various thoughts

Drew R Whitehouse
23 Apr 1997 15:40:29 +1000

>>>>> "Richard" == Richard Coleman <> writes:

    Richard>    I guess this is a religious issue, but in this case it
    Richard> probably doesn't matter much.  Much of the philosophical
    Richard> arguments between Lispers and Schemers are
    Richard> minimalists/pragmatists arguments.  I believe working on
    Richard> a lispOS is more in the Common Lisp philosophy in this
    Richard> regard.

    Richard> 4) How ambitious to be.  I'm a big believer in starting
    Richard> small in order to get started.  If you set initial goals
    Richard> that are too hard, frustation will set in.  A good
    Richard> starting point might be a small, minimal Linux
    Richard> distribution with pre-built Lisp (or Scheme or scsh) and
    Richard> lots of Lisp pre-installed.

	Hello Richard, 

	  One system that is quite interesting and possibly addresses
many of the issues of scheme vs common lisp is the bigloo system
( which has many things
going for it. I just did a small (mostly meaningless) test - a
stripped hello world executable is 9124 bytes.

Quoting from the README -

"This directory contains the source files of Bigloo, a Scheme system
which includes a compiler generating C code and an interpreter. Bigloo
is the tool of choice for the construction of small autonomous applications
in Scheme. Bigloo is conformant to
IEEE Scheme standard with some extensions:
        Rgc, a lex facility.
        Match, a pattern-matching compiler.
        Foreign languages interface.
        Module language.
        Extension package system.
        An Lalr facility.
        An Object sytem.", the author, has been friendly
and helpful whenever I've need help, which I think is quite
important. The fact that the system is designed from scratch for tool
building rather than just for educational or academic purposes is a
plus IMHO. (As someone pointed out recently in the endless Ousterhout
vs Lisp threads - scheme authors, usually CS people, don't get
promoted for writing and supporting good, complete software. I guess
this is why Ousterhout left for the .com world).

	RScheme may be a possibility but the documentation is sparse
and the system seems pretty slow in comparison. Gambit is very nice
but the copyright is onerous (from the README - "The Gambit system and
programs developed with it may be distributed only under the following
conditions: they must not be sold or transferred for compensation and
they must include this copyright and distribution notice.  For a
commercial license please contact `'.")
Mzscheme is nice, very well documented, has a module system and a
(bloated) GUI, but is interpreted (as is scm, siod, guile, elk etc). I
think common lisp systems, as they stand at the moment, are just too
large and hairy. Just a few thoughts from someone who has played with
many of the current scheme scheme systems floating around the net.....

;; Viz programmer, Australian National University Supercomputer Facility
;; Unsolicited commercial e-mail will be billed at $250/message.