STRENGTHENING THE INFORMATION PROCESSING STRUCTURE OF THE AGENCY

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78-04723A000400010004-8
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 6
• Document Creation Date: December 9, 2016
• Document Release Date: July 19, 2001
• Sequence Number: 4
• Publication Date: September 5, 1969
• Content Type: MF

Approved For Re1se 2001/08/08*ft1AADP78-04723AM400010004-8
,
j.
S SEP 19q.
MEMORANDUM FOR: Chief, Information Processing Staff,
OPPB
SUBJECT : Strengthening the Information Processing
Structure of the Agency
1. We have been asked to submit ideas about WHAT:
a. IPC responsibilities should be within the Director-
ate;
b. the responsibilities of the IPC group should be as
a group;
c. the responsibilities of the Information Processing
Staff, OPPB should be; and
d. the responsibilities of the Executive Director-
Comptroller should be.
Ideas about how_ to strengthen the overall IPC structure and performance
are also solicited, but will be specifically addressed at a later time.
2. I believe our approach to the management of information pro-
cquoEug Activities in the Agency has been backward. The concern has been
wLth proliferating computer acquisitions and burgeoning costs. Conse-
quently management attention has focussed on hardware. We have backed
a few steps away from insisting on centralizing all computing capability
ol! the Agency under single management, but we still attempt to lump
tether into one program Category for planning and budgeting purposes
all activities which use computers. The Information Processing and EN'?
pLoitation Category is a hodgepodge. It is a strange mix of automated
and manual activities; of functional units and organizational components.
The logic of what it contains is equally as obscure as the logic sup-
$jr ting the omissions.
3. Computer systems are versatile. They will do a wide variety
oF things from mundane clerical operations to exotic and sophisticated
sOicatlfte calculations. Lsseatially they are collections of hardware
arid SQ ware configured in a way to permit them to be uaed to solve
so4,10 operational problem or serve some management objective. There in
;Approved For Release 2001/08/08 : CIA-RDP78-04723A000400010004-8
SECRET
Approved ForRVOse 2001/08/0e:wo1A-RDP78-04723A064400010004-8
no reason for them to exist in an environment unto themselves. There
is nothing about them which gives any element of commonality to the
things for which they are used. Because computers instead of an acre
of back to back analysts are used to retrieve information, for example,
is not a reason to divert the responsibility for the management and
productivity of information retrieval systems to the manager of the
computer center, or to dilute that responsibility by sharing it. The
manager who has the responsibility for the function has the responsi-
bility for the system used to perform that function whether it requires
slide rules, calculating machines, punched cards, or computers. He is
responsible for what goes into the system, and what comes out. This
is not to say, however, that he must have command jurisdiction over the
computer center. It simply says that he cannot relinquish his responsi-
bility for the function to the manager of the computer center no matter
where that center may be located organizationally.
4. The judgment of whether or not to use a computer to solve a
problem is in the first instance the responsibility of the manager who
has the problem to be solved. He must evaluate the alternatives and decide
whether a manual, a computer, or some method in between represents the
most effective solution operationally and economically. Because computers
are versatile and available simple as well as highly complex problems
gravitate to them. Computer equipment is used too often simply because
it's there with time available,'or because the working level operator who
has a problem also has a friend who is a computer programmer. Together
they work out solutions to fragmentary problems in a fragmentary way with
the result that valuable human as well as machine resources become com-
mitted without the exercise of management judgment about the cost, the
resource allocation, the effectiveness or even the utility of the solu-
tion and its relationship to larger parts of the system. Once an appli-
cation is on a computer it stays there regardless of how it got there or
what its utility may have been when it was first installed, and regard-
less of whether it continues to serve a valuable purpose or has outlived
it usefulness, Managers and operators almost never give their atten-
tion to a critical review of on-going applications for the purpose of
discontinuing those which have been overtaken by changing requirements.
Iteis for these and other similar reasons that the number of computers
groiwn and the associated costs mushroom, but centralized management
attention Co computing equipment is not the way to get at problems of
thtil sort, Segregation and separate identification of functions which
uee coeouting equipment and their merger into one information proces-
sing cauldron is not the way to get at it either. To lump imagery
anaymis together with, printiug services into one program category
simply beeeuse they both use computing equipment or even because they
are both, in a manner of speaking, processors of information does not
meke them a management homos;eeeity. The value of computer systems must
beimeasured in terms of the effectiveness of the operational or manage-
ment systems they serve. It cannot be measured in terms of numbers of
1
1 2
Approved For Release 2001/08/08 : CIA-RDP78-04723A000400010004-8
SECET
Approved For ReAseise 2001/08/08 : CIA-RDP78-04723A0441400010004-8
computers and their costs alone. There is no such thing as an ADP pro-
gram in the Agency, and we should Stop trying to manage it as though there
were such a program. Instead, we have a wide variety of operational and
management programs which use ADP and we should concentrate on their ef-
fective management of it.
