Re: Bug#1014908: ITP: gender-guesser -- Guess the gender from first name
Steve McIntyre <steve@einval.com> writes:
> IMHO there are 2 points to an ITP:
> * to save effort in case two people might be working on the same
> package
> * to invite discussion on debian-devel / elsewhere
> If people post an ITP and upload iummediately, then I don't think that
> helps on either count.
> If the only reason for the ITP is to make lintian quiet then I think
> that's a total waste of time - it's following a guideline blindly
> without understanding the reason for it.
> How do others feel?
I feel the same way, although I have for a long time been dubious of the
benefit of debian-devel review for ITPs (and, to be honest, the benefit of
WNPP in general apart from orphaning, although sometimes ITP and RFP bugs
are a convenient central place to document all the reasons why packaging
some specific piece of software is really hard), so the whole system feels
kind of creaky to me.
ITPs do occasionally catch things that really shouldn't be packaged, and
we don't have another good mechanism for doing it. But the whole process
as we currently follow it feels oddly dated and manual and sometimes like
a box-ticking exercise. (It also adds a lot of noise to debian-devel from
the perspective of, I suspect, most participants. But we've talked about
that aspect of it before, and there was some moderate desire to see the
new packages flow by.)
Given that new packages as uploaded (a) include nearly all of the
information in an ITP in a more structured form, and (b) have to flow
through NEW anyway, I do sort of wonder if it would make sense to notify
some mailing list of every new source package, extracting similar fields
and the top entry of the changelog (which hopefully has some explanation
for why the package is being packaged for Debian, and we could encourage
people to do that), and then use the time the package sits waiting for NEW
review as the window for people to raise concerns.
That doesn't address the locking purpose of ITP (avoiding duplicate work).
I'm not sure how frequently ITPs are effective at doing that. It feels
like the percentage of the total software ecosystem that Debian is
packaging is smaller than it used to be (we've grown but free software has
grown way faster) and most of the places where I'd expect contention to
happen are handled by language packaging teams that probably have their
own processes.
--
Russ Allbery (rra@debian.org) <https://www.eyrie.org/~eagle/>
Reply to: