Re: Stricter policy when including BIG modules

[Danilo Segan <danilo gnome org>]

Murray Cumming <murrayc murrayc com> writes:
>
> We have very few must-haves for the release schedule, because must-haves
> are an obstacle to time-based releases. One of those must-haves is "an
> acceptable level of stability", for instance. I don't think that "being
> translated into languages X, Y, and Z" should be one of those
> must-haves.

No, I never said that.  Evolution is fairly well supported by the
translators, but this support may have come at the price of
translations of regular desktop in the 2.6 timeframe, because of late
withdrawal.

>
> If we required 100% certainty then we would never do anything. The
> schedule manages the risk and copes with what happens.
>  

Well of course, I never asked for such strictness.  This is supposed
to be about VERY (and I mean VERY) significant modules in terms of
work.  Translating complete Evolution probably requires 100 manhours
(my rough estimate is, therefore, two weeks of full time work).  No
other module in entire Gnome CVS is as big as it is, except for Gimp
and Gnumeric.  So, the policy I suggest would only apply to Evolution,
Gnumeric and Gimp, if they were ever proposed into Gnome. (Of course,
there will probably be more modules on par with them in the future,
but with over 150 modules that are translators currently concerned
about, this is still a minority.)

> I don't think that you have actually suggested any particular action or
> process. Feel free to describe it in 1 or 2 official-sounding sentences
> so we can see what it feels like.

Ok, let me try again.

"If a module proposed for inclusion in Gnome requires substantial
effort by many subteams of Gnome development (such as translation,
documentation, UI review teams), and there's a real risk of that
module being withdrawn after the final module list is announced, then
don't include that module in the release."

I don't know how "official-sounding" this sentense is, since that's
not my field of practice, but I hope I at least got the point across,
and someone will come up with something better.

And now some interpretation:
I admit all of the adjectives here ("substantial (effort)", "real
(risk)") are pretty vague, but in case of first, it is easily
estimated by the teams in question (ask any translator, and they'll
gladly tell you that translating Evolution is a substantial effort),
and in the case of latter, I guess having UI redesigned only after
Gnome 2.4 was released, and a big number of substantial changes to
come in 4 months before feature freeze are enough of "real risk".
So, the criteria would get down to a simple: is this "major impact"
module already stable enough at the time of final module list
decision?  

Yes, this big module would have 1 month less than the other components
to get to a feature-full, stable state, but that's the price of it
being so huge.

And this is supposed to concern only the inclusion of *new* modules.


Hope I'm a bit clearer this time.

Cheers,
Danilo