2009/2/21 Ray Dillinger <[email protected]>:
> because they thought it was ratified before the process was finished.
>
> The committee seemed to be in a rush to publish the standard,

This, in the light of memory, was probably my greatest reason.
Developing the spec is not the same as a commercial development. The
deadline had no connection to anything other than the groups own
desire to bring the process to a halt in a certain time frame.

> word count, but it contributed to a feeling that crucial choices were
> made without considering (or acknowledging) the community's input.

Actually I am not over-fussed about "community input". Most of the
editors are significantly smarter than I - or at least they have been
thinking about these topics in greater detail over more years and I do
largely trust them for technical accuracy.

> In asking for statements from dissenting voters, the R6 committee
> compounded this error by asking people to articulate the very
> disagreements and alternatives that they had expressly failed to
> give rationales for rejecting.  Thus, the dissenters were explicitly
> given the last (and only) word on the matters

Requiring the reasons for dissent makes sense on many levels, but it
was not taken far enough by the-then SC. Given the names and nature of
the dissent, there were some people whose dissent *alone* should have
been enough to send the document back for another round. But the
problem of the arbitrary deadline made that impossible.I believe that
one of the key tasks of the SC is to maintain the technical and
aesthetic integrity of the language. That belief has many implications
I will not delineate here.

But I will say a bit about "aesthetic integrity", and why Olin Shivers
moved up my ballot with his statement. This also goes to why I wanted
Jeff Siskind on the SC as well. Whoever mentioned that Christian's
agree on the fundamentals but disagree as to the exact members of the
set of fundamentals was precisely correct in his analogy. There
certainly fundamental technical principles that define the language,
but more important (and more unifying I think) is the overall flavor
of the language - as exemplified by the ubiquitous invocation of
Clinger's now-famous quote. "Expressive Parsimony" might be another
way to put it. Depending on which use a programmer finds for the
language, there might well be different answers to constraints, but I
think that we all have a fundamental appreciation of the aesthetic -
or we'd be programming in CL.

In light of that aesthetic, I think that now, more than ever, a clear
separation of core and periphery is mandatory. There are exactly two
things that need doing: and extensible type system and some kind of
lightweight, compositional, modularization system. The fact that there
is no consensus on the latter only makes clear the fact that it is
hardly a *solved* problem yet (and my actual position on this is
long-publicized: lambda is the ultimate module system).

> R7 should try to avoid some of the processes and pitfalls that led to
> such an unfortunate and disruptive reception of R6.

Whoever of the nominees said that a clarity of goals was an important
missing feature, was also very much on track.

david rush
-- 
GPG Public key at http://cyber-rush.org/drr/gpg-public-key.txt

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

Reply via email to