Martin Cooper wrote:
No, "prohibited" would probably be too strong (PROBABLY)... And yes,
I'd agree that if you know there are dozens of committers ready to
provide support, that's a bit of a different story too. But can you
really say such a discussion usually takes place before a vote? Is the
question: "are there at least a few people ready to support this?"
actually asked before a vote is called? That would be atypical in my
experience, based on the project I've been involved in.
For anyone paying attention to the project, the question shouldn't need to
be asked. It should be clear from the activity on the mailing lists. What do
you think, Frank? Are there people here that will support Struts 2.0.12 or
2.1.1? I think you know the answer to that, but can you point me to the
discussion thread where that was established?
Of course I know the answer to that, and the fact that there's no
discussion required supports my point. A +1 apparently means to enough
people that they intend to support a given release. Exactly what I've
been saying all along.
Assuming that's the case then, it's the *implication* of what a +1 means
that's important, which I believe was Ted's point.
If what's the case?
That a discussion of who will support the release doesn't typically
occur before a release.
What's important about a +1 vote from a PMC member is that it counts towards
the minimum three +1 votes required for the release to happen at all. The
_implication_ of what you are saying is that, without at least three PMC
members - not just committers or interested contributors, but PMC members -
who are ready to support the release, that release cannot happen at all.
I had to think about that one for a few minutes :) No, what I'm saying
is that NO ONE should be voting +1 for a release that they are not
themselves willing to support. Now, let me be clear: this in no way
ensures that everyone who votes +1 WILL support the release. Clearly
some won't, for one reason or another, and that's OK...
Let's try it another way... you earlier implied that there are plenty of
people willing to support 2.1.1, and I agree 100%. But how do we *know*
that? What if everyone who votes +1, in their own minds, thinks to
themselves "yeah, I'm voting +1, but TO ME that doesn't mean I'm going
to support it, and further I have no intention of supporting this
release"? That release is then left unsupported by those that voted to
release it, a situation that should never happen (and as you pointed
out, Apache members are typically responsible, so such a thought would
likely never cross their minds). That's also why there's X number of
votes required I think: if each one implies they intend to support the
release, it's likely that at least ONE person actually will, and the
release will be supported. Without that implicit meaning, it's a
crap-shoot. (And yes, it's also true that even with this implication in
place, a release could wind up going unsupported, but history has shown
that's just not likely).
Should we declare Struts 1 dead? Do we have three PMC members who are still
willing to support further releases of it?
That's a loaded question... do we have even three *PEOPLE* still willing
to support further releases of S1? :)
I left the part above where you said "...a +1 vote is *not* an assertion
that the voter, specifically, intends to provide such support". I would
contend just the opposite is in fact the case, but I'll now qualify it
slightly in light of your reply: in the absence of a discussion before a
vote where it is determined who will provide support other than the
person casting a +1, then that +1 does in fact *imply* that person
*specifically* intends to provide support. Stated another way: a person
voting +1 cannot *assume* there will be support provided by others, that
would potentially be a big disservice to the community at large when
they discover no one is in fact willing to support the release.
See above. I'd bet that, on at least one version of Struts in the past, you
have done exactly that - "*assume*d there will be support provided by
others".
You're right, and I'd even take it a step further: I have *always*
assumed there would be support provided by others because I've always
assumed a +1 vote by any responsible Apache member implies they intend
to support the release. Don't forget, I'm not a committer, I'm not an
Apache member in any way, so me casting a non-binding +1 vote means
squat other than "yeah, one extra set of eyes has looked at it and
thinks it looks good". Rest assured, if I was a member, my +1 vote
would *always* mean I'm willing to support the release, otherwise I'd
abstain.
The fact that projects at Apache do not typically push out release which
are not then supported pretty much supports this: I think most Apache
members voting +1 are not only saying "I believe the code is ready for
public consumption" but are also by implication saying "...and I'm ready
to back up that belief with support".
They probably are. But that's a long way from that being a requirement,
which is the point on which this conversation started.
Not to be Clintonesque or anything, but I think it depends on what your
definition of "requirement" is :) Should you not be able to cast a vote
until you've FAX'd in some notarized certification of your intent to
support a release? No, that'd be over the top. But, should it be a
safe assumption by all members of the community that your +1 means you
are willing to support the release? Yes, I believe so. Further, I
believe all the projects that have worked well to this point are
evidence that most people feel that way, whether it's ever been written
down somewhere or not.
Martin Cooper
Frank
--
Frank W. Zammetti
Author of "Practical Ajax Projects With Java Technology"
(2006, Apress, ISBN 1-59059-695-1)
and "JavaScript, DOM Scripting and Ajax Projects"
(2007, Apress, ISBN 1-59059-816-4)
and "Practical DWR 2 Projects"
(2008, Apress, ISBN 1-59059-941-1)
Java Web Parts - http://javawebparts.sourceforge.net
Supplying the wheel, so you don't have to reinvent it!
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
