On Mon, 9 Oct 2000, David Grove wrote:

> On Monday, October 09, 2000 7:12 PM, Nathan Torkington [SMTP:[EMAIL PROTECTED]] 
> wrote:

> How about an open, crossplatform mailing list for issues, with a
> mechanism on perl.org for public voting on larger issues.

In a volunteer organization, you can't reliably translate votes on issues
into action.

To be concrete, consider a perl5 example.  Suppose that there was a vote
to change from metaconfig to autoconf for perl5.  Suppose further that the
current Configure maintainer didn't know anything about autoconf and
judged himself unable to complete the task in any timely manner.  What
would happen next?  Probably nothing, unless an autoconf champion steped
forward and volunteered to take over configuration issues.  But in that
case, what was the point of the vote?  At any time, _anyone_ can step
forward and champion a cause.  If he or she makes a good case and
implements it well[*], then it will probably be adopted.

Similar concerns hold for other issues.  Without a champion to implement
the changes, nothing is likely to happen.  With a champion, no vote is
needed--the champion can just go ahead and try to do it.

There are, of course, cases where things are not clear cut.  Suppose, for
example, that someone proposes and implements a feature (e.g. safe
signals) but it has a drawback (e.g. 10% slowdown).  Should that feature
be included in perl?  That's a tough question involving difficult
tradeoffs.  It's not at all obvious to me that voting is a better way to
resolve such issues than the current process of allowing the pumpkin
holder to decide.

[*]Of course "implements it well" is a subjective criterion, but one
important aspect of it ought to be a reasonable plan for the long term
maintenance of the feature (usually the champion saying "I'll do it.")

    Andy Dougherty              [EMAIL PROTECTED]
    Dept. of Physics
    Lafayette College, Easton PA 18042

Reply via email to