After many years of trying, Branden Robinson has finally won the Debian Project Leader election. Linux Magazine has an in depth interview with Robinson about his plans as DPL, the problems that face Debian, and what it’s like to finally win the election.
Interview with Branden Robinson of Debian
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Please get sarge out NOW !
Just as the ship is sinking he gets the helm..
Debian is in need of a turn around and everyone loves a good comeback story..
MrX
Debian’s been around a long time. I don’t suppose the 500 odd developers who voted in this election are just going to give up.
Also, I’m not convinced Ubuntu has so much of an edge on the other Debian based distro’s. It may have a lot of money but I don’t see it killing Debian.
I’m glad that he sees value in a calendar-based release process. That’s one of the main attractions of Ubuntu for me. He’s also right that length of the interval between releases isn’t that significant. The key is that the interval be predictable.
I’ve worked for companies that did feature-driven releases, and others that did time-driven releases. In one case, the company changed from feature-driven to time-driven. The improvement was remarkable. Contrary to intuition, the feature-driven approach resulted in the slowest adoption of new features, and the time-driven approach added features at a much quicker pace. If the number of tasks is fixed, the schedule will always slip. If the time is fixed, then it forces you to prioritize and drop tasks. Completing a small number of important tasks is far better than continuous, never-completed work on many tasks.
Do we know what the “voter turnout” was in the end? After that big fuss about how 2/3 hadn’t voted (even though it was only a week in).
“Total unique votes cast: 504, which is 52.227% of all possible votes.” From the link in the article…
http://lists.debian.org/debian-vote/2005/04/msg00028.html
looks like something is wrong, the weeky Sarge ISO has been stopped since 24 March…
maybe Debian is about to make a serous leap forward and make an official release of Sarge, i wish Debian good luck…
The installer team is probably just to a point that they don’t need any public tests at the moment or they are satisfied with their code. There are still 103 release preventing bugs to be fixed before Sarge is release — http://bugs.debian.org/release-critical/
That was a pretty good interview. Dude sounds like he is sane and understands where Debian is and needs to go. The next year should be interesting.
“There are still 103 release preventing bugs to be fixed before Sarge is release”
With the exception of key packages such as those that make up the base system, RC bugs are generally release-critical only for the packages they’re filed against (and dependencies etc…) as opposed to the distribution as a whole. Quite a few (most?) of those 103 RC bugs aren’t release blockers at all. If they’re not fixed, the packages in question can be dropped.
I interpret the page to mean that there are 103 bugs that are preventing Sarge from releasing — period. It seems there are 561 other bugs that would cause packages to be dropped but the 103 are stopping the whole show. I could be interpreting that page incorrectly but I don’t think so.
I’m glad that he sees value in a calendar-based release process. Contrary to intuition, the feature-driven approach resulted in the slowest adoption of new features, and the time-driven approach added features at a much quicker pace.
Agreed. As good and diligent as Debian developers are, the project and developers just seems to need some true incentives and goals to reach bigger goals (like releases) a bit faster: like a fixed time-based release cycle. People and projects tend to work like that, you know…
I think that one problem of Debian may have been that the developers themselves are quite happy with running Debian Unstable on their desktops, and they know how to avoid the bugs there with aptlistbugs etc. (and I’ve grown to do much the same on my home desktop). Quite a few Debian developers probably even run Xorg on their Debian desktops already although even Unstable has only XFree86 still. Servers are a bit different story, but I bet that quite many Debian developers run something else than pure Debian Stable on their servers too, maybe mixed Testing. Again, they know how to overcome the potential hurdles, so why would they feel the need to have Sarge out as an official release?
The whole Debian project needs to make the necessary decisions, changes and other work in order to release Stable(s) faster. But it won’t necessarily ever happen until the project as a whole is *obliged* to do the necessary work.
So, the best and perhaps also the only solution to have a faster Debian release cycle, is to decide to have timed releases. As to the time scale, I think that one year might be a rather natural and ok period between Debian releases(?) – if fewer months (6-11?) would be out of question.
…Just think about all sorts of successfull projects of the largest scale in the history, like space missions, war operations etc: usually they did have fixed time goals besides of necessary feature goals.
A good example from the history:
If the Allied forces (including sarges…) in the World War II had decided that they would attack Normandy only after every tiny detail of the attack would have been assured to work perfectly beforehand (so a bit like “it will be ready when it is ready” philosophy), we Europeans, might live under Nazi rule now… But they decided that on a certain day they would attack; they did it, successfully. Sometimes, and usually especially the bigger the goal is, you just need to decide that you will be obliged to reach that goal in certain fixed time to really reach the goal at all.
LM: Out of more than 900 eligible Debian Developers, only 504 valid votes were recieved – is this a sign of apathy on the part of the Debian Project, or do you think this is a healthy turnout?
—
Well perhaps if you spell-checked before you published people wouldn’t chuckle?
received.
