gdb-discuss
[Top][All Lists]
Advanced

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

Re: [Gdbheads] Let's resolve this quickly


From: Jim Blandy
Subject: Re: [Gdbheads] Let's resolve this quickly
Date: 26 Mar 2004 14:55:11 -0500
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3

"Eli Zaretskii" <address@hidden> writes:
> Wake up, the Committee, whoever you are!  Open your eyes, the FSF and
> the GNU Project!  Whatever alternative of the suggested solutions you
> like, let's implement it, but by God, let's do that quickly!  Let's do
> it today and then let's try to put all this behind us and get back to
> our business of maintaining GDB.


I'm less experienced than Eli, but as I see it, one stumbling block is
that the committee has no agreed-upon procedure for making decisions.
I've been trying to get GCC's procedure, so we could start with
something that actually has been observed to work, but as I said, it
doesn't look like they have anything written down.  Maybe we can agree
on something together quickly, and revise it as needed.

Here's a proposal:

    A vote begins with a proposal posted to address@hidden, whose
    subject begins with "call for vote", and containing a proposal for
    people to support or disapprove of.  A committee member votes by
    replying to that message saying "I vote yes." or "I vote no.".
    The vote continues for at least a week (40 business hours), or
    until at least half the committee has voted one way or the other.
    If half the committee has still not voted on a proposal after four
    weeks (160 business hours), the proposal is dead.

Here's the thought behind the proposal:

This doesn't talk about the discussion of a vote.  I don't think we
need any procedure about that: anyone can post to address@hidden
about anything they like, and discuss it as they please.  People can
even reply to "call for vote" messages with non-vote messages --- say,
suggesting revisions of the proposal.  Committee members can prolong
the discussion period by simply abstaining from the vote.  If half the
committee members feel the issue is never sufficiently discussed, then
the proposal dies.

This rule has a quorum: one half the committee.  Should that be
higher?  Lower?

Votes are public.  Should they be private?  I think they should be
public; the decision is being made with the public's trust.

Votes must be explicit and unconditional.  It must not require
judgement or interpretation to count the vote.  Other messages are
commentary --- which is fine --- but not votes.  Votes may contain
explanations or other comments.

If there's general agreement among the committee members that the vote
shouldn't proceed immediately (perhaps someone who will have valuable
things to say is on vacation), then committee members can simply put
off voting on the proposal until it's appropriate to proceed.  If they
want to get the "one week" rule clock started again, they can simply
re-post the proposal; people will reply to the new thread, if they
agree it was reasonable to do so.  A vote message applies to the
speciifc proposal message it's a reply to.

This rule doesn't address the question of how to choose between more
than two proposals.  Voting rules for such choices are more complex,
and although there are plenty of useable sets of rules, none of them
are entirely satisfactory (Google for "Arrow's Theorem").  In our
situation, I hope we can ignore that problem for now.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]