5. The problem of controlling the numbers of computers is essen-
tially an inventory problem. The problem of controlling their costs is
essentially an accounting problem. These are relatively simple problems
deserving only of the simplest possible solutions. Perhaps we should
be seeking a way to adapt the FAN account and sub-object class mechan-
isms to accomplish these purposes rather than attempt to extract the
computer usages from all of the operational and management applications
they have in the Agency. In any case, accountability is essentially a
clerical operation and it should be treated clerically.
6. If we are really going to manage and control computing equip-
ment in the Agency, its costs as well as its growth in numbers, we have
to find some way of getting at the source of the problem. We have to
find some way of getting at the source of the problem. We have to find
some way of bringing management judgment to bear on the decision to use
the computer in the first place. There is an over riding concern that
the use of computers be limited to problems which justify the expendi-
ture of computing power and the scarce human resources necessary to
operate and maintain it. I repeat: the judgment of whether or not to
use a computer to solve a problem is the responsibility of the manager
who has the problem to be solved. He must evaluate the alternatives and
doeil.de the use of a computer represents the best solution. We have to
find some way to ensure that this responsibility is conscientiously,
prudently, and judiciously fulfilled.
7. BOB Circular A-71 dated March 6, 1965 assigns to the heads
of agoncies responsibility for administration and management of ADP
activities. Key words in that document are: mission accomplishment;
utlitty; increased effectiveaess; greatest return; higher productivity;
ineeerated systems; systems management; program accomplishment; use;
jo.r.nt use; impact on the work force; and planning. If we can find a
way to bring responsible management judgment to bear on these factors
611:1ng the program plannine,. stages bnfore problem solutions become com-
milaed to computers we will have solved the problem of proliferation.
We ooe not interested in trying to manage equipment in a vacuum. We
arki interested in mission accomplishment and the best management we can
bring to bear to achieve that objective regardless of the tools used iu
thrA'process. We need a mechanism which will determine when computing
cqufpmenc will best serve the purpose; when it will not; what we should
do,rbon it will; and what we shouLd do when it won't.
8. Inherent in the management and control over the uses of
computing equipment is the need to plan for its use as far in advance
as Tensible. This is not, and we should not attempt to make it, a
Approved For Release 2001/08/08 : CIA-RDP78-04723A000400010004-8
r... -.r..
? Approved For Rase 2001/08/08 A+AtrtrAD78-04723AD49400010004-8
separate exercises. Plans to use computers can only be made when func-
tional and operational plans have reached the point in their development
where the manager begins to identify how he proposes to construct and
conduct an activity. While it may not be possible to identify potential
computer impact in the initial formulatipn, the conceptual stages of a
plan, it should be possible to recognize them very early in the develop-
ment of operational plans and programs. In any case, it is vital that
potential ADP applications be identified as early in the planning pro-
cess as practicable. This suggests the need for a very close relation-
ship between the program planners and the information processing
coordinators, particularly if we acknowledge that it isn't always reason-
able to expect the operating manager to identify them because, he may not
be attuned to computer potential and capability.
9. While it is important for the Information Processing Coordin-
ators to have a close working relationship with the program planners,
we dare not rely entirely on the PPD cycle and mechanism for the iden-
tification of potential ADP applications. Isolating a problem, getting
it defined, determining the desirability and feasibility of a computer ?
solution, and designing a system to satisfy the requirement is a dif-
fiCult and time consuming process which will very often cause the
. development of ADP applications to get out of phase with the PPB cycle.
Accurate resource forecasting can only be done after systems have been
.designed which may further distort the time phasing relationship with
the PPD mechanism. This suggests that we cannot rely entirely on the ?
PF mechanism for the early surfacing of potential ADP applications ?
but will need a separate mechanism closely interrelated with it and
overlapping it at certain points LI the cycle. ?
10. Early identification of potential ADP applications is not
. .
only'necassary to permit reasonable forecasting of computer resources,
Also to permit logistics, communications and security planning. The
information processing mechanism as well as the PPD system must provide
for the early identification of plans and requirements to all of these
-components so their planning and programming can include provision for
:ono:to, power and air conditioning supplements, Screen rooms, building .
