASKS
Data Specification
3.2.4.1.
-55-
Approved For Release 2000/05/0%E .RDP78-03952A000100030001-9
Approved For Release 2000 BT CIA-RDP78-03952A000100030001-9
STATSPEC
mandatory header indexing since these documents would
be selected and marked for indexing by the information
analysts. Only the "non-repository" documents selected
for content indexing and storage in the Document Delivery
System would require header indexing. Non-repository
documents which do not pass through regular dissemination
or are obtained by subscription, will pose a special
problem for the initial system. These documents will
have to be identified and copies directed to the ~
The general flow of IR's,finished intelligence, cables,
FDD and = published materials should pose no
dissemination or procurement problem. However, a
number of document series (FDD scientific abstract
items, SI teletypes) which could eventually be received
in machine-readable form will have to be provided for.
The system personnel will have to (a) accept the idea
that these materials will not be indexed until they
can be machine prepared or (b) allow for additional
manpower and indexing time to input these documents,
or (c) make provision in Phase III to prepare these
inputs for system acceptance in machine language by the
end of Month 18. While the latter of these alternatives
is the most desirable, it will require special study
and the allocation of development manpower to make
IMPLEMENTATION TASKS
Data Specification
3.2.4.1.
Approved For Release 2000/0MU8 : CIA-RDP78-03952A000100030001-9
SECRET
25X1A
Approved For Release 2000/05/&cr'f4-RDP78-03952A000100030001-9
these inputs ready by the start of _ testing
activities.
Disposition of TS documents and certain special
categories now controlled in SR will also require study
during the Phase III effort. A program for dealing with
these classes of documents will have to be worked out
before the start of Phase IV.
Even with information specialist marking of
documents for indexing, the document flow intended for
coding may prove too deep for available manpower. At
a minimum, present NODEX standards, and indexing levels
applied to various document series by the various
registers, must be examined to assist in determining
whether any major document series can be excluded from
indexing or, if indexed, to what level indexing must
be applied.
25X1A
3.2.4.2. Data Selection
This is one of the most difficult of the tasks to
be undertaken during Phase III. A selectivity standard
based on current operational needs can be developed if
the information specialists are asked to only mark items
to be indexed in their documents and the indexing and
trFLnscription of this material provided by a separate
IMPLEMENTATION TASKS
Data Specification
3.2.4.2.
-57-
Approved For Release 2000/05/0%i(g-RDP78-03952A000100030001-9
Approved For Release 200(5868 : CIA-RDP78-03952A000100030001-9
25X1 B
group. Indeed, if this system proved satisfactory to
the information retrieval personnel, it might never be
necessary to specify a set of general system selection
criteria for named object indexing except, perhaps, for
some series. The danger in this approach is that the
information specialist, with a free hand to mark items
for someone else to process, may use very little
discrimination in what he marks for indexer attention.
A thorough review of processing criteria by geographic
area for and installational indexing would
25X18 have to be made by the =and FIB management with
25X1 B
25X1A
have to be summarized and compared with CHIVE minimum
requirements and with the needs of the information
registers. Some series could be handled at the header
level only, still others would require only minimal
named-object indexing. Others (diplomatic lists, plant
selectivity criteria defined by the working personnel.
FIB appears to have made a good start in this direction
already, and _ could probably formalize its processing
requirements for as an initial attack on this
problem.
Levels of data specification by document series,
as currently applied in the various registers, would
IMPLEMENTATION TASKS
Data Specification
3.2.4.2.
Approved For Release MOAN= : CIA-RDP78-03952A000100030001-9
SECRET
Approved For Release 2000/05/6?c'TA-RDP78-03952A000100030001-9
brochures, finished reports) could be given
header indexing but could be retained in hard copy for
the topic or area to which they pertained.
25X1 B
3.2.4.3. Map Library and Graphics Register Inputs
During the Phase II activity, informal agreements
were worked out between OCS, ORR/Map Library, and OCR/
Graphics Register to provide for the development of
compatible index standards which would enable either
the CHIVE analyst, or the graphics or map analysts, to
search the central computer system for index records to
maps, photos and films. A major Phase III task will
be the coordination of index development, the establish-
ment of index levels, and arriving at index compatibility
with Map Library and GR so that computer inputs can be
ready for acceptance by October 1966. Development
Division personnel will be assigned to this task and
persons familiar with the GR and Map Library systems
and requirements will, presumably, be loaned to the
development activity by the map and graphics components.
