COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	export[w85,jmc]		Control of Soviet access to computer technology
C00021 00003	Anecdotes:
C00022 ENDMK
C⊗;
export[w85,jmc]		Control of Soviet access to computer technology

	Transfer of computer technology now involves the repeated
copying of successively updated versions of complicated hardware
and software.  While knowledge of the general principles of a new
product is still helpful, the ability to copy newer and newer versions
exactly saves enormous time and cost.  This development offers new
possibilities for controlling the transfer of such technology to
the Soviet Union and for selling it to them on advantageous financial
political and arms control terms.  We can make it harder to steal the
technology and increase the advantages of buying it legitimately.

	Here are some relevant facts.

	1. It is not feasible to keep secrets indefinitely.  Only delay
is feasible.  We will later discuss the importance of delay.

	2. Computer technology is becoming more complicated and
more interconnected.  Transfer of such technology increasingly
involves the precise copying of complicated hardware and programs.
Knowing in principle how something works is still useful, but
there is an increasing premium on precise copies.

	3. Using a program on a copied computer architecture depends
on the architecture being an accurate copy.  Otherwise, it is difficult
and time-consuming
to distinguish bugs in the program or the programs that interact with
it from bugs in the architecture.  If precise copies are available,
people with lesser ability, intelligence, training and knowledge of
details can do the adaptation.

	4. Programs are often built to interact with previous
programs.  The availability of accurate and stable versions of the
previous programs is often the limiting factor in the development
of new programs.

	5. The debugging cycle is often a very large part of the
cost and time of developing a new product.  Knowing the general
performance specifications of a product permits developing a
similar product, but then the similar product must still go
through an expensive debugging cycle.

	6. Another area in which precise knowledge is important
involves communication and other interaction protocols.  Precise
specification of programming languages is important if they are
to be used.

	7. Typically hardware and software undergoes further
debugging even after it becomes commercially available.  Prompt
access to updated versions is important.

	8. The Soviet Union often copies American (I don't know
examples of other foreign) hardware.  They copied the IBM 360
architecture, the D.E.C. (Digital Equipment Corporation) PDP-11, and are
now copying the D.E.C. VAX.  Within the last year attempts to
smuggle advanced versions of the VAX through South Africa, Germany
and Sweden were frustrated.

	9. The reason for copying is to reduce the debugging cycle
and to be able to use foreign software.  It is important for them
to copy as advanced versions as possible in order to reduce the
amount of subsequent changes and debugging required.  Therefore,
if preventing smuggling the VAX 785 delayed their ability to
copy it by six months this was important to them.  They were forced
to copy the older versions they had already smuggled, and they
will have to update them in the future.

WHY SHOULD WE RESTRICT THEIR ABILITY TO COPY?

	1. There is the military reason.  The older their military
technology, the less money the West has to spend on defense, and
the less our society is dominated by defense considerations.

	2. When they copy they should pay for the right to do so,
just as any company in any country must pay the original developer
of the technology.

	3. They want our technology, and not only for military
purposes.  We want many things from them, and the more we can
prevent their stealing technology, the more they will have to
pay to get the technology legitimately.

HOW CAN WE RESTRICT THEIR ABILITY TO COPY?

	Companies that develop new products keep proprietary information
for reasons not connected with the Soviet Union.  They hope to
be able to meet certain market needs before their competition.  To
this end they do the following.

	1. They keep certain information "Company Confidential" for
prolonged periods.

	2. They make their suppliers and consultants
 sign non-disclosure agreements.

	3. They deliver prototype products with non-disclosure agreements.

	4. When they deliver programs, they often don't give source
code, making copying difficult and expensive.

	5. They patent hardware and copyright software and sue infringers.
As explained above, as hardware and software become more complex, the
value of copyright protection increases.  Hardware ideas, principles
and functions can be protected by patents, though apparently not
software ideas, principles and functions.  Copyright only protects
the precise realization of the idea.

	These actions are often effective in delaying the ability
of commercial competitors to copy.

	All the above forms of protection work against the Soviet Union
except the last.  So far as I know, the Soviet Union has never been
successfully sued for patent or copyright violation.  Nevertheless,
the Soviet Union has often bought technology licenses, because these
licenses include help in transferring the technology.

	In so far as the Soviets can be induced to move further in
the direction of buying licenses rather than illegally copying, the
agreements can include clauses in which further transfer is dependent
on their refraining from illegal copying of variants of the same
technology.

HOW CAN WE IMPROVE THE SYSTEM WITH REGARD TO SOVIET COPYING?

	1. Documents giving preliminary information about products
should contain a standard notice urging particular care in keeping
out of Eastern Bloc hands until the official release of the product.
While even released products are subject to export control, such
control is much more difficult once mass release begins.

