ftpmasters' job and the DPL (was: Question for candidate Towns)
Anthony Towns <aj@azure.humbug.org.au> wrote:
> Kalle Kivimaa wrote:
>> anyway: what do you think the NEW issue is an example of?
>
> Not having enough time in the day.
>
> The resolutions to that are:
>
> (a) reprioritising things
> (b) making more time available
> (c) making things take less time
> (d) training other people
>
> I don't think (a)'s plausible in the case of NEW; there don't seem any
> obviously unimportant things that're being mistaken as being important
> and sucking up too much time. (b) is generally tricky, and is mostly a
> motivational issue, and thus generally not a short term one. (c) and
> (d) are both ideal, but also involve taking time out now for a gain in
> future (ie, scripting or rethinking things to work out how they can be
> done better, or finding and mentoring newbies to help with the tasks)
> -- and when the problem is a lack of time right now, they're hard to
> implement.
That sounds plausible.
> Most of ftpmaster are at the release cabal meeting Steve announced on
> -project yesterday; I expect the current ftpmaster issues'll be
> resolved before we're done on Monday anyway. I would have liked for
> them to have been resolved already, but I find it very hard to push
> for things like that when messages like Martin's to -project [0] are
> more or less left to stand.
I don't like the tone used in that thread, and I don't want to repeat it
here. But some of the concerns raised seem valid to me. Let me give
one example, an unimportant one in hope not to start a flamewar, and
because I happen to be involved. At the end, I'm going to ask some
questions to you as a DPL candidate; the others can of course also
answer if they want to.
- Early in January this year, I uploaded a version of tetex-bin that
produced a binary package for the kpathsea library with a new soname,
and a new package name. Shortly after this, there was a discussion on
-devel about exactly such things (in which you, Anthony Towns, were
involved), and as a followup I wrote a mail to ftpmaster, saying you
should reject that package because I did the naming wrong.
- After some more thinking and reading of the thread, I changed my
mind. I wrote again to ftpmaster, saying this, and that my case was
probably one of the "special" ones were the naming I chose was
acceptable.
- However, I never received an answer to that second mail. Since it
wasn't important for the release or any bugfix to get that package in
soon, I didn't complain; but in fact I was kind of puzzled to not get
any answer. I was unsure whether I should proceed uploading packages
with the naming I thought correct, or finally make the change to the
other scheme you said (on -devel) was "usually" correct. I also
didn't know whether the reason why the package was not processed had
anything to do with the fact that this issue might be controversial
(rumor goes that "easy" packages are processed first).
If the issue would have been time-critical, I would not have known
whether asking again was a good idea, or would actually annoy the
ftpmasters, and slow down processing.
As a DPL candidate, do you feel that such a feeling of uncertainty, and
of being at the mercy of somebody you can't talk to, occurs frequently
among developers, and if yes, do you think it is a problem for their
motivation to do Debian work?
Do you think that the uncertainty about the criteria for accepting or
rejecting a new package can make developers to hang on to badly designed
old solutions, instead of proposing and implementing/packaging something
really new?
If the answer to any of these is yes, do you think the DPL can do
anything about it? What?
Regards, Frank
--
Frank Küster
Inst. f. Biochemie der Univ. Zürich
Debian Developer
Reply to: