COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "let.pub[let,jmc]" source
C00005 ENDMK
C⊗;
.require "let.pub[let,jmc]" source
∂AIL Dr. H.E. Bruderer↓Institute of Computational Linguistics
↓Finkenweg 3↓Postfach 149↓CH-3110 Munsingen/Bern↓Switzerland∞

Dear Dr. Bruderer:

	The Advice Taker was meant as a general problem solver, but
implementing it depended on obtaining satisfactory axiomatizations
of common sense knowledge.  I have been working on this since 1958,
and I believe I have made progress, but I am still not ready to
write a general problem solver based on these ideas even though
I think it is a good approach once more progress has been made
on the theoretical problems.  My 1977 IJCAI paper surveys results
obtained so far and contains references.  In the early 1960s, Fisher
Black, at Harvard, wrote a program he called the Advice Taker which
solved some easy problems, and in the late 1960s Cordell Green pursued
some of the same approach at Stanford and SRI.

	SHAKEY was an SRI project, not a Stanford University Project,
and Peter Hart at SRI could give you more information.  However, I do
know that SHAKEY has been mothballed.  They may give various rationalizations,
but I believe the truth is that the Government, in a misplaced
fit of short range practicality, refused to fund further work, and
demanded other projects that promised quick practical results.  Of
course, the results were not forthcoming.

.sgn