COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	GOODMAN%uamis@rvax.ccit.arizona.edu,
C00008 ENDMK
C⊗;
GOODMAN%uamis@rvax.ccit.arizona.edu,
DUANE.ADAMS@c.cs.cmu.edu, BLUMENTHAL@a.isi.edu, DONGARRA@anl-mcs.arpa,
 GANNON%RDVAX.DEC@decwrl.dec.com, JAHIR@athena.mit.edu, HEARN@rand-unix.arpa,
 JLH@sierra.stanford.edu, JMC@sail.stanford.edu, KNEMEYER@a.isi.edu,
 MCHENRY@GUVAX.BITNET, OUSTER@ginger.berkeley.edu, Ralston@mcc.com,
 CWEISSMAN@dockmaster.arpa
comments
global[e88,jmc]		Comments on draft of Global Trends in Computer Technology

1. I think the report assumes to much about what the attitude of the
U.S.  toward the Soviet bloc is likely to be during the lifetime of
the report.  If a definite attitude were to be chosen, I have no
quarrel with the moderately hardnosed line of the report, but with
only a little more work the report could be useful over a spectrum
of attitudes.

	To take one view, suppose the U.S. decided to help the
Soviet Union a little.  How would we do it?  In my opinion, the
thing to do would be to encourage use of normal commercial channels,
especially including post-sale service for both software and
hardware.  Such an attitude might arise in the U.S. Government
if major progress is made on disarmament or, even better, the
Soviet Union were unilaterally to reduce the fraction of its
GNP going into the military.

	Also one can imagine swaps involving access to technology for
the Soviets in exchange for much freer access to Soviet territory, e.g.
the elimination of closed cities.

	On the other hand, one can also imagine things getting
much worse if Gorbachev were forced out under the nationalist slogan
that he was selling out to the West.  The we and our allies would
be inclined to more severe restrictions, and the report should also
provide some technical basis for that.

2. The section on software is rather strongly oriented toward
the "strongly typed" style of programming languages.  LISP and
Prolog are not mentioned and AI is not mentioned in that section.
I'm not sure that it matters very much, but I suppose the Government
might want guidance at some point on the export control of LISP,
Prolog and expert system shells.

I recognize that I should have written a section on AI, but I
couldn't bring myself to do it, especially as I have no current
knowledge of Soviet AI.

3. Standardization didn't begin 15 years ago.  IBM tape formats
became standard in the late 1950s.  The 360 became somewhat of
a standard in 1964 with RCA and Japanese companies deciding to
make compatible machines.  The Soviet decision to standardize
on 360 architecture dates from 1965 or 1966.  Algol was an
attempt to standardize programming language for numerical
computation.

4. Software is easier to control than the report suggests, because
adopting it depends not merely on acquiring a copy but on maintaining
access to updates, documentation and bug fixes.  I remember being
told that often the Soviet copies of the documentation don't correspond
to the software they have managed to copy.  Adopting a major piece
of software may involve a commitment to continual adaptation to
hardware, operating system and network changes.  If the co-operation
of the original software producer isn't available, the user may
have to cut loose from compatibility at some time.  My impression
is that the Soviets have often had to do this.

This means that offering regular access to software maintenance
can be a bigger carrot than the report indicates.  Of course,
not all Soviets are aware of this, but some are.

5. The BESM-6 was not up-to-date when it came out because of
its limitation to a 32K word memory.  This was just after the
360 offered a 32 bit address, cut down to 24 by IBM shortsightedness.