Re: [HACKERS] Happy column dropping

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
>> I certainly don't want to discourage Peter, either, and perhaps
>> was a bit too harsh. But release of a feature this half-baked
>> would fit the stereotype many people have held towards free,
>> open source software, and postgres in particular.

> Who said something of a release?

Um, you do recall that we are one week from feature-freeze for 7.0 beta,
don't you? It's mighty late in the cycle to be committing code that you
are not expecting to release in more or less its current form. The
reason the howls have been so loud is that because of the calendar,
everyone is assuming that you intend to release this code more or less
as it stands.

If that was *not* your intent, perhaps you had better pull the code out
until after we fork the tree for 7.1 development.

> Whatever happened to release early, release often?

The Postgres project has generally adopted a more conservative approach
to releases, because we know that people entrust critical data to
DBMSes. We don't have anything that corresponds to a development
release series; *all* our releases are supposed to be "stable releases".
Of course we don't always get there, but that's the idea.

Jan has been muttering that we ought to have some means of dealing with
code development that spans multiple release cycles, ie, CVS branches
for work that is not expected to be part of the very next release.
I've found it hard enough to keep track of tip vs. last release branch,
but maybe something like that is needed. It would let people push code
out for review without implying that they think it's good enough to go
into the next release.

regards, tom lane