From: Brian Harvey <[email protected]>
Subject: Re: [r6rs-discuss] Proposed features for small Scheme, part 9: modules
Date: Mon, 12 Oct 2009 18:10:50 -0700

> > I personally need it to ensure that the
> > program I installed on a client machine includes everything
> > it needs).
> 
> I would say that if you know how to say "client machine" you are the
> very paradigm of a WG2 customer!

I disagree.  The very reason I bet my living on Scheme is that
I believe small core language and industrial strength can get
along.  Of course I don't expect WG1 Scheme provide everthing
I need to write commercial apps.  I don't even expect WG2 Scheme
does.  I do expect WG1 Scheme provide a basis, on top of which
various SRFIs can be provided.   So, I'm flexible about moving
some features out of the language, but I'd like to make sure
that the language won't make it impossible (or very difficult)
to provide such features on top of it.

Module system can be in a language with a simple form as in John's
proposal, or can be outside of the core language as a separately
defined meta language (maybe in SRFI).  I'm open to either way.
I've been living mostly happily with R5RS + very simple module system
+ cond-expand + other SRFIs, and I don't think it's unreasonable to
put a simple one into a language.  But if people insist Small Scheme
shouldn't have phase distinction, it might be actually a good idea
to kick the meta information outside of the core language.  Then
the behavior of module system may be explained as a simple-minded
preprocessor that takes bunch of Scheme code fragments and
generate a monlithic source that can be executed by simple Scheme.
(Implementation can manage Scheme sources more cleverly, of course.)

--shiro

_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss

Reply via email to