Newsgroups: comp.lang.scheme
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.mathworks.com!newshost.marcam.com!charnel.ecst.csuchico.edu!csusac!csus.edu!netcom.com!NewsWatcher!user
From: hbaker@netcom.com (Henry Baker)
Subject: Re: Removing READ
Message-ID: <hbaker-2202951621300001@192.0.2.1>
Sender: hbaker@netcom3.netcom.com
Organization: nil
References: <dig-Scheme-7.45@mc.lcs.mit.edu> <9502160422.AA05023@clark.lcs.mit.edu> <hbaker-2202951004460001@192.0.2.1> <bakulD4FAss.C7H@netcom.com>
Date: Thu, 23 Feb 1995 00:20:00 GMT
Lines: 38

In article <bakulD4FAss.C7H@netcom.com>, bakul@netcom.com (Bakul Shah) wrote:

> hbaker@netcom.com (Henry Baker) writes:
> 
> >One look at all of the bletcherous '.xxxrc' files proves this statement
> >dreadfully wrong.  The world would be dramatically improved if nearly
> >_all_ programs parsed s-expressions for '.xxxrc' programs instead of
> >re-inventing the parsing/language wheel for every single init file.
> 
> This is an appealing idea but usually .<program>rc style files
> use a little language that is better suited *for describing
> simple things*.  Scheme/Lisp usurp a number of characters and
> that forces one to use a much more verbose syntax.  I find syntax
> style such as
> <keyword>: <value>
> much more pleasing and easier to read/modify.

This isn't very general.  For example, it doesn't provide any
mechanism for nested expressions (e.g., for specifying nested
menus, which some programs provide in completely 'soft' form).

> Where these
> .<program>rc files go wrong is in inventing their own syntax
> for doing complicated things.  For that Scheme or Self or some
> such language would be perfect.  (and no, the idea of using two
> languages does not disgust me!).
> 
> What would be more interesting is the use of Scheme as a
> *communication* language; where programs throw back and forth
> s-exprs (instead of sending just data in some arcane &/or
> proprietary formats) to communicate with each other.  Perhaps
> s-exprs encoded in some sort of a compact/efficient binary
> format.

I believe that a number of 'parallel' Lisps do exactly that -- pass
S-expressions back and forth.  The 'Linda' language can be seen as
a kind of broken version of this, except with a pattern-matching scheme
to retrieve S-expressions from a shared pool.
