Re: Scheme pre-R6RS Workshop at ICFP - What is the Point?

[Alan Bawden <Alan@lcs.mit.edu>]

   Date: Wed, 27 May 1998 07:01:29 -0400
   From: Richard Kelsey <kelsey@research.nj.nec.com>

      Before I spend any more effort (and money) discussing records,
      exceptions, modules, objects, binary I/O or Unicode characters, I
      would like to know more about what our -goals- are, and about the
      -process- we plan to use to achieve them.  Do we want Scheme to
      remain a teaching/research language that can only be used for
      real-world programming with great difficulty?  Or can we come up with
      some way to satisfy both camps?  Do we want to change the way
      language decisions are made?  Abandon consensus?  Do everything under
      the IEEE process?  Appoint a Scheme Czar? Freeze the language and go
      home?

   Why do you think discussing these issues will be any more fruitful than
   a renewed discussion of records?

Maybe it wouldn't be any more fruitful, I don't know.  But more talk about
records is highly unlikely to be fruitful until we have some agreement
about the -direction- the language is going.  If we agree that Scheme is a
teaching/research language only, then I may conclude that records should be
kept out of Scheme.  If we decide that Scheme consists of a simple core
language, plus a rich library of utilities, then I'll have to decide what
piece of a record system belongs in the core.  If I knew what mechanisms we
were going to use to make decisions, then I'd understand how to argue my
point of view effectively.  Etc.

In my opinion, if we can't have a fruitful discussion about the process,
then we are screwed, and we should just give up.

				- Alan