It is not planned that the map and graphics document
files be made a part of a central document delivery system.
The physical proximity of these files to the central system
is, however, desirable. GR is already located in an area
IMPLEMENTATION TASKS
Data Specification
3.2.4.3.
Approved For Release 2000/05/e-RDP78-03952A000100030001-9
Approved For Release 20008$798T: CIA-RDP78-03952A000100030001-9
25X1A
adjacent to the first floor location which has been
suggested as a possible site of a central document store.
Conceivably, the small Map Library Headquarters collection
could be located nearby.
3.2.4.4. Vocabulary Control Files
The construction of indexer tools and vocabulary
control files would be among the very first implementation
tasks. These are discussed below.
3.2.4.4.1. Area Dictionary
The area rule problems as applied during the CHIVE
Indexing Experiment will be examined. Intelligence
Subject Code area rules would be modified and a special
area indexing procedure generated especially for CHIVE's
indexing problems. A Master Location Dictionary for
based on the ISC dictionary and NIS Gazetteer
would be developed.
3.2.4.4.2. Intelligence Subject Code
Suggested revisions to the ISC book would be
collected and incorporated into a CHIVE version of the
ISC. In addition, the ISC would be expanded to include
those subject/codes peculiar to Comint materials.
IMPLEMENTATION TASKS
Data Specification
3.2.4.4.2.
-60-
Approved For Release 2000/05/08 : CIA-RDP78-03952A000100030001-9
SECRET
Approved For Release 2000/05/o9cTA-RDP78-03952A000100030001-9
3.2.4.4.3. Occupation and Function Code
Special occupation and organization function listings
for personality and installation indexing will be developed.
The ISC in its present form would not be used for this
purpose.
3.2.4.4.4. Organization Identifier File
25X1A A - Organization Identifier File will be
developed consolidating the organizational information
available in SR, FIB, and _ cross-referenced and
non-redundant, and reformatted for computer. This
would probably be the most sizeable of the file
building tasks and could be expected to extend through
the first six months of the Phase III effort. Specialists
25X1A on -organizations would be borrowed from appropriate
OCR divisions to advise during this activity. One of
these people would probably be identified as the
organization dictionary editor of the future
Branch. At least two subtasks would be involved in this
effort: (a) establishing a liaison procedure with SR,
25X1 B
25X1A
25X1B FIB and Mto pick up newly identified - organizations 25X1A
and to record changes in established ones, and (b) analysis
of the political organization problem peculiar to 25X18
category files to determine whether these files are
IMPLEMENTATION TASKS
Data Specification
3.2.4.4.4.
-61--
Approved For Release 2000/05/0%E RDP78-03952A000100030001-9
Approved For Release 2000Q': CIA-RDP78-03952A000100030001-9
25X1 B
suitable for incorporation in the basic organization
dictionary or whether they are more efficiently maintained
in manual form.
3.2.4.4.5. File
The possibility of reformatting the
li
ti
25X18
s
ng
for use both as an index to
and as
25X18
a basic reference aid must be studied.
The entire card
deck for each - entry could be rearranged so that
25X113
all elements of information coded to the
ld
25X18
cou
be listed after the name entry.
3.2.4.4.6. Header Data Dictionary
Creation of a more detailed Header Data Dictionary
is required to establish standardized control terminology
for recording and displaying data contained in the
bibliographic or header portion of documents.
3.2.4.5. File Conversion
Present planning calls only for initial conversion
of those EAM-based OCR files which are required by CHIVE
to provide vocabulary control aids for the system. These
are listed in the preceding section. Conversion of the
EAM card base for the Intellofax, SR, FIB, or
document or presents a number of severe
IMPLEMENTATION TASKS
Data Specification
3.2.4.5.
