Per Bothner scripsit:

> I'm reasonably content with how R7RS-small turned out.  It's a good
> size, and a sane/coherent updating of R5RS and (a trimming-down of)
> R6RS, though of course we can quibble about various issues.

WG1 thanks you, Alex thanks you, and I thank you.  I see Kawa is getting
close to R7RS-small; is there a list somewhere of things not yet implemented?

> There is something to be said for either a "benevolent dictator" or
> a small committee of experts.  Voting on features by the community
> at large (i.e. anyone who cares to vote) does not seem the way to
> design R7RS-large.  I think it's fine to do a poll to get a sense of
> the community, but it should never be deciding.

And yet that is how WG1 worked, very nearly.  Everyone who wanted to
join was admitted without exception, mostly at the beginning but one
or two later on.  Two people were removed by the Steering Committee in
the early stages.  A few more left either openly or silently (by never
voting at all).  As time went on, fewer and fewer voted: up to 16 on the
first ballot, 7 on the last few.  Whether those people were "experts"
is a question.

Because of the declining vote situation, and because the large langauge
will naturally include many parts in which different people are involved,
it makes sense to me to have a rolling membership, with people allowed
to join at any point, when the matter being discussed is of interest
or concern to them, and (effectively) drop out at any point.  That does
amount, at this early stage, to "anyone who cares to vote".

I do have the right by charter to remove members/voters who I think are
problems: who don't know Scheme at all, or who appear to be sock puppets,
for example  (Anyone I remove can appeal to the Steering Committee.)
I hope never to have to use that power.  Overall, I think a balance
between users and implementers is a Good Thing, though in the nature of
things there are more users than implementers.

> Perhaps we should aim for a more modest r7.1rs with a few optional
> additions.  Perhaps every other year we could have a new 7.x point
> release with some modules we can take more time to get consensus for.
> Instead focusing on features perhaps it is more important to find a
> standards language and framework for optional features and modules.

Right now I'm just balloting a few points that are optional in R7RS-small
and either mandatory or optional in R6RS: the numeric tower, Unicode,
IEEE floats.  That's because I don't have time/energy right now to work
on SRFI 113.  I hope the logjam will break soon.

The bulk of WG2 work will consist of deciding which SRFIs to add to
the large language either as options or as mandatory parts.  I expect
that these libraries will be mostly optional.  Some will be mandatory
either because many other libraries depend on them, or because a modern
application developer would just naturally expect to find them in any
language and they can't be written in R7RS-small.  (Networking is the
obvious example.)  Of course, new SRFIs will be grown concurrently with
WG2, and they will go through the regular process and remain SRFIs for
use outside R7RS-large as well.

Consequently, I plan not point releases but rolling editions.
The first or Red Edition of R7RS-large will contain various modules
that have existed as SRFIs for a long time, are widely implemented,
and are uncontroversial.  Candidates for the Red Edition appear at
<http://trac.sacrideo.us/wg/wiki/ConsentDocket>.  Then without
revisiting any of that, we'll move on to the Orange Edition, adding
yet more modules from SRFIs and from R6RS.  Eventually we'll reach the
Ultraviolet Edition (so called because it is currently not visible) and
declare the process complete.

--
John Cowan          http://www.ccil.org/~cowan        cowan@ccil.org
He made the Legislature meet at one-horse tank-towns out in the alfalfa
belt, so that hardly nobody could get there and most of the leaders
would stay home and let him go to work and do things as he pleased.
    --H.L. Mencken's translation of the Declaration of Independence

_______________________________________________
Scheme-reports mailing list
Scheme-reports@scheme-reports.org
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports