COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	I just got your  first set of comments Tuesday, and am
C00008 ENDMK
C⊗;
I just got your  first set of comments Tuesday, and am
glad to seee that they are indeed criticisms. I don't agree with
all of them  and indeed some are based in misunderstanding of what I
was trying to say. But if you are confused then so too will be the
thousands of students who buy the book.   Before I attempt to 
untangle some of the confusion I would like to clarify one of my 
worries.  Since your comments and criticisms are strong I worry that
you might lose patience with this reviewing process. I would hope not;
the detail you have given the first sections has been valuable and I would
not like to lose the rest of the review. However if you do not feel that
you have time to complete the manuscript please say so soon. I have until
January to get a completely revised manuscript to McGraw-Hill; between
full-time work and part-time teaching, its going to be close. I have been
chasing Ed Feigenbaum around trying to get some money to pay reviwers, but
he seems to spend most of his time in airports. He will be back next Tuesday
and I will camp on his door-step again.

Thus the main point of this note is to thank you for your
comments and to make sure that they aren't your last ones. 

quick comments on some of you comments before i have to run. I'll try to
get some moe coherent remarks to you later.

The audience is meant to be uppper-division to beginning grad level student
unaware of LISP. The point is not to make them into LISP hackers but to
introduce as much of what (I think) is interesting in CS. I am very
much opposed to the Knuth-data structures approach to the subject; and
am equally opposed to the Weissman lisp primmer approach to programming languages.
I learned LISP from LISP1.5M and don't see what everyone is bitching about.
(blush: i also programmed on the 704..."address register" indeed!!!)

I didn't use a "top" element since I would rather use complete partial orders.
The arguments for "top" have never been very convincing. error situations
should be modeled as error situations; perhaps with Milner's or Milne`s
methods; don't know. I did not use a coalesced sum since I would rather
keep the ideas of strong-typing around in the introduction; but alas
i did screw things up.

Oh, my!! " m-expresssions are hard to read"; I go the m-to-s expression route
because that best shows the power of representing program as data structure.
(my that is a vacuous statement...  best thing to do is let you read next two
chapters)

Yes indeed lisp is a machine language; i point that out farther along.
i don't think that point has much meaning until the student understands
eval.

Yes cuteness will go. much is the remains of being lecture notes; much is
just my obnoxious behavior.

I don't think a univesristy text book is the place to discuss the system level
details of LISP. I.e. how to run  Interlisp or Maclisp; that discussion
is best done with an on-line terminal and a manual. i am against
teaching  any language, even LISp, as the main diet of a college course;
you learn it by doing it. 

Damn, gotta run!! I hope you will continue the review; your comments are 
indeed valued.  I will pry money out of Feigenbaum.

				john