25X1 B
25X1 B
Approved For Release 200W=08 : CIA-RDP78-03952A000100030001-9
SECRET
CONFIDENTIAL
Approved For Release 2000/05/%@ RDP78-03952A000100030001-9
problems and will require careful examination of the
most feasible conversion routes for the most active
files. This matter is being studied and some further
discussion of the problem and its approaches is given
in Volume V.
25X1A
3.2.4.6. Information Files
EDP-based summary information files would be
created by the information specialists assigned to the
Test Branch. These files would be begun on the
basis of an established requirement for the maintenance
25X18 of such files in= FIB, or any other OCR component.
25X1A An all-source M would face at once the problems of
25X18 all-source and installation summary files.
These would include organizational and target installation
listings, some of which may be identical with or derived
from, identifier lists. Construction of non-redundant,
25X18 summary
could be begun as soon as experimental indexing was
started. Construction of these files would make use,
wherever possible, of work already done in _ and FIB. 25X18
Other summary files, such as (a contribu- 25X18
tion only), leader appearances, travel listings, may
also be required. In addition the initial system must
IMPLEMENTATION TASKS
Data Specification
3.2.4.5.
Approvedw;\I~jIs
05/08 9T4-RDP78-03952A000100030001-9
wl--v I __,_
Approved For Release 200 C
M 000100030001-9
be prepared to simulate or specify procedures for the
preparation of finished intelligence reports and meeting
production deadlines for the products cited above.
The system should also experiment with the preparation
of summary information files from unsynthesized data
in the index records, and should attempt to gain experience
in preparation of these outputs. Timeliness, both of
input and output, will be a significant factor in judging
the success of the unsynthesized file approach in filling
requirements for such closely scheduled system products
as leader appearance listings.
Information files will be constructed at the lowest
possible classification level. Here, the problem of
all-source inputs will be a significant factor, since
the files must be ready to serve a wide range of system
customers who will want products at the all-source level
and below.
3.2.5. PROCEDURES TASKS
The tasks described below are related to the problems
of how the system is to operate once it enters an
experimental indexing mode and thereafter. A number of
activities touched on in preceding sections, experimental
IMPLEMENTATION TASKS
Procedures Tasks
3.2.5.
cOivFIDENTIAL
Approved For Release 2 1h8 : CIA-R - 9 2A000100030001-9
Approved For Release 2000/05/ -RD 9I4'kEYWl&30001-9
use of remote input/inquiry devices, for example, might
also be considered procedures tasks.
3.2.5.1. Indexing/Transcription Procedures
An indexing guide was developed for the OCR/CHIVE
Indexing Experiment (see Appendix 5.C). For an operational
system the manual needs considerable expansion to cover:
- Indexing procedures unique to - as an area. 25X1A
- Special indexing rules which should be developed
out of the experience gained in the experiment.
Document selection procedures unique to as
an area.
Header data and content data transcription sheets
have also been developed, but experimental use indicates
25X1A
some need for revision. A single header data transcription
sheet was developed for all sources for use in the experiment.
It has been recommended that during Phase III several forms be
developed to accommodate various series and that/the use
of continuous roll forms be investigated as an aid to
expedite typing. Both the header data form and the
content data form which is used to retype the content
data from the indexer's handwritten transcription form
will have to be developed with the requirements for
character reader input in mind.
IMPLEMENTATION TASKS
Procedures Tasks
3.2.5.1.
PU41 -65-
Approved Fo I U16 -RDP78-03952A0"00100030001-9
Approved For Release 200
CI
3.2.5.2. Other Forms Design Requirements
25X1 B
3.2.5.2.1. Summary File Forms
and installation summary files may require
special input/output forms. The content requirements of
these files can be established by the appropriate OCR
divisions. However, considerable study of formatting
problems may be required.
3.2.5.2.2. Transaction Processing
A general form on which to record-information about
processing activity will also be required. A version of
the input transcription form can be passed through the
page reader for computer action; however, additional data
about the query transaction must be captured. Possibly
some variation of the data collection forms used in the
OCR/CHIVE Indexing Experiment can be employed initially
and revised as new requirements come to light.
3.2.5.2.3. Document Requests
Another form will be required for communication
with the Document Delivery System when large numbers of
documents or files are requested and it would be
burdensome to transmit document numbers by telephone.
