On 2-Nov-08, at 12:40 PM, Mitchell Wand wrote:

> I would very much like to hear whether any others feel that this is  
> a perceived advantage for supporters of the language described by  
> the R6RS.

I agree with Will.  I know some "real Schemers" who did not
participate in the R6RS process because they did not want to "spoil
the party" for those pushing for R6RS.  My feeling is that the R6RS
electorate contains proportionately more pro R6RS people than the
general Scheme community.  So the process becomes biased if you make
it easier for them to vote.  Moreover people who did not vote on R6RS
may feel excluded from the community you want to poll because
automatically including the R6RS electorate gives the appearance that
this community is the one interested in R6RS.  Lets aim to be
inclusive!

I also like Will's suggestion to allow people to register and to vote
at the same time (on one form).  This will simplify the voting
process, for the voters at least, and hopefully attract a broader
participation.

Beside these changes to the voting process, I would like to see some
changes to the organization of the Scheme standardization committees
and their duties.

The RnRS standards with n<=5 evolved from committee work under the
unanimous consent rule.  Even though there were many RnRS authors, the
language avoided acquiring the inconsistencies that are typical of
"committee designed languages" (for more on this see Paul Graham's
essay "Five Questions about Language Design"
http://www.paulgraham.com/langdes.html).  These previous Scheme
standards were true to Scheme's principal design philosophy that
"Programming languages should be designed not by piling feature on top
of feature, ...".

The R6RS was designed by a smaller committee yet it suffers from
several inconsistencies and unwarranted bloat.  To avoid this in the
next design iteration I believe the committee members must agree from
the start on the design philosophy and the nature of changes which are
sought.  There may be more than one way to achieve this.  I propose
that the steering committee should take on a more active role in the
design process and ensure that the fundamental features of Scheme, if
you wish its soul, is preserved.  I see the editors committee as
implementors of the design goals and constraints set forth by the
steering committee.  There must be frequent and regular interaction
between the steering committee and the editors committee.

As I understand it such a change in the committees' organization is
possible since once it is elected the steering committee can change
the charter.  However it is not clear to me that the current steering
committee election process is appropriate because it does not include
any measures to promote the election of a committee that works well
together.  The single transferable vote mechanism, with which I have
no experience, appears to promote proportional representation.  This
is not the most important property to optimize.  A steering committee
with divergent opinions would be a disaster, yet nothing in the
election process guards against it.

For this reason I believe that the electorate should not vote for
individual members of the steering committee.  Instead, nominees
should be given some time to get to know the other nominees and to
form coalitions with the nominees they would like to work with.  Once
this is done the electorate can place their vote for one of the
coalitions, and the one with the most votes gets elected.  This
procedure will yield a much more effective committee.  Moreover the
electorate will have a much clearer idea of what their vote
represents.  It also reduces the likelihood that members of the
committee will resign due to friction with the other members.

Marc


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

Reply via email to