[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Scheme-reports] Comments on revised WG1 charter

In what follows, I am speaking as an individual.  My views
may not coincide with those of the Scheme Language Steering
Committee as a whole.

Compatibility between the work products of WG1 and WG2 could
be weakened from "must" to "should", but those who advocate
that weakening should take the following into account.

The word "must" constrains WG2 as much as WG1.  In practice,
it would probably constrain WG2 more than WG1, because WG1
is supposed to finish its work earlier than WG2, which would
give WG1 whatever power might accrue from being the first to
finish its semantics.  It is of course possible that WG1
might fail to meet its schedule while WG2 meets its schedule,
which would then put WG1 in the position of negotiating or
conforming with a concrete proposal from WG2 instead of the
other way around.  It seems to me that one solution to that
potential problem is to develop the sense of urgency and
compromise needed to stay on schedule.

Note also the section on "Coordination with Working Group N"
and the provision for having some people belong to both
working groups.  I believe those provisions reflect the
importance of compatibility between the work products
produced by the two working groups, at least as judged by
the Steering Committee.  I also expect the chairs of the
two working groups will work together to achieve the goal
of compatibility.

I would also point out that, even if "must" were changed
to "should", the Steering Committee might not endorse work
products (by either working group) that place unnecessary
obstacles in the path of the other working group.  Had the
Steering Committee not believed compatibility between the
work products to be delivered by WG1 and WG2 is technically
feasible, it probably wouldn't have proposed draft charters
that use the word "must" when mandating that compatibility.

It seems to me that the perceived impossibility of achieving
compatibility between the work products to be delivered by
WG1 and WG2 derives from several questionable assumptions
about the work products to be delivered by WG1.  It also
seems to me that some of those assumptions involve changes
to the semantics of IEEE/ANSI/R5RS Scheme for which there
may be little actual support within the Scheme community.
If my impressions are correct, then leaving the word "must"
in place might help to avoid fruitless attempts to use the
WG1 process to mandate problematic changes to the semantics
of Scheme.

(speaking for himself)

Scheme-reports mailing list