IMPLEMENTATION TASKS
Procedures Tasks
3.2.5.2.3.
Approved For Release
W
5/08 : CIA-RD0
k&A30001-9
Approved For Release 2000/05/TA-RD-0
3.2.5.2.4. Error Correction
An error correction form may also be required so
that items in index records rejected as incorrect by
the computer can be corrected by the indexer with a
minimum of reindexing or retyping.
3.2.5.2.5. Index Instruction
A form may be needed through which the information
analyst can specify his
to the indexing staff.
input or maintenance instructions
In many
documents themselves may not be
instances,
analyst/indexer communications is,
territory, and detailed procedures
30001-9
as yet,
marking the
This area of
unexplored
specification
area will become a major system requirement.
3.2.5.3. Operator/User Communications
No novel operator/user communication techniques are
planned for the initial system. Communication with
system customers will take place by mail, by personal
customer visit, by black and gray phone communication,
and through use of the Headquarters Building tube
systems. Agency mail and courier service will be used
25X1A between ~ and remote locations.
The major task in the communications area will be
25X1A to define and simplify ~ contact with customers who
IMPLEMENTATION TASKS
Procedures Tasks
3.2.5.3.
Approved F ~JpCQ~jA
or
Approved For Release 200%jY0
C~6eNlRDEfl1qA(00100030001-9
are not cleared at the all-source level. A special work
area will have to be established for customers who
cannot visit the Branch itself, and _ analysts will
have to meet with customers in this area to discuss
requests, make files available, and explain query products.
Establishing feedback procedures to record and
describe customer reactions to the developing system
will be another essential system task. Feedback can be
captured verbally, by CHIVE analyst comment on his
requestor's reactions, and als by followup forms attached
to distributed query products. Circulation of a
questionnaire to known system customers at the conclusion
of the initial operational test period will also be
considered.
3.2.5.4. Maintenance/Query Procedures
An extensive set of file maintenance and query
capabilities has been designed for the initial EDP
system. Appendix 7.A. describes the query language
by which the information analyst will request the
performance of these functions by the EDP system. The
optical page reader will accept the user coding sheets
for maintenance and query as well as for index record
input. Experimental usage of the query language will
IMPLEMENTATION TASKS
Procedures Tasks
3.2.5.4.
Approved For Release 200
CIA8I9521100030001-9
Approved For Release 2000/0,5/R'
3P0;00100030001-9
be required to determine whether improvements can be
made in its ease of use and capabilities. Possible
changes might involve altering the syntax of the
language, the coding forms, or, most importantly, the
functions performed by the language.
The structuring of new files, new output reports,
and new input formats will be carried out initially by
the programming staff by way of the assembly/compiler
languages of the system/360. As additional operating
experience is gained, higher level, problem-oriented
languages will be developed for the structuring of files,
inputs, and reports.
3.2.5.5. Management Reporting Procedures
Throughout the implementation phase it is assumed
that both periodic and ad hoc reports to OCR and OCS
management will be made through CSS, ~ and Development
25X1A
Division personnel. Written reports will be produced
as various milestones are passed or as management
approval for design decisions is required. = will
25X1A
report routine management information on work accomplished,
personnel actions, etc., in the manner of a line OCR
component. As soon as experimental indexing has begun,
logical products of this activity, e.g., volume figures,
IMPLEMENTATION TASKS
Procedures Tasks
3.2.5.5.
Approved For Release 2000/05/0,
IA-RDP78-03952A000100030001-9
r
Approved For Release 2000
''"'CIA-RDF7--03952A000100030001-9
indexer rates, key punch times, and transaction speeds
can be collected. The initial component will have the
capacity to produce manually any of the data now reported
routinely to OCR management by its operating components,
and to generate automatically information on machine
activity of interest to OCS management and to design
personnel.
3.2.5.6. Procedures Documentation
Input, maintenance, and output procedures for the
CHIVE document index files will be summarized and
described throughout the initial implementation period.
