[Axiom-developer] Re: votes

[Martin Rubey]

address@hidden writes:

> There are 97 people subscribed to the Axiom mailing list.
> Please explain how you plan to garner 49 votes for anything.

It is quite OK to abstain.  I'd suggest to

* have no vote for trivial patches (fix typos, for example) or if nobody
  disagrees.

* have a short voting period for small and medium size patches.  Say, 24 or
  maybe 48 hours for the patch concerning the Taylor series powering.

* have longer voting period for large and huge patches. Say one week, or maybe
  10 days for voting on changing the build system to autotools, or to switch
  from autoconf to asdf.

A vote is triggered by anybody on axiom-developer saying: stop, I'd like to
vote.

If somebody says that a patch breaks this or that, this would be discussed
first.  Sometimes it will be possible to agree whether this or that was broken
before the patch and the patch only surfaced it, or whether this or that was
ok, but the patch breaks it.  Only if such an agreement is not found, a vote is
necessary.

Consensus remains most important, i.e., if a vote is not absolutely clear, it
may be better simply to start a new branch and keep it in sync.  subversion or
svk should make that quite easy.  (In any case, I'd like to reduce branching
and merging to an entirely technical, not emotional issue.)

Three remarks:

1) thanks for asking for clarification
2) interesting to know that there are already 97 subscribers.  I would have
   guessed fifty.
3) I'm open to discuss any other democratic system.  However, I would like to
   keep refereeing and voting two distinct processes.  Peer review is a great
   thing, but it's not a substitute for democracy.


Martin