[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Proposed Justification Letter

I agree that RPG's letter should be held in reserve, for the reasons you
say.  I think that most of the community will simply be happy to see
that the committee is chosen, and the main pressure will be to proceed
to the next step.  Let's wait a few more days and see if there are any
other questions or proto-complaints about our procedures, before we do

Gee, don't you remember Alan Bawden?  Do you still have a copy of "Votes on
the First Draft Common Lisp Manual?"  He was one of the major contributors.
But, here's as much of a resume as I can think of:

Alan Bawden wrote the third and definitive version of the Lisp READer in
the MIT Lisp Machine system (Symbolics, TI, LMI), with a lexical
analyzer driven by an FSM produced by a "read table compiler", which is
something like a standard one (a la Aho and Ullman) but with several
tricky modifications because Lisp syntax required some extensions.  He
also wrote the PRINTer.  He wrote DEFMACRO and the backquote facility,
and was the main designer of them, based on various earlier and less
elegant hacks that had been floating around the Lisp community.  He was
the first person to figure out and understand nested backquotes.  He
wrote the full version of DEFSTRUCT.  He also wrote the microcode
implementation of bignums for the CADR/LM-2.  He and I did a lot of
design and implementation work on the package system, going through
about four different philosophies and implementations.  He worked out a
lot of the basic problems and various strategies for dealing with them.
He also did some amount of general system maintenance and bug fixing,
although not to the extent that Moon and I did.  All this work was done
at MIT and is/was used by all three companies.

In addition, he wrote the initial bignum implementation for the 3600,
during a summer job at Symbolics; that was his only official connection
with Symbolics (and his only connection now is that he sometimes sends
extremely critical complaints and bug mail).

He was also instrumental in developing BSG's Lisp course (along with
BSG, ACW, and myself), and he taught the course one year.  A friend
who took the course found it extremely clear.

At MIT, he also spent a great deal of time experimenting with toy Lisp
implementations and compilers, to compare various strategies and things
like that.  I am not familiar with what this was all about (Moon may
know more).  In fact, since I left MIT in 1979 I have not been very
familiar with what he's been up to.  The biggest thing has been his work
on the Connection Machine, in which he developed ideas about how to find
a clean abstraction that was at once close enough to the level of the
hardware that it could take advantage of the extreme parallelism, yet
close enough to the level of Lisp that it was relatively easy to
understand and program in.  His Master's thesis was on such work.  It
included other things, too, but I haven't read it so I don't know any
more details.

Also at MIT, he worked with Glenn Burke (and someone else?) on a package
of macros and things to make some advanced Lisp features available in
Maclisp.  These included defstruct and defmacro or something like that;
I never used Maclisp at that time so I don't know the details.  They
published a little reference memo as a MAC technical memo.  I think he's
done a bunch of other things like this but I can't remember more than
this.  I've known him for 11 years now and it's hard to think of
everything.  I think he's done other things with Glenn.

Bawden is a mathematician at heart.  His extreme clarity of thought and
presentation, and deep understanding of complex issues, were extremely
helpful in dealing with many of the projects mentioned above.  (At MIT
he got all A's, except for a couple of B's in some humanities courses,
majoring in Math, and he got unsolicited letters of recommendation from
math professors.)  However, he is also an excellent engineer when he
wants to be, which makes him unusual as someone who can appreciate (and
grok and criticise) highly theoretical issues and highly practical
issues both with great skill.  He is a serious system hacker, and was
one of the main people who recently brought up ITS on the DEC 2020 at
the A.I. lab.

The only criterion that I think might give us trouble is "prominent",
"internationally known", etc.  I don't see how to make a case for this
one.  However, while this criterion is good and important, in my opinion
the other criteria and the factors above make Bawden an excellent