Now onto a serious note:
If I’m not mistaken Branden is in charge of the X-Windows project debs. If the rate at which he is moving from XFree86 to X.org will be indicative of his work as Project Leader we should expect 3.2 to be released around 2008.
Don’t get me wrong, Branden has the years of effort for the title. The title is useless if his actions don’t refine, accelerate and predictably produce revisions of Sarge that will draw more users to the Debian base.
Sid is excellent, for the most part–Eclipse 3 being missing still pisses me off but that’s not Branden’s ownership: perhaps as leader he will give them “encouragement” to move Eclipse 3 into Sid?
complain complain complain……
debian is the mother and should be respected!!!! ubuntu may be great but without debian it wouldnt BE!
debian is trying to take care of details so they can speed up releases but it takes time to do all this, yes we know that sarge is over due, but I consider this acceptable if we manage to get all our ducks in a row and get the infrastructure in place to make everything run smoother….
take your time, get it right, get the infrastructure in place, be happy
long live debian – heck it has lived longer than most and I predict it will be here looooong after most 
Most of the people that think Ubuntu is the best are only speaking from a (nieve) desktop users point of veiw. In the server arena Debian is the best, having stable maintainable servers is critical for many organisations.
Debian does need to speed up its release cycle as many organisations are now using Sarge on their servers as they can’t afford to be held back by the old software that Woody uses.
Another thing I think Debian should do is promote/educate users more about stable/testing/unstable. People seem to have a misconceveived beleif that if they use Debian they must use stable, I think this is just because they don’t know about the other branches.
you sould use stable because if you dont you would not get any security update (timely)
Heh, I liked Robinson’s remark that “Debian’s *unstable* distribution is the wellspring of Ubuntu’s stable one”. As much as I admire the amazing achievements of Ubuntu/Kubuntu so far, I’ve noticed that some over-enthusiastic claims made about Ubuntu’s stability are clearly rubbish. When you use bleeding edge software, it simply cannot be stable. In Debian the claims of stability come from extensive testing and bug-fixing.
I was quite amazed when I read that Ubuntu takes packages from Debian’s Unstable branch, then “stabilizes” them and makes them available as the “universe” repo. I installed Ubuntu Hoary to see this miracle myself, only to find out that the packages in “universe” are marked as not supported by Ubuntu developers. So, this “stabilizing” actually meant that the Ubuntu team did nothing at all to maintain these “universe” packages.
At the same time I used a Debian Testing system and I noticed that the packages on my Debian system were mostly more up-to-date than the same “universe” packages on Ubuntu Hoary, plus these packages on my Debian Testing system received very active maintenance and bug-fixing while the Ubuntu “universe” packages just lay there — untouched, unmaintained and unsupported.
Well, since I don’t use Gnome or KDE and since most packages I installed under Ubuntu came from the “universe” repo, I decided that Debian was, after all, a better choice for me.
Please don’t take this as an excuse for starting yet another pointless Ubuntu vs. Debian debate. Debian and Ubuntu are different distros and they serve different audiences. I wouldn’t hesitate to warmly recommend Ubuntu/Kubuntu to anyone who uses Gnome or KDE. The only point I’m trying to make with this post is that software just doesn’t magically become “stable” — all software needs a lot of real-life testing and bug-fixing to become truly stable. So there’s a good reason why people trust Debian’s stable releases: they are tried and tested and TRULY stable.
Can somebody please post the article in its entirety? I’ve been trying for hours and can’t get in the site. It has been osnewsed.
thx
A 52% voting turnout seems to be about the same ration that a typical US president election gets, so what is everyone worrying about? The land of democracy and freedom… 😉
Anyways, I just wanted to add to the discussion that there, in my view, would be a very sorry and poor GNU/Linux world without Debian. Without Debian’s wonderful software infra-structure, free projects like Ubuntu, Knoppix, ClusterKnoppix, Quantian, Kanotix, Feather Linux and still others would surely fade away. Debian has, above all other “wheel-distros”, made it possible for small groups of individuals to create truly amazing OS distributions specifically tailored for highly different tasks, from tiny USB-key OSes to self-configuring high-performance clusters. It’s great to be able to have a Knoppix rescue disk around, boot a Feather Linux USB key, run Kanotix on a laptop or execute ten parallel week-long scientific computations with Quantian. These projects actually WORK! This is arguably much because the project leades can focus on the fun part, to implement their specific ideas (be it the most intuitive placement of the Gnome trash can as in the case of Ubuntu, or to collect and integrate a couple of gigs if scientific software as in the case of Quantian). They do not need to bother that much about package management and package bug testing – because their are all Debian, and uses the Debian repositories. The fact that Debian has spawned so many projects is truly a sign of technological maturity and stability. Any Debian distro moving away from Debian will have a hard time.
