Re: New procedure for proposed modules

[Vincent Untz <vuntz gnome org>]

Hi,

Sorry for the late answer.

Le lundi 03 avril 2006 �0:04 +0100, Andrew Sobala a �it :
> Vincent Untz wrote:
> > Hi all,
> >
> > The GNOME Release Team, after some discussion, has come up with a new
> > procedure for getting modules accepted into the next release of GNOME.
> > It is our hope that the new procedure will lead to less confusion. This
> > is more of a tweak than an overhaul, taking into account lessons learned
> > from the past.
> >   
> As I understand it, this implicitly makes module inclusion slightly 
> different for applications and libraries. A question to ask if I've 
> understood this correctly:

Right.

> If you're an application, then you get your module synchronised with the 
> release cycle by 2.15.1. If you're crackful, the community should voice 
> dissent straight away; if it doesn't, then you're accepted. By 2.15.4, 
> you can only be vetoed for technical reasons (ie. those in GEP 10).

Hrm. In the real world, I guess the answer could be "no". It's possible
to not be crackful, and have something we don't know how to handle or
that doesn't get integrated so well, or that doesn't work out really
well. So it can be vetoed for non-technical (or really,
not-that-technical) reasons after 2.15.1.

But ideally, yes, you'd be right :-)

> If you're a framework (library, "su"-framework, etc.) , then the above 
> applies, but you should get yourself used by other modules. By 2.15.4, 
> you could be vetoed for technical reasons (GEP 10), but in general if 
> you have widespread adoption by maintainers then that is a de-facto 
> acceptance by the community. On the other hand, if you haven't got 
> yourself used by other modules, it begins to look like the 
> service-as-implemented isn't really wanted in GNOME.

Having widespread adoption is not enough. libnotify is a good example:
it has a good adoption (or at least, it starts to have) but some of us
think we shouldn't include it since it should really be part of one of
our existing libraries (think Project Ridley).

> To summarise, the community makes an initial decision on applications 
> and libaries straight away, but libraries also have to prove themselves 
> as useful in-the-wild.
> 
> Is that correct, or have I got the wrong end of the stick?

The initial decision on applications is just what it is: initial. The
application still has to prove itself. However, this is far more
critical for a library since, in this case, there's a lot of work to get
it used by other modules.

One of the main points of this "new" procedure is that we know really
soon what is proposed and thus, the community can (and should) give
initial feedback soon. If the feedback is clearly negative, then it's
pretty clear. The initial feedback can then help the authors of the
proposed module to improve the module so that everyone agrees it should
be included.

I hope I'm not answering wrong. I'm pretty sure other people of the
release team will correct me if I do, though ;-)

Vincent

-- 
Les gens heureux ne sont pas press