COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	I would favor a small required reading list.  Ideally it should be one
C00005 ENDMK
C⊗;
I would favor a small required reading list.  Ideally it should be one
book.  Asking the student to use his judgment is not fair, because the
faculty is not bound and is unlikely to respect his judgment.

change the title from "Application of formal logic to AI" to
Epistemological problems of AI and the use of formal logic" and
change the first 3 sentences to
The problem of describing facts about the world including the effects
of actions has been studied apart from specific problem solving
programs.  This work has used first order logic to express facts
about the world.  These issues are discussed in [McCarthy and Hayes]
and in [Hayes DEFENCE].  Many current issues are discussed in [McCarthy
5IJCAI].

The most accessible reference to my reply to Weizenbaum is the
Stanford AI memo containing reviews by myself Buchanan and Lederberg.

I think the syllabus should downplay philosophical and social issues,
because the we certainly don't want to grade students on their own
views, and the field is fluffy enough as it is without letting or
requiring students to be able to regurgitate various people's views
on the issues.  If the faculty wants students to have exposure to
these issues it should require attendance at a seminar or lecture
series and take attendance but not examine.

I guess I would advocate omitting or drastically condensing the
sections on philosophical implications, political and social
implications and history and politics of the field.  However, on
philosophical implications there is a new important reference,
namely Aaron Sloman's "The Computer Revolution in Philosohy".

I regret the demise of MICROplanner which had a clearer and simpler
structure than its successors as well as having more actual use.