Re: Questions for Martin
* Raphael Hertzog <rhertzog@hrnet.fr> [2003-02-20 23:49]:
> You explain that too many external projects do not make the effort
> to integrate their work within Debian and you want to do better in
> this area ... but do you have some specific external projects in
> mind already ?
Well, I know of two cases were it seems to work pretty well. One is
Skolelinux and the other is Trinity College. Also, when Progeny used
to work on their own distribution based on Debian, they contributed
patches back to the project. However, while Lindows makes source
available, I don't really see many patches from them in the BTS.
Similarly, I don't see many patches or bug reports from Knoppix or
Knoppix users, although they are based on Debian.
There are other efforts which go wasted, but I couldn't tell you any
exact projects now (partly because I'm behind a really bad Internet
connection at the moment). However, I know that there is at least one
company using Debian in an embedded system. I'm sure they made some
modifications, but they were never contributed to Debian. Perhaps
this is because they were only hacks (like removing /usr/share/doc
instead of implementing an option in dpkg to actually exclude specific
directories or files). However, for the greater community, it would
be better do implement it properly and give it to others. I'd like to
encourage that type of behavior.
> Don't you think that there are logical explanations why their work
> is not always integrated into Debian ? And if you identify problems
> that make it difficult to integrate their work into Debian, do you
I think there are good or fairly good reasons in some cases. Also,
there might be reasons that Debian makes it difficult to integrate
work from others. For example, it's clear that by our policy the
woody backports won't get included in woody. However, that's not an
example of a wasted effort; I think the backports are good (I use the
backported Mozilla myself). However, there are many examples where
there is no good reason. One example would the the unofficial KDE 3
packages. They could have been done together with the KDE package
maintainers, but they weren't and now there are lots of upgrade
issues.
> You're speaking of a possible certification program in cooperation
> with hardware vendors ? Can you elaborate on that ? I find this idea
> a bit strange ... I don't see that as very important but maybe you
> can convince me of the contrary.
One huge problem Debian has is that many people go to big hardare
vendors and want Debian pre-installed. However, those companies don't
offer Debian due to various reasons. In most cases, it's a combinatin
of a) Debian is not a company so we cannot make a support contract
with them and b) Debian does not have a certification program so we
cannot guarantee that our hardware will actually work with Debian
("we" being the hardware vendor).
I think we should try to address both of these issues. A first step
will be to talk to vendors and figure out what they actually need, and
how we (or someone else) can offer that (This might actually be an
opportunity for Debian Developers to get support contracts or set up a
company offering commercial support).
> When you speak of the internal functions, you're mentionning
> sponsorship, new maintainer and mia developers. That's what you know
> best. But did you identify other problematic parts in the inner
> working of Debian ? What do you think about debian-admin,
> ftpmasters, RM and so on ?
I cannot give you a status report of how those are working. Although
I know that the ftpmasters are doing a very good job (for package
removals and NEW packages, partly due to the new archive tools), I am
not in a position to say much of the others. However, if I'm elected
as DPL, I will talk to them to get status reports and to see how I can
help them. Bdale actually asked all people listed on
www.d.o/intro/organization for a status report at the beginning of
this year, so I will ask him for this information as a starting point.
However, since I have to ask all delegates whether they are still
interested in what they do, I can ask for status report at the same
time.
> You're encouraging other people to become mini-DPL or mini-RM,
> acting like responsible people who can do almost everything provided
> that they are reasonable, polite, etc. Don't you fear the chaos ?
> Don't you think that giving official titles to people encourage them
> to work on a specific task ? Don't you think that having a team
> instead of a single RM would be better ? What about a vice-DPL ?
I dont fear the chaos at all. Debian works by having people who work
on various things and coordinating with others. I'm also not conviced
you need an official title in order to get work done. In fact, my
work on inactive maintainers showed the contrary. In any case, I
would support a team for release management; however, that's really up
to the current RM to decide. Also, while I encourage other people to
help with release management tasks (and other stuff), I also
acknowledge that we need one RM to make final decisions.
I'm not sure we need a vice-DPL; however, a secretary might be a good
idea, depending on the actual workload. I cannot say anything about
this yet, however, since I haven't been or are a DPL... ;-) However,
I'm sure there are quite a few mails sent to leader@d.o which take up
much time and which don't really have to be answered by the DPL.
--
Martin Michlmayr
tbm@cyrius.com
Reply to: