Re: Vote on Windows installer links

Simon,

> The popularity of the motion is precisely the thing we wish to vote on.

The popularity is completely irrelevant. Bruce may have suggested that
it was, but if so he's wrong.

Parallel query is extremely popular. Built-in replication is extremely
popular. Infinitely scalable replication is extremely popular.
However, popularity isn't going to get any of those things built, or for
that matter make the last one possible.

So unless someone organizes an installer team which isn't funded by
EnterpriseDB, then the far more appropriate path is to call up Larry
Allston at EDB and work something out. Please take note that the MSI
installer was dropped *precisely* because of lack of maintainers, so
we're not exactly hip-deep in Windows packagers.

> I requested we find evidence as a formal poll for the reasons you
> suggest, and others. I would still be happier if this was a formal poll
> rather than as an email from me, but if you would like me to raise this
> on other lists, please confirm that.

Not really. Again, without an alternate installer team, there's no
point. It doesn't matter how much anyone wants the moon in a box.

> The vote was requested on the advocacy list (which is where concerns
> were originally raised by Andreas) so that I am not accused of
> grandstanding. ISTM a valid list for this vote.

I agree that this is the right list to discuss this concern. I just
repeat that there is no point in voting.

> Chicken and egg. It was suggested that I was the only person that held
> the view as stated in the motion and was challenged to prove otherwise.

I think that Bruce wasn't being helpful and has contributed to this
thread spinning out of control. You've established that several other
people think that the current installer setup is not ideal, which is
sufficient for discussion of alternative methods of producing an
installer. Beyond this, the only useful further discussion would
involve a discussion of how to produce an alternate installer.

> You are the only person to mention that the ballot seems unclear, out of
> more than 20 people responding so far, so there is no evidence to
> support your view. If you choose not to vote, that is up to you.

Just because people vote on something doesn't mean that their +1 means
what you think it means.

In any case, we have absolutely no ability to take a deciding vote on
any mailing list, except may be core, because nowhere does PostgreSQL
have a defined electorate. We operate by discussion, clack,
volunteerism and consensus. The most you can have is a straw poll.

--
Josh Berkus
PostgreSQL Experts Inc.
www.pgexperts.com