[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Question for the candidates: Custom Debian Distributions

Enrico Zini wrote:
This is to introduce the topic.  My question is: what vision do you
have for the future of: Custom Debian Distributions; partly-derived
distributions such as Ubuntu; and forks such as Knoppix or Linspire?

First, for the record, my response to LWN on the Ubuntu question was:

>> Some have claimed that Ubuntu is hurting Debian by drawing developers
>> away from Debian. Do you agree with this? How should Debian work with
>> distributions that are derived from Debian? Is Debian
>> "infrastructure"?
> The only way Ubuntu can draw developers away from Debian is by
> providing a better environment for hacking -- whether that be by
> paying for the work, or being more fun, or being more satisfying, or
> all of the above. I think it's great that there are projects that some
> people find more enjoyable than Debian, and the great thing about free
> software is that those of us who prefer Debian can just take the work
> they do for Ubuntu and use it ourselves. And vice-versa, too -- all
> without anyone being unhappy about code theft or having to involve
> lawyers or formal agreements or anything of the sort.
> I think Debian works quite well both as a distribution of its own, and
> as infrastructure for other distributions; I hope it will improve as
> both.

But to be more specific, I think Debian's had problems for years in dealing with derivatives -- not so much in making them possible, I think we do a great job of that, but rather in taking advantage of them and making it easy for derivatives to contribute back to us.

The first example I can remember is the work VA Linux and Joey Hess put into the "slink and a half" release, which was basically an update to slink while we were still working on potato. That was fairly well done and tested, and probably could and should have allowed us to make a lot of stable users happy by giving the access to some more current software, without distracting us from our work on potato, or lowering our standards for stable software. Unfortunately it turned into a non-starter, with Debian being unable or unwilling or both to even work out how to accept the contribution.

Likewise, in spite of Stormix, Progeny and others building useful graphical installers for Debian, we were unable to get any benefit at all from those efforts. That example has been improved by the development of debian-installer though, which derived distros have been able to contribute back to, and in some cases have taken the lead in developing.

We've not managed to work out a way of integrating the cool live CD concept from Knoppix into Debian proper either; though in a sense Knoppix itself is pretty much a "proper" Debian live CD anyway. Ubuntu's also managed to create live CDs for their distro. Having this integrated into Debian properly would have a few benefits: it'd imply support for more architectures, it'd make it easier for derivatives like debian-edu to get a live CD for free, and in general it'd let people not have to worry about trading other cool Debian features to get and maintain a live CD.

Ubuntu brings that question into somewhat starker light; both because there are more contributions going into that project to feed back into Debian, and because Ubuntu is an active development project, there'll keep being new things to feed back into Debian. If we're not successful in managing that, it'll probably be bad for both Debian and Ubuntu; bad for Debian in that we'll be missing cool features, and bad for Ubuntu in that they won't be able to leverage the work Debian does in maintaining old stuff, in order to focus on developing neat new stuff. In the worst case, doing this badly could create problems for the viability of both Ubuntu and Debian -- Ubuntu in that paying people to do all the work Debian volunteers currently do might be impossible or at least unaffordable; and Debian in the simple sense of why work on Debian if Ubuntu does everything better?

One approach we could take here could be to make some space in the archive for packages that derivatives have had to patch, and that haven't been merged into the main Debian package yet; if we can manage to deal effectively with the other components of that: dealing with multiple active versions of a package in the BTS, dealing with packages with multiple maintainers working separately instead of together, potentially dealing with limiting uploads to authorised maintainers for some derivatives, and of course any other complications that might arise, then I think both Debian and its derivatives would starting reaping the benefits fairly quickly. I suspect there are other approaches we could take too, either instead of or together with the above, and whatever approaches we take, it'll definitely involve working closely with the derivatives to work out what will actually be useful and effective for them.

Long term, I hope derivatives will be unsustainable, in the sense that there just plain won't be any reason for them to want to work outside of Debian. Short term, there's plenty of reasons why that's a sensible approach, and I think we should be paying attention to that, and learning from it.


Reply to: