On 3/9/06, Ted Husted <[EMAIL PROTECTED]> wrote:
>
> On 3/9/06, Martin Cooper <[EMAIL PROTECTED]> wrote:
> > I do agree that we should make sure people understand that it's not a
> > release, but I don't think we need to assume that the user@ audience is
> too
> > dumb to recognise the distinction between a Test Build and a Release.
>
> I wouldn't use the word "dumb", but I do think it is cavalier for us
> to assume that most users are aware of the distinction between a "test
> build"  and a "beta release". The term "test build" is something we
> made up, and it's doubtful that many other people actually understands
> what it means.


Why are you bringing this up now? This is a rehash of a discussion we had
years ago. The first announcement of a Test Build was sent to the user list
in February 2004, over two years ago. I'm not aware of problems over the
last two years that would lead us to need to rehash it all over again now.

Whether a build is a "release" is an important distinction to an
> Apache project, and I would not want to make a practice of doing
> anything that might dilute the concept of an Apache Release. We
> understand that a test-build is not a release, but people seeing an
> announcement on user@ might not share that understanding.


We've been doing it for two years. It's not like it's a new concept. And our
messages have been very explicit about a Test Build not being a Release.

If people are not ready to vote a test-build to GA, then there is no
> harm in voting it beta first and announcing it. (Especially since we
> aren't voting on the plans.) We can vote on a build as many times as
> we like, to either upgrade or downgrade the quality.


I think you're missing my point here. I'm not talking about people on dev
not ready to vote to GA, but rather that we would be limiting the feedback
we gain on the test build before we have to vote on it. IMHO, the more
opportunity for feedback _before_ the vote, the better off we are.

If we are going to announce unsanctioned builds to user@, then,
> please, let's at least call them *ALPHA* builds, so that these
> unilateral "trial balloons" are not confused with an actual beta
> release, approved by the PMC.


This would be much worse, in my opinion. Trying to explain to people that an
Alpha build is not a release but a Beta build is a release is going to be
*much* more confusing to people than explaining that a Test Build is not a
release. To have Alpha and Beta designate builds of different legal status
would be very bad.

And besides, as Wendy pointed out, an Alpha _is_ a release, and would
therefore need a vote from the PMC. The whole point of Test Builds is so
that we can push a build out for experimentation without making any up-front
declaration as to its quality. It also empowers all of the Struts committers
to lead us down the "release early, release often" path, which is A Good
Thing.

The scheme we've had in place for the last two years is working just fine.
If it ain't broke, don't fix it. And IMO this is one thing that just ain't
broke.

--
Martin Cooper


(And, here, I'm just speaking in the general case, not to any specific
> build.)
>
> -Ted.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>

Reply via email to