Indexing manuals will be created and kept under
revision. In addition, a query guide, which will also
include access and maintenance routines for summary
files, vocabulary control files, and other identifier
listings, will have to be drawn up. Capabilities of
and access to external OCR files of interest to the
CHIVE system will also be described, as will procedures
for interaction or feed-in to on-going OCR programs
which may be influenced by CHIVE developments.
IMPLEMENTATION TASKS
Procedures Tasks
3.2.5.6.
Approved For Release 20
: CIA-RDP78-03952A000100030001-9
Approved For Release 2000/o5' TA-RDP
3.2.6. PROGRAMMING TASKS
"Ar 0001-9
3.2.6.1. Program Design
The CHIVE System Design has specified an overall
system flow, system loads (input, processing, and
output), system priorities, system queues, etc. This
information will be used in Phase III for program design.
The program designers will specify the logical
program units that must be written. These program
specifications will delineate all of the relevant
parameters associated with each program unit. These
parameters include;
input data formats, devices and volumes;
processing requirements for file building,
maintenance and query;
output devices, volumes, and formats;
storage allocations; and
interface between program units.
3.2.6.2. Development Programming
This task involves the translation of the CHIVE
Program Design into an operational program. The first
step in accomplishing this task is to prepare flow
charts of all EDP functions. These flow charts will be
reviewed by project management to insure that the
IMPLEMENTATION TASKS
Programming Tasks
3.2.6.2.
G F1DENT!A 1_
Approved For Release 2000/05/0 RDP78-03952A000100030001-9
CONFIDENTIAL
Approved For Release 2000, b1 CIA-R 78-03952A000100030001-9
system design has been properly interpreted. The next
step is coding into an appropriate language which can be
read by a computer processor. Testing will then be
performed on program units which will later be assimilated
into program modules for further testing. While a program
unit performs limited functions such as sorting, merging,
and report structuring, these units will be combined into
program modules (such as output processing, file searching,
and file maintenance) for further testing.
3.2.6.3. System Testing
During the program development task a set of
objective criteria will be established to be used as a
final test of the programming system. This test will
demonstrate to management the capabilities of the CHIVE
Operational program.
The complete program system must finally be tested for
proper interaction among the program modules and to insure
that the program units and modules properly interface with
the operating system. Input conditions are established
and data used to simulate operational conditions. This
testing will uncover flaws in the program design, flow
charting, and coding efforts. These errors will be
corrected under this task.
IMPLEMENTATION TASKS
Programming Tasks
3.2.6.3.
COW
Approved For Release 20GG108ft : CIA-RDP7$=0-3952A000100030001-9
DEN /
Approved For Release 2000/1014-RDP78-03952A0001 60030001-9
The final phase of testing will consist of using the
previously developed set of criteria for system testing in
order to secure management approval of the systems operation.
3.2.6.4. Support Programming
The development of an operational EDP system requires
a variety of computer programs other than the operational
program itself. Support programming on CHIVE has been
carried on for several months in support of the document
indexing experiment. This effort will be continued and
will expand to include the programs necessary to convert
existing files into formats that can be efficiently
processed by a computer as well as the development of
support files such as authority lists. The capabilities
of existing programs will be exploited before special
purpose programs are written since efficiency of program
operation is not a significant factor in this task.
3.2.6.5. IBM System 360 Software
Major decisions have yet to be reached on the degree
to which the more sophisticated IBM programming packages
can be exploited in the development of the CHIVE EDP
system. The target dates for availability of these
packages are such that severe delays could be encountered
in providing the initial capability if full reliance on
IMPLEMENTATION TASKS
Programming Tasks
3.2.6.5.
ApprovedFor` l elea$e 2000/05/
RDP78-03952A000100030001-9
Approved For Release 200
CIA-FJ
IBM software is planned. While reliance on these
00030001-9
programs is advisable because of their power and the
need for Agency programming standards, the need to get
on with the job may require that some initial work be
undertaken without these programs being available.
Approved For Release 20
IMPLEMENTATION TASKS
Programming Tasks
3.2.6.5.
03952A000100030001-9
Approved For Release 2000/05/08 :CIA-RDP78 (DEjjf?0
01-9
SECRET'
Approved For Release 2000/05/08: CIA-RDP78-~UWWJj O01-9