geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aaron Mulder" <ammul...@alumni.princeton.edu>
Subject Re: Frustrations of a Release Manager
Date Fri, 09 Jun 2006 00:44:34 GMT
Boy, you _are_ a little frustrated.

I think it will help to have the schedule of the release.  No one can
claim IBM has a secret agenda if the time line is laid out there.  And
it's easy to wink if no one has any idea what the deadlines we're
working toward are.

I propose for 1.2 we drive it more by time than by features.  That is,
we lay out a schedule including builds every 2-3 weeks, initially
milestone builds, becoming beta and RC builds.  We try to get people
to test and provide feedback on the builds as we go, and we expect
that we'll have some issues early to mid-way through the process but
have clear dates for feature freeze, no bugs fixes except blockers,
branch for the next release, etc.  We may have to adjust the schedule
depending on what develops, but at least we'll know what we're
targeting at all times.  Then we'll have to huddle after the 1.2
release and decide whether this was an improvement over the 1.1
process or not, and decide how to go for 1.3/2.0.

I ended up using the SuSE 10.1 betas and RC's during their
dev/test/release process, and it went very much like the above
(including at least one adjustment to the schedule in mid-stream).
But it was nice to have a schedule laid out, to always be able to see
the date the next test build was expected, etc.  I'd like to give it a
try.

We may have to play things a little by ear in deciding how to deal
with plugins and which releases we try to create in-place upgrades for
(vs. fresh install only), but I think that's all manageable.

Thanks,
    Aaron

P.S. Maybe we'll get to see if the "Chariot agenda" is any better than
the "IBM agenda."  :)

On 6/8/06, Matt Hogstrom <matt@hogstrom.org> wrote:
> I previously sent out an e-mail regarding the freezing of the branches/1.1 code so we
can march to a
> release.   I expect you could tell by the tone of the e-mail that I was very frustrated.
 My
> frustration arose from a private e-mail asking if IBM had some secret agenda and if that
was the
> reason I was trying to move the release out.  This really set me off.
>
> You will have to be the judge as to whether IBM yields any undue influence on the project;
I believe
> the answer is no.  From my perspective, IBM goes out of its way to make sure we do not
"force" the
> community.  We carefully consider how we interact in the community.  not only about how
we conduct
> ourselves but what the appearance of our actions might be perceived.  Yes, we base a
server offering
> on Geronimo and we do have a vested interest in seeing the project succeed; however,
so do a number
> of other companies.  It is always a challenge to balance contribution and influence to
ensure your
> growing the ecosystem and I think we do a damn good job of it.  IMHO we actually are
too conservative.
>
> The reason I'm trying to move the release along is because it has been OVER FIVE MONTHS
since we
> have given our users anything else to look at.  Does anyone remember who they are?  These
are the
> developers who we're trying to create something for.  Developers that will be interested
in using
> our project.  I don't know about you but in this fast paced world of development people
don't sit
> around waiting for FIVE MONTHS for anything.  They will choose something else and then
you've lost
> them.  You've broken their confidence on your ability to deliver and consequently they'll
be less
> likely to believe you'll deliver when you say you will.
>
> I am not pushing the release because of some secret IBM agenda.  I'm embarrassed that
WE can't seem
> to deliver something.  Originally we said end of January and then discovered that we
had some
> refactoring to do.  It will only take a few weeks we thought.  Two months later (and
stalled
> development on new innovation) we set a target date of April 28th.  Yes, I chose the
date but it was
> four weeks from the day I suggested it; we seemed to have consensus.  Unfortunately we
found that
> our changes (they were the right thing to do) caused us lots of heart burn in CTS testing.
 Few
> people were able to help with that for whatever reason so it was a long slow grind. 
We burned up
> April and then started into May.  Java One was in there for a week so we basically lost
two more weeks.
>
> Continuing to try and get a release out we've diagnosed performance problems, survived
an SVN outage
> at Apache, a Codehaus outage, I've written numerous e-mails about getting 1.1 out and
yet we are
> still not in lock step as to what we're trying to accomplish.
>
> In short, yes, I am frustrated not because some secret IBM plan is not coming to fruition
but
> because the community, of which I am apart, is dysfunctional to the point of laughability.
 We are
> now under Review and Commit.  We're not doing well there either but that is likely a
separate thread
> for discussion.  I know we're a volunteer organization but I hope that being part of
a project makes
> us a team.
>
> If we cannot begin to operate like a team its going to be a slow painful process.
>
> So, to answer the question about why I'm pushing to get a release out, there is no secret
IBM plot.
>   I simply want to get this release out to break the log jam on development.  In one
e-mail thread
> it was posed that new features were being added to a frozen branch and that was met with
a wink.
> One person simply said maybe we need balance, will you help me on the next release? 
Wow, what about
> this one.  Are we a team or a bunch of people that act as an Army of One.
>
> I'll post another note with the schedule of the release.
>
> Yes, I'm frustrated.  I know others are as well.
>
> Matt
>

Mime
View raw message