Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!wuarchive!decwrl!argosy!freeman
From: freeman@argosy.UUCP (Jay R. Freeman)
Newsgroups: comp.lang.scheme
Subject: Re: Free Macintosh Scheme
Message-ID: <613@argosy.UUCP>
Date: 13 Jul 90 00:36:08 GMT
References: <601@argosy.UUCP> <12383@yunexus.YorkU.CA> <606@argosy.UUCP> <305@fjcp60.GOV> <612@argosy.UUCP> <ALMS.90Jul12192621@brazil.cambridge.apple.com>
Sender: news@argosy.UUCP
Reply-To: freeman@cleo.UUCP (Jay R. Freeman)
Organization: MasPar Computer Corporation, Sunnyvale, CA
Lines: 30

In article <ALMS.90Jul12192621@brazil.cambridge.apple.com> alms@cambridge.apple.com (Andrew L. M. Shalit) writes:

>  If you want to make your Scheme a real contribution, you should
>distribute it as source code.  That way people could use it as
>the kernel of applications, and also include it as an extension
>language. (and also extend it themselves!)
>
>     -andrew

Possibly sometime.  The code needs cleaning up and regularizing, which
I do piecemeal as I rework and add to it.  The implementation is in
some sense not done (though it is pretty complete "R3" Scheme as is)
-- I am adding features to the user interface, and I would like to get
my compiler to work well enough to release; that may require internal
changes for support.

Furthermore, MPW C has been rather a moving target:  I am now porting
to 3.1 from 2.0.2, and have encountered subtle bugs due to changes in
the Pascal glue.  (And floating-point is mostly broken in MPW C 3.1,
but that's another story.)

And how many people would digest 40000-odd lines of stuff well enough
to make changes and applications.  (Since I knew my implementation was
going to be mostly interpreted for a while, I wrote the built-in
Scheme procedures in the underlying C, rather than in Scheme itself.
This approach leads to lots of C ...)

                                            -- Jay Freeman

	  <canonical disclaimer -- I speak only for myself.>