/Andreas
People, please get some idea, and look at the numbers:
http://azure.humbug.org.au/~aj/blog/2005/04/14#2005-04-14-sarge-v-h…
As of 2005-04-04,
Of 3136 packages in Ubuntu Hoary main,
663 (21%) Ubuntu Hoary is exactly same as Debian Sarge
1565 (50%) Ubuntu Hoary is newer than Debian Sarge
908 (29%) Debian Sarge is newer than Ubuntu Hoary
And that’s “main”, where Ubuntu supposedly “concentrate”. In case of “universe”, Ubuntu get worse score.
I cannot understand how other people can think Debian Sarge is very out of date with respect to Ubuntu Hoary. That just doesn’t make sense. Perhaps a bit out of date, definitely not very.
I think debian went bsd when they relesed a developmental release of debian distro running the freebsd kernel. what happen to sarge?
This is false. Very false. So I have to call BS.
At the same time I used a Debian Testing system and I noticed that the packages on my Debian system were mostly more up-to-date than the same “universe” packages on Ubuntu Hoary, plus these packages on my Debian Testing system received very active maintenance and bug-fixing while the Ubuntu “universe” packages just lay there — untouched, unmaintained and unsupported.
Actually there are a team of hard working people that work on the universe:
http://ubuntulinux.org/wiki/MOTU
Unsupported just means that no paid Ubuntu devs work on the universe. Its also a way to simplify support. It does not mean that the universe gets no love…
> This is false. Very false. So I have to call BS.
Actually, it is not entirely false. It’s certainly the fact that Debian testing is more up-to-date than Hoary universe, period. Since MOTU team is currently smaller than whole army of Debian developers, this is to be expected.
Also, it is the fact that Debian testing does receive very active maintenence and bug-fixing.
The only thing false in what he said is that Ubuntu universe is unmaintained. That’s false, but that doesn’t make the whole comment BS. Get calm.
> Unsupported just means that no paid Ubuntu devs work on the universe. Its also a way to simplify support. It does not mean that the universe gets no love…
Unsupported means “maybe it will work, maybe it wont” (personal experience)
Sorry, but there’s just a handful of volunteers trying to keep a few thousand packages in “universe” in shape (the original work done by Debian developers, of course)
https://launchpad.ubuntu.com/distros/ubuntu/+bugs
See for yourself what gets fixed
For example, I am trying to get “hardware-monitor” to work in Hoary, but no matter what I try I can’t. It works in Arch/Gnome 2.10 and Warty… Friendly ppl at MOTU IRC are sympathetic but can’t help.
Sorry.I didn’t mean to be rude. It was just a dramatic portreyal of Hoary that I did not like.
For example, I am trying to get “hardware-monitor” to work in Hoary, but no matter what I try I can’t. It works in Arch/Gnome 2.10 and Warty… Friendly ppl at MOTU IRC are sympathetic but can’t help.
See if the backport team can help you..
I apologize for my slight mistake. It just felt to me like these packages were unmaintained. I’m glad to hear there’s at least good intentions among the Ubuntu community to maintain the universe packages. Hopefully these good intentions will in time grow into a serious effort.
Debian GNU/kFreeBSD has nothing to do with the release of Sarge.
It’s also not an official port (yet), so, no, Debian has not made any official release of a GNU/kBSD system.
What seems to make the most sense is for the Ubuntu folks to focus their efforts on the packages most relevant for desktop users. This way, even though Debian will have more up-to-date packages overall, Ubuntu will (may?) have more up-to-date desktop-related packages. I don’t know too much about Ubuntu — maybe they’re already focusing on things this way.
Seo — I’m curious: if you limit your analysis to desktop-related packages, how do the two distro’s compare?
Debian GNU/kFreeBSD has nothing to do with the release of Sarge.
It’s also not an official port (yet), so, no, Debian has not made any official release of a GNU/kBSD system.
I think it does has something to with it, because they are working on it instead of trying to release sarge.
GNU/kFreeBSD is mostly a single man project, with a few interested developers joining in. It has nothing to do with the sarge release. Right now there are three main problems blocking sarge:
– the infrastructure to support security updates for sarge is still not complete.
– Due to downtime the ARM autobuilders have a huge backlog.
– There are still about a hundred release-critical bugs in sarge. Every week some are fixed and some new bugs come to light.
If I’m not mistaken Branden is in charge of the X-Windows project debs. If the rate at which he is moving from XFree86 to X.org will be indicative of his work as Project Leader we should expect 3.2 to be released around 2008.
X.org hasn’t been merged yet because they’re waiting for the modular release so that the packages can be done properly. I do believe there were X.org debs created for sid but they never got pushed in because that would have caused more problems than it fixed as every new package in sid would have been built against X.org making them either force X.org in sarge (and thus have to rebuild a ton of sarge apps) or leave sid incompatible with sarge making it impossible to get new packages into sarge.