On 6/15/07, Thomas Lord <lord at emf.net> wrote:
>
> Paulo J. Matos wrote:
>
> > This is my first post to this ML but I'm following the discussion
> > since the beginning and I can't understand your reasons at all.
>
>
> I've enclosed a guide that might help.
Here's a guide that attempts to provide more explanation, rather than
persuasion.
I would say that there are, broadly, three sets of people who find
R6RS controversial as currently constituted.
- People who agree with, or at least accept, the rough outline of
R6RS (Unicode, syntax-case, libraries, records, etc) but who disagree
strongly about particular technical matters - strongly enough that
they think the draft *might* need to be rejected and revised. I would
say that Will Clinger is a representative member of this group.
- People who think the concept of R6RS is legitimate, but who disagree
with the technical content of the largest changes in R6RS (Unicode,
syntax-case, libraries, case-sensitivity, etc). These people tend to
have different opinions about how Scheme should evolve (more low-level
programming support, or first-class environments, etc), and often are
associated with a more Lispy approach to language design. This group
of people has certainly always been a part of the Scheme community,
but with the demise of the consensus requirement, they seem to have
lost the struggle for R6RS. I think Chris Hanson and 'felix', the
author of Chicken, are members of this group.
- The final group is people who think Scheme standardization, of the
sort represented by R6RS, is a fundamental mistake. They think of
Scheme as a language design idea, not as a particular programming
language, and thus making particular language design choices within
the "Scheme" space, and decreeing them better than others, is a
fundamental misunderstanding of the role Scheme should play. They
like the SRFI process, because it allows some people to say that they
like one point in the design space, and others to prefer another,
without closing off anyone's freedom in the slightest. The
representative member of this group, of course, is Tom Lord.
Having described these groups as fairly as I can, not belonging to
them, let me say a bit about why I think the latter two are misguided
(while I disagree on most of the technical matters with the first
group, I think they are an important and useful part of the Scheme
community and the standards process).
The second group (which hasn't made much of an appearance in the R6RS
process) has been overtaken by events. The major Scheme
implementations and communities had already decided the broad outlines
of R6RS, before the editors even began their work. The time to oppose
syntax-case, and modules, and Unicode, has now passed - they have been
successful and widely adopted. The alternative to adopting them was
adopting nothing.
The third group has been overtaken by different events. The Scheme
community no longer consists of a small number of people, all of whom
are associated with an implementation that they take part in, as it
might have in 1992. Most people who use Scheme don't implement
Scheme, they use it to write real programs. And they thus want
Scheme, as a language, to evolve to help them write those programs.
The production of a new language design, whether it is called Scheme
or not, doesn't help them, since they've already chosen the language
they want. What does help them is evolving the language they've got,
and making it more portable and better-specified.
These claims offer little consolation to the people who've been left
behind by the decisions made by the R6RS committee, but they perhaps
offer some explanation of why things have turned out as they have.
--
sam th
samth at ccs.neu.edu
Scheme User since 1995
Received on Fri Jun 15 2007 - 15:30:08 UTC