[Framework-Team] Re: Updated PLIP review deadline

Martin Aspeli optilude at gmx.net
Thu Feb 14 20:44:12 UTC 2008


 > >  > That doesn't make sense, does it?
 > >  >
 > >  > If you give final votes/recommendations on Monday the 18th of
 > >  > February, and then you want to merge and freeze by Friday the 22nd,
 > >  > that leaves ... four days for people to react to your 
recommendations?
 > >  > That's insane!
 > >
 > >
 > > well, the reviews (incl their corresponding notes) will have to be
 > >  submitted by saturday, and the authors will also be notified via
 > >  trac.  that would give an additional weekend to react, well perhaps
 > >  just sunday.  i agree that this is still leaving very little time to
 > >  respond, but that was also the same with the original schedule we all
 > >  agreed upon.  and besides, afaik all but one PLIPs have been reviewed
 > >  with notes posted, so there was some additional time available.  and
 > >  yes, in theory new issues could be brought up by reading the notes and
 > >  thinking of something the first two reviewers have missed, but i think
 > >  the chances here are pretty slim.  imho they're so slim, that adding
 > >  another week before starting the merge doesn't justify them.

I still don't think this makes any sense. If you want to have
everything *merged* by Friday the 22nd, then you (Wichert) will have
to start doing those merges way before then. We'll all bed rushing
around like mad, things will drop out for no good reason and bugs will
creep in. This is precisely what we tried to avoid with a "small"
release. You can't decrease the scope by 50% and decrease the time
available by 90% though.

For the record, no-one's told me anything about my five
bundles.There's been no centralised, proper communication with PLIP
authors. Unless they are vigilant and watch Subversion and Trac for
notes the trickle in at a random intervals, they are unlikely to have
realised if they need to react.

The Framework Team needs to bear some responsibility for the quality
of the release, and that includes setting realistic expectations of
PLIP authors. It also includes giving clear communication about what
those expecations are, and taking into account that people will be
doing this work in their spare time, with other time pressures.

For example, if you'd met you original deadline, I would've had more
time, since that deadline was set at a more advantageous date. Right
now, I'm exhausted and fed up.

It leaves a really bad taste in my mouth that we have these
discussions at nearly every single deadline we set. What if I'm busy
this Sunday - potentially the *one* non-work day you give authors to
react to the final reviews? It's the first weekend since the PSPS. The
deadline recently slipped with lacklustre communication as to why,
when the new deadline was, whether it was being adhered to and what
would happen next. You are expecting PLIP implementers to be just
sitting by ready to jump when you manage to get the reviews done.
That's in no way fair.

I have minor changes to make to my PLIPs, as would be expected. If I
want to have a social life this weekend, I'll have to drop two or
three PLIPs out of Plone 3.1, PLIPs that I put a lot of work into
already and that people are likely to need. I am pretty sure I'm not
the only one who hasn't been planning my life around a slipped review
deadline.

 > >  > You need to at least give a weekend in-between. I thought that there
 > >  > were at least two weeks beteween final votes/recommended changes and
 > >  > final merges.
 > >
 > >
 > > the 22nd is "just" the date of the alpha freeze, the first rc gets
 > >  frozen on the 29th.  that should leave enough time to work things out,
 > >  esp since i think it's fine to still put in minor (and very serious)
 > >  fixes between the two.  plus like already said, it's the same as the
 > >  original schedule (i.e. long-known ;)) except shifted two weeks.

Seven days from alphas to rc? You've got to be kidding me ... We've
never, ever managed anything even remotely like that. If you do things
in a rush, quality suffers. And for what? A made-up deadline that
wasn't thought through properly? We can slip one week now and give
people a fighting chance to get things right, or we can slip three
weeks later as our RCs suck and throw up bugs, or we can be right on
target with a buggy release.

 > >  > I don't think you can realistically freeze the alpha before 
March 3rd
 > >  > if you want all the packages in for the freeze.
 > >
 > >
 > > i'm not sure, but i guess that's really wichert's call anyway.

*shrug* .. I've got half a mind to just not care anymore. What will
you do if you reach the 22nd and you have two or three bundles merged?

Martin





More information about the Framework-Team mailing list