	2. Companies should be urged to know their customers and
report suspicious requests for large volumes of information to
the Commerce Department or the FBI or appropriate authorities in
foreign countries.  It is infeasible and undesirable to question
individual requests for brochures, but when information is requested
in volume and detail than would require many people to evaluate it,
then a supplier should try to know what business his customer is
engaged in.

	3. For some products, like new microprocessors, it may be
feasible and desirable to supply preliminary products in variant
versions that will be incompatible with production versions if copied.

WHAT SHOULD BE OUR LONG TERM OBJECTIVE?

	The above ideas are relevant to any reason for restricting
technology transfer.  However, here is one idea that has perhaps
not been sufficiently explored.

	The Soviet Union is a closed society.  Most of it is off
limits to foreigners.  All industrial activity is a military secret
to a lesser or greater degree.  Soviet citizens have legal restrictions
on even social contacts with foreigners.  Besides the restrictions
on information for foreigners, the country is internally secretive.
For example, there are even secret criminal laws.  A person can be
convicted of a crime without being allowed to know the precise wording
of the statute he is alleged to have violated.

	Our ability to monitor Soviet compliance with arms control
treaties is limited by this secrecy.  There is no doubt that the
secrecy is somewhat of a military asset to the Soviet Union, although
communism has many additional reasons for secrecy.  If we want
them to relax their secrecy, we must pay.  They won't do it merely
to establish mutual trust, which they don't value anyway, nor will
they trade information they successfully keep secret for information
they can readily steal.

	However, our ability to delay their acquisition of our technology
is a valuable bargaining tool.  We can try to buy their secrets with
our secrets.  Naturally our companies want money too, but we can couple
Government permission to sell technology with relaxation of secrecy.

	We might trade technology for the following.

	1. Opening more Soviet cities to foreign visitors.

	2. The right to visit the places where the technology we
supply is used.

	3. The right to information about what is produced using
technology we sell them.

	4. The grand trade would be full rights to purchase Western
technology on the same basis as Western companies in exchange for
as full access to Soviet society as Western countries have to
each other.  Such a trade would be in the interest of Soviet
prosperity, though not in the interest of Soviet aspirations to
world domination.  It isn't likely to happen soon.

	Finally, getting full co-operation with making the Soviets
pay for technology they get from us depends on explaining what is
to be gained and overcoming misconceptions.  These explanations
should include the following.

	1. The Soviets copy extensively, and it is important to them.

	2. The importance of delaying access to technology they
will get eventually.

	3. The potential financial benefit to companies of making
it necessary for them to buy technology and products on the same
basis as anyone else.

	4. Companies should know about how many times the Soviets
have gotten samples supplied in a mistaken belief that this would
lead to larger sales.

	5. The importance of anything that can be done to open
the Soviet Union to maintaining peace.

	6. The political fact that unilateral concessions to the
Soviet Union are almost never reciprocated.  It is impossible,
according to Soviet ideology, to recognize as genuine a gesture
of good will.  A foreigner, who is not entirely subservient, can
get credit at most for being "realistic".

	This essay assumes a certain negative view of the Soviet
Union and communism that has been adequately supported by other
authors in other publications.  It does not assume they are about
to attack us.

A PROPOSAL FOR AN OFFICE OF COMPUTER TRADE

	After finishing the above, it occurred to me that a more
systematic approach to computer related trade with the Soviet
Union might be valuable.  It involves creating an organization
combining the information required to make such trade advantageous
to us.  It might be a Government office or it might be a non-governmental
organization operated by some institution like the U. S. Chamber of
Commerce.

	It would interact with companies doing computer business with
the Soviet Union, companies wishing to do such business, and companies
with technology the Soviets might covet.  Its first object is to frustrate
and deter Soviet attempts to steal technology or to swindle U. S.
companies out of it.  Its second object is help transfer to the Soviet
Union such technology as in our interest to transfer under terms that
make this transfer advantageous to us.  These terms include both price
and political and information concessions.  It might meet also with
representatives of
Soviet organizations, e.g. the State Committee for Science and Technology
and the USSR Academy of Sciences.
Within the U. S. it would have to interact with the Commerce Department,
the Defense Department, the State Department and the Ofice of Science
and Technology and relevant Congressional committees.  Since there would
be a substantial possibilities of turf battles within the Government,
making it a non-governmental organization might have advantages.
Anecdotes:

Use of East German purchased IBM computer to debug Soviet Bloc
software and peripherals.

Fredkin efforts to sell Soviets on the importance of microcomputers.

Their unsuccessful attempt to smuggle VAX 785 on top of successful
smuggling of other hardware.