COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "let.pub[let,jmc]" source
C00009 ENDMK
C⊗;
.require "let.pub[let,jmc]" source;
∂CSL Dr. Norman Metzger↓National Research Council
↓2101 Constitution Avenue↓Washington, D.C. 20418∞

Dear Dr. Metzger:

	Here are a few comments on the %2Five Year Outlook . . .%1
emphasizing
Schwartz's section on Artificial Intelligence

	There is little wrong with what Schwartz says, but it
doesn't express the current directions of research in artificial
intelligence in the way that the other sections of the report
describe their fields.

	The statement on page 85, line 13 %2"Arbitrary collections of axioms
that express the laws of particular mathematical and nonmathematical
domains can readily be written in this calculus"%1 is misleading.
No-one has yet succeeded in so writing the major common sense facts 
of causality, space, time, knowledge, belief and goal-seeking in
predicate calculus or any other formalism in such a way that it
is just a search problem to find consequences.  Writing these facts
and formalizing the non-deductive, i.e. non-monotonic, methods
of common sense reasoning are the major epistemological problems of
AI.  Philosophers recognize them too, but can't yet solve them either.

	A sentence is repeated at the bottom of page 89 and the top of page
90.

	The linguists, especially Chomskyans, will flip at any hint of
ascribing any credit to AI for applications of context-free
grammars.  Most programming languages don't have context-free
grammars, even ignoring requirements for agreement of the usage
of an identifier in its different occurrences.

	Here are some unsolicited comments on other sections.

page 6, line 14

	Following the market projections is a case of the blind leading
the blind.

page 23, line 16.

	I doubt that biomass is cheaper than petrochemical feedstocks now
or in the near future, and I would be very surprised if industrial
scale biochemical reactions have "no real by-products".  These are
minor points, however.

page 26, line 9

	This paragraph seems naive.  In many of the discussions of recombinant
DNA, "the public" has meant ideological lobbies.  The sense in which
there "are discussions between science and society" is unclear and
probably can't be clarified.  While human genetic engineering may
eventually require laws, etc., not many people interested in politics
will give them much attention along as they are hypothetical.  This
is right, because our society is beset by more immediate problems.
Imagine the result of having asked each of the candidates in the
recent presidential election for a statement of his position on
human genetic engineering.  If any reply at all were received, it
would have come from the least busy speechwriter.

	The remark applies to most demands that an issue that will arise
in the future be discussed "by the public or by society" now.
My experience is with demands for the discussion of AI.

page 40, et. seq.

	The discussion of energy storage ignores nuclear fission energy.
This will lead to a perhaps justified suspicion in the new
Administration and its Congressional supporters that the scientists
often let their ideological beliefs unduly influence their scientific
statements.  The omission is particularly striking on page 41,
line 18 and at the top of page 42, but it is noticable elsewhere too.

page 61, line 14

	This discussion presupposes that a "Law of the Sea" agreement
is required before submarine resources can "legally" be mined.
I think this isn't actually so either in U.S. law or previously
accepted international law.

page 64, line 14

Rutile should be identified as a titanium ore.

The section on the Space Shuttle seems to take for granted that
the main motivation for man-in-space is scientific.  I think
the motive of exploration as distinct from scientific investigation
will become frankly dominant in the 80s.

	I hope these comments are helpful.

.sgn

cc: Jack Schwartz