The decision making process in the Debian process is not fast enough and can be held up by a few crappy and difficult people. Ubuntu fixes that, and does not try to support 15k packages.

Ubuntu is agile and can try more new things to see what works. Ubuntu can show millions of Open Source buyers to hardware and software makers, while hiding our few malcontents. It could be a beautiful OS ecosystem if both distros play nice — so that the fruit of the Ubuntu tree can fertilize the soil of Debian.

Ð Ubuntu syncs a lot from Debian, and needs to try to give back more without sinking into the dysfunctional culture of the project.

Ð Debian, on the other hand, needs to take the major changes from Ubuntu that are working (like Upstart?).

Actually he hasn’t forked Debian, he’s created a branch. There’s a big difference. A fork has a life of it’s own and over time will diverge (e.g. Mandrake is a Red Hat fork that is now quite different from Red Hat). A branch is like a fork, except that it periodically gets remerged. In Ubuntu’s case, each release is a snapshot of Debian unstable that’s been stabilized for a time. The next release resynchronizes with Debian. This allows Ubuntu to have packages that are almost as fresh as Debian unstable without having to worry about things breaking every once and a while (the whole purpose of unstable is to precisely be able to test out new changes and break things if needed).

Building Ubuntu off of Debian stable misses the point of Ubuntu (always use the latest and greatest). Building it off of Debian Testing wouldn’t work unless the Debian Testing conformed to Ubuntu’s model (i.e. Debian Testing is a 6 month snapshot of Debian Unstable and Debian Testing started getting security patches).

Personally, I think it would be a *great* idea if Debian changed this way (it would make having regular Debian stable releases *a lot* easier). However, it’s not Ubuntu’s place to push Debian around. Debian is a big meta distribution that’s been around for over a decade and has a social structure that dwarfs Ubuntu’s. Ubuntu may get all the glory but Debian is the silent workhorse that makes it all possible. If Debian sees the benefits of making Debian Testing a 6 month snapshot, there’s little reason why Ubuntu would choose to create a Debian Unstable snapshot branch. Ubuntu would likely continue to create a snapshot of GNOME or KDE (since it’s unlikely that Debian will synchronize it’s Testing snapshot with GNOME), but the size of the branch would be a lot smaller and would be usable by any other Debian Testing distro.

For Dapper (and other releases that will be supported for a long time) this is a very good idea.

Debian’s next version(etch) will ship with 64 bit(officially) support… but debian’s current version runs fine on 64 bit and you can download the 64 bit version from http://www.debian.org. And from your comment “I’d rather see canonical helping debian in moving to new platforms (64-bit, embedded etc)” … debian supports the most architectures out of any distribution…it supports i386, AMD64, PowerPC, 68k, SPARC, DEC Alpha, ARM, MIPS, HPPA, S390, IA64 … So I doubt debian needs to move towards new platforms… Ubuntu forked for this reason particularly(among many), that is doesn’t have to support all of the architectures that debian does. Ubuntu only supports 3 architectures… i386 powerpc and amd64.

## For Montana
“SPM for Ubuntu doesn’t necessarily mean the particular program is available for Debian. It does from the other side, however: that is, from Debian to Ubuntu.”
This depends on which version of Ubuntu and Debian you are speaking about. And generally this statement is false. Ubuntu only gives _official_ support to a small amount of packages. You have to manually edit sources to add other packages.(which aren’t support[if you pay for support?]) Debian has over 15,000 packages in main…which can be accessed by default. And if you are running Debian testing(which a lot of people run on their desktops) you could easily have programs in debian etch that aren’t in Ubuntu and vise versa. So while most programs that are in debian are also in Ubuntu…not all of them are.

And both these comments so far seem to miss the point of the article. The author seems to suggest that Ubuntu should remerge with Debian…this way efforts would be more combined.(many of the Debian developers are ubuntu developers..and vise versa)

Mark Shuttleworth is a debian developer. This is how he started…and he wanted faster release cycles essentially. And debian does not want to go that fast. So mark forked. And created Ubuntu. Mark Shuttleworth could have shaked up the project and made things go faster…but he didn’t want to disturb/ruin the project as most debian developers are volunteers and they program debian in their spare time. Where mark hired 20ish full time developers to work on Ubuntu full time. ubuntu is based off of debian. Without debian ubuntu would have a hard time surviving/existing at all. So Ubuntu team not only needs to manage their own project. They need to make sure debian is moving in a certain direction…so they can guarantee Ubuntu’s sucess in the future also. This is difficult to do…as debian has stipulation in their contracts about corporations(Ubuntu/Canonical Inc.) not being able to control/mess with their project too much.(This is good as it doesn’t allow microsoft to destroy them, heh) I believe Mark’s move to fork debian was not a good idea. I think he should have taken a risk and tried to shake up the project. As things the way they don’t seem to be the optimal productivity that they could be. I think it would be possible for him to remerge with debian…and would be more likely now..as he has proven himself and his worth. And proven the project(Ubuntu). I do not think he will do this though. And that is ok…things can be forked…that is the whole point of GNU/Linux … and both projects are doing very well and are very sucessfull.

I think there needs to be more structure in the way Ubuntu and Debian are tied to each other…and the developers need to not think of each other as competitors…but as collaborators. But again both projects are doing very well…and not much needs to change. They are obviously moving in the right direction. It just takes time.(and now money?)

I’d rather see canonical helping debian in moving to new platforms (64-bit, embedded etc)

ikd

Good topic. We need more of these!