LETTER TO (Sanitized)

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78B04770A001500060041-6
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 4
• Document Creation Date: December 28, 2016
• Document Release Date: September 26, 2005
• Sequence Number: 41
• Publication Date: June 28, 1968
• Content Type: LETTER

Approved For Release 2005/11/21 : CIA-RDP78B04770A001500060041-6
June 28, 1968
On June 19th, I visited the
to review the optical system of a
The group described in detail the optical
system they p 6n to use in this scanner. The correlation
principle was also described.
It was apparent that most of the efforts in this
contract were limited to the mechanical and the electronics
of the device. The budget on the Optics was limited and
it was decided to use on-the-shelf optics. It is this
part of the problem that I will report on.
As a general principle, it is difficult to use
off-the-shelf items for a sophisticated optical system.
The reason is that optical systems are almost always de-
signed for a specific task and they are carefully optimized
to meet requirements. This means a careful compromise has
been made.
In mocking up a series of lenses, it is extremely
unlikely that all the lenses will be used in the manner
for which they were designed. Therefore, there is practi-
cally no possibility that an optical system mocked-up out
of shelf items will be an optimum design. There is, however,
justification for considering the use of off-shelf items
under the following conditions:
NGA RevieWAQ9Phe%r Release 2005/11/21 : CIA-RDP78B04770A001500060041-6
 Approved For Release 2005/11/21 : CIA-RDP78B04770A001500060041-6
June 28, 1968
Page 2
1. Optical elements would be very expensive.
2. The optical requirements are modest.
3. A short time requirement.
4. A limited budget.
It appears that these conditions for 0 except
for 2, are true. Therefore, it appears reasonable that
lelected to use off-shelf items.
If shelf items are to be used, it is necessary
to set up an exact bench test of the system and determine
if the optical system performs. Since almost nothing is
known of the off-shelf item, it is essential that the mock-
up duplicate exactly the optical system required. It is
exceedingly dangerous to extrapolate from the mock-up to
the final system. This procedure sounds cheaper and easier
to do than it really is. There is a grave risk that the
mocked-up system will not perform according to specification.
The reason for failure will be obscure because nothing is
known about the lenses. There is an almost irresistible
urge to conclude that the failure is due to the mock-up not
being made up as well as the final system will be. A
particularly inviting hazard occurs when the system almost
meets the specifications. It then becomes very difficult
to say: "Stop! This is not sufficient. We will have to
redesign a new system." The mock-up, almost meeting the
specifications, provides strong pressure to convince the
project leader to accept the design "as is", in place of
starting all over again from scratch. Experience shows
that the project managers will divert their attention to
convincing themselves that it may be good enough or that
by hunting a bit further they may find the lenses that they
need.
If a project manager decides to go the route of
using off-the-shelf items, it is tremendously important to
carry out the following procedures:
1. Test the individual lenses and thoroughly
understand how the lenses will interact
with each other.
2. Put together a mock-up of the complete
system.
Approved For Release 2005/11/21 : CIA-RDP78B04770A001500060041-6
 Approved For Release 2005/11/21 : CIA-RDP78BO477OA001500060041-6
June 28, 1968
Page 3
3. Thoroughly test the complete system
to be sure it meets with the specifications.
4. Never plan on an optical system that has
not been checked. out in advance.
If these procedures are followed, one finds that
on-the-shelf lenses do not offer the advantages most
people think they have. When off-the-shelf optics prove
to be successful, it usually indicates that the optical
system problem is not very dema:tiding.
rules:
SUMMARY
does not have sufficient evidence to predict
that this system will perform as well as will be needed.
They indicate that there are no real specifications. The
instrument was built to compare stereo scanning with monocu-
lar scanning. The stereo scanning adds a great deal of
optics. Even with an optimum design, it would almost cer-
tainly degrade the imagery seen through a monocular system.
group essentially violated all of these
1. They did not have tests on the individual
lenses, and were quite unaware of the type
of aberrations, vignetting or pupil
distortions. They had done essentially no
analysis which would enable them to
estimate the field curvature, secondary
color, pupil aberration, or distortion.
2. They had not put together a complete mock-
up to prove that it would work. They did
have a similar system mocked-up as a
monocular instrument, but there was suffi-
cient difference to cast doubt on the
validity of the comparison. Even with
this, however, they had not made sufficient
measurements to prove to me that the system
could meet the specifications.
Approved For Release 2005/11/21 : CIA-RDP78BO477OA001500060041-6
 Approved For Release 2005/11/21 : CIA-RDP78B04770A001500060041-6
I _T
June 28,
Page 4
I suppose there is a chance that the stereo system, with
inferior optics, may enable people to perform better than
without stereo, but if the stereo system does not, then
what? Will the whole project fail? Will it be possible to
attribute the failure to optics or electronics? It would
have seemed to me that more consideration should have been
given to the optics for this instrument, when the whole
device depends so heavily on subtle visual effects.
RECOMMENDATION :
I Ishould be asked to provide adequate informa-
tion on the optics and to provide a more defensible argument
that the optical system they propose is adequate.
Finally, my note is hastily written and appears to
be very harsh on To be fair, I should say that their
approach is what one usually finds in design engineering
groups. I was disappointed that 0 had not pressed the
optics more vigorously, but I am afraid the same approach
would have been taken at most companies capable of doing
such a large project. It is also well to remember that
hindsight is far easier than foresight.
Before I could recommend what you people should do,
I would have to spend much more time than I did.
Sincerely,
Approved For Release 2005/11/21 : CIA-RDP78B04770A001500060041-6