∂28-FEB-75  1518		network site MAXC
 Date: 28 FEB 1975 1519-PDT
 From: MOORE at PARC-MAXC
 To:   JRA at SU-AI
 
 Yes, I got your previous message.  Sorry I didn't respond.  Things
 have been pretty hectic here too.  I like the idea of introducing
 sequences and illustrating the representation problem when you
 introduce sexprs.  I would caution you against getting too abstract
 but I trust your sense of balance.  Reworking the value thing
 is also a good idea.  Its such an important concept that you
 want to be sure its understood.  As you have already
 noted, it can be used a vehicle for introducing a lot of other
 ideas as well.  I would like to see the ms again when you have completed
 the next round of changes and are vaguely satisfied with it, if
 possible.
 J
 -------


∂14-JAN-75  1425		network site MAXC
 Date: 14 JAN 1975 1429-PST
 From: MOORE at PARC-MAXC
 Subject: SUPERLISP
 To:   JRA at SU-AI
 
 I have been sent a copy of the draft from Blake Vance.
 Is it up to date?  (Unfortunately, I took it home so I
 can't compare it with the copy Bruce gave me.)
 
 I agree completely with your comments about LISP and
 introductions to data structures.  Of course, that
 says less about the validity of your arguments
 than it does about my own bias.  I have long thought that
 LISP should be the primary vehicle in introductions to
 computer science in general (or at least that majority of
 computer science that can be divorced from hardware).
 The enlightenment that comes when you finally understand
 LISP is quite overwhelming, especially if you thought
 you knew computing.
 
 I agree with Knuth that your letter will invite countless
 responses of the form:  "LISP is full of shit, but XYZ-1.5
 is perfect".  That, to my mind, is one of the most
 discouraging aspects of computer science today.
 Furthermore, although the resulting discussion will
 be beneficial I am sufficiently pessimistic to conclude
 that you won't really win.  There are two
 problems.  The first is the incredible inertia that must
 be overcome.  For example, I think of LISP as an interactive
 language and unless one can actually get his hands
 on SEXPRs and SEE it happening, half of the joy is lost.
 So you have to sell both interaction and LISP implementations
 to an awful lot of people with fairly heavy investments
 in IBM cannon-fodder facilities.  Secondly, and perhaps this
 reflects my own cynicism, even if you "won" and
 LISP was adopted, I fear the same thing that happened to
 the "new math" would occur:  Namely, if you force people
 who don't understand a conceptualization of a subject to
 teach it that way, they will make a mockery of it and produce
 students who are only more confused than ever.  Of course,
 if every student had access to a '1O and LISP, you could
 throw the instructors out.
 
 J
 -------