Re: Postgresql Materialized views

Sean Utt wrote:
> My point is simply this: The lack of a clear formal process for
> feature requests leads to this degradation in the conversation.
> Without a formalized structure, the conversation devolves rapidly into
> an argument over semantics and word choice. It is not my contention
> that the "core" developers need to be different in any way. It is also
> not my contention that the users need to be different in any way. It
> is my contention that the "process" currently generates more ill will
> than it needs to, and needs to be replaced. The problem is a systemic
> one. There needs to be a more formal structure put in place than just
> the -hackers mailing list. There needs to be a way to evaluate the
> demand for a specific feature as well as the benefits and the effort
> it will require. It needs to be done in as neutral a way as possible.
> In order to be effective, it will have to be driven into being by the
> developers, because they will be the ones who can hamstring it -- not
> the users.
What sort of structure are you envisioning? Features do make it into
PostgreSQL - PostgreSQL has had, and arguable still has a more complete
feature set than well funded alternatives such as MySQL. Their is a TODO
list that both grows and shrinks with each release. I have myself seen
major changes in 8.0, 8.1, 8.2, and 8.3 that have personally benefited
me. What problem needs to be solved?

I agree with your sentiment. For somebody who wants to request a
feature, expect it to be taken up by a champion (other than them), and
monitor it's regular process, no clear infrastructure seems to exist.
However, I question the relevance or value of such a system. My own
initial contribution was a reaction to the notion that somebody should
be able to demand other people to volunteer their time to work on
something. It may have been unfair and cold to the original poster, and
for this, I apologize. I don't believe the user community should
necessary be able to demand or prioritize a feature unless they are
willing to put up resources to support the effort. Resources usually
means either people or money. If people truly have a strong business
case for a feature, there are several qualified companies willing to
take their money and turn it into something real. If people are not able
to produce such a business case or justify the expenditure of funds,
attracting volunteers to do the same work requires a very different
approach. It requires zeal, compelling argument, and a reasonable amount
of ego stroking or challenge. Putting in place an official process can
have the opposite effect. People such as myself are very willing to
volunteer efforts in an informal manner without formal deadlines or
processes, because we enjoy it. Forcing the regular amount of red tape
many of us need to cut just to get our jobs done at work is not very
motivational for people such as myself. Now, while I have contributed to
other open source projects, I have not personally contributed much to
PostgreSQL. The core PostgreSQL contributors would have to make their
voice heard. I think, though, that telling them that they must work on a
certain feature, because that's what the users are asking for, is the
wrong approach. Not to say that is exactly what you are requesting, but
I suggest that is where you are leading.

Cheers,
mark

--
Mark Mielke <mark(at)mielke(dot)cc>