'alterations and special construction, establishing the communications
: network and all of the other supportin actions necessary to put a com-
puter application into practical operation. Early information is vital
'tothese components, but it should be regarded as equally vital to the
-component proposing the installation in order to ensure that every pre-
caution is taken to have the facilities ready when the application is
the to becomeoperational.
11. In summary, we should abandon the Information Processing
au4 Mcploitation category because it doesn't make sense; we should
abandon the attempt to pull together into one lump all systems which
u*1 computing equipment ati a mans of managing and controlling .prolif-
oration of computers and costs because that doesn't make sense either;
wo should seek a simple mechanical method for maintaining an inventory
a44 accounting system; we should direct our attention to a system for
Approved For Release 2001/08/08: CIA-RDP78-04723A000400010004-8
Approved FortVease 2001/08/08: CIA-RDP78-04723741160400010004-8
find a suitable way to relate to program planners to assure the earliest
possible identification of programs which have potential ADP impact; we
should find a way to assist managers to make prudent judgments about
adopting computer methods; and we should find a way to monitor the exer-
cise of management judgments and ensure their judiciousness without
looking over the manager's shoulder.
12. .If this is a reasonable description of WHAT the IPC mechanism
should be concerned with, we can begin to address the question of what
the responsibilities of the various echelons should be.
13. While some of the activities of the Office of Communications
and all of the computer support rendered the Support Directorate by the
Office of Computer Services are included in the Information Processing
and Exploitation Category of the program, none of the Support Information
Processing Systems (SIPS) are included. The Support Directorate is the
only Directorate which does not have its own general purpose computing
equipment. With the exception of special purpose gear in the Office of
Communications, we are only indirectly concerned with hardware manage-
ment. Our attention has, and will continue to focus on managing and
controlling the use of computing equipment at the source, that is with
the component or manager who has a problem to be solved.
14. In October 1963 the DD/S directed a memo to each of the
Support offices saying that centralized Directorate cognizance and broad
management control over all information processing activities in the
.Support Directorate is imperative. - (Information processing was defined
to include all automatic and electronic data processing and related
manual or machine activities except communications automatic switching.)
-The memo also said that all proposals for the development of new or sub-
tantial modification of on-going information processing systems bust be
? subjected to an overall review and adjudication of priorities from the
Directorate point of view. This requirement for review applies whether
the proposal is for the conduct of a feasibility study, the acquisition.
of equipment, the modification of existing systems or any other aspect
of System conception or change. All proposals and requirements for
information processing service or assistance are to be taken up with the
DD/S.-IPC before assistance is sought from other Agency facilities or
contractors. They are to be prepared for consideration in accordance.
with. criteria and standards to be established by the DD/S Information,
Processing Coordinator. The IPC is to assist in problem definition, in
determining what facilities can best meet specific requirements,. and in
thu. development of requests for project approval. He will recommend to
tho DNS priorities for project development. The DD/S IPC was also
ma4e responsible for keeping the DD/S informed of all Information pro-
cessing activities of interest to tlAs Support Directorate in the Agency
as well as the intelligence community.
15. In addition, last Fall each of the offices was asked to
Approved For Release 2001/08/08 : CIA-RDP78-04723A000400010004-8
CUT
: Approved For RtipOe 2001/08/08:CiA-RDP78-04723AQQA400010004-8
identify separately all plans in their program submissions which they
expected would have ADP implications. Each of the programs was also
reviewed by the Chief, Plans Staff in the Office of the DD/S with the
same objective in mind. We have compiled these into a separate docu-
ment of Support Information Processing projections for FY 1971-1975.
Time has been too short to have a good assessment of the value of this
exercise, but I believe it can be shaped into a significant planning
tool. Whether we will do the same thing again this year or try some
other way to get at the source of applications is still to be decided.
The point is that any Support component considering a new computer
application is expected to present the proposal to the IPC first.
16. I have the responsibility for the personnel management of
support careerists assigned to the SIPS .ask Force, and for approving
requests for ADP related training for anyone in the Directorate. Per-
sonnel management is a serious problem not only in finding the right
people but in trying to provide for their reasonable career develop-
ment and progression.
17. I believe this covers pretty well what the responsibilities
of the Support Directorate are and ought to be. I doubt that it would
be productive to suggest what the responsibilities of the IP Staff, the
IP Group, and the Executive Director should be until I have had the
reaction of the other IPC's to the approach discussed in this paper.
DDS/SSS:RRW:lja(5 Sept 69)
DistribUtion:
Orig IPC Coordinators-3
1 - SSS Subject
1 - 555 Chrono
25X1A
Information Processing
Coordinator, DD/S
6
Approved For Release 2001/08/08 : CIA-RDP78-04723A000400010004-8