From phelps@palm.CS.Berkeley.EDU Fri Dec 18 12:55:35 1992
To: fateman@cs.Berkeley.EDU
Subject: CGOL notes
Status: R

In the rush to finish up the paper I forgot to include two important notes
about the CGOL conversion.  First, when I rewrote the r-e-p loop I took out
the error handling catch-throw construct, so now an error falls back into
Lisp; it is a simple matter to put this back in.  Second, looking through
the source file you will see a couple areas in all uppercase, which stand
out from the sea of lowercase; these are for let and arrays, which have
changed enormously and for which one needs to decide whether it would be
better to construct some artificial interface to the numerous optional
arguments (as in making arrays) or better to have the user escape directly
into Lisp (with the ! mechanism).

-Tom

From fateman@peoplesparc.Berkeley.EDU Fri Dec 18 14:47:18 1992
To: phelps@palm.cs.berkeley.edu
Subject: Re:  CGOL notes
Cc: fateman@peoplesparc.Berkeley.EDU
Status: R

if you want to incorporate your notes into your original tex file -- and
it would be nice to make it into one file (no includes) -- it would help
make re-distribution easier.  I understand the variations on arrays, but
why is "let" a problem?  (I haven't looked at your code..)
RJF


From phelps@palm.cs.berkeley.edu Fri Dec 18 15:14:09 1992
To: fateman@peoplesparc.Berkeley.EDU
In-Reply-To: Richard Fateman's message of Fri, 18 Dec 92 14:50:50 PST <9212182250.AA05181@peoplesparc.Berkeley.EDU>
Subject:  CGOL notes
Status: R

The next two e-mail messages have (1) the paper as a complete unit and (2)
the accompanying bibliography (.bib) file.

The problem with let is that, currently, it's used for a weird lambda form.
What's used as local variables aren't really local at all, with the dynamic
scoping making them visible everywhere during their extent.  I think there
should be a new command that looks like this:

local a:=5**2, b, partial_sum:=0, list_so_far:=[];

which is a CGOL-copacetic version of Common Lisp's let.  But right now I am
tired of working with CGOL (the project took much more time than I
anticipated; it wouldn't be too much of an exaggeration to estimate 100
hours, including the writeup).  As I mention in the paper, you could go
arbitrarily far in constructing new syntaxes for all the new features of
Common Lisp.

-Tom