COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	jeh@gvax.cs.cornell.edu
C00004 ENDMK
C⊗;
jeh@gvax.cs.cornell.edu
mcderm[s87,jmc]		Comments on McDermott's opus on AI

1. Make the title "AI as a branch of computer science".

2. Remove the assertion that computer science should really be
called informatics from the first paragraph.  I agree with the
remark mainly, but this isn't the place for it.  Perhaps the
transition McDermott wants could be accomplished by some such
phraseology as "Computer science is as much concerned with the
form of expression of information as it is with the algorithms
that process it.  Indeed the subject is called `informatics' in
much of Europe.

3. At the bottom of page 1, the sentence "We can define ... " is
fuzzy and not quite correct.  Better would be

AI is concerned with programs that achieve goals in certain kinds
of open-ended "information situations" where there may not be a
complete model of the situation in which the program has to act.
These are the situations in which human intelligence is most
effective and which offer the greatest difficulties for computer
science.

4. Apart from the above remarks, I think the essay is somewhat of
a sampler, but I think it will do.