Is Debian Dying?
Off the Beat: Bruce Byfield's Blog
Steven J. Vaughan-Nicholls created a stir this week when he marked the release of Debian 6.0 by wondering if the distribution was still relevant. He was refuted by Joe Brockmeier, and the discussion spilled over on to Facebook, where a number of journalists (including me) speculated freely. But the noticeable lack of hard facts disturbed me, so I decided to see if I could find any indicators of Debian's health on-line.
On Facebook, Vaughan-Nicholls explained that his basic question was, "'Where will the new Debian developers come from to keep it going?' I see 20-something developers working on Android or Ubuntu, Debian, not so much."
That seems a reasonable place to begin, but it needs some qualification. For one thing, I don't think that anyone interested in Android is also going to be interested in Debian. For another programmers do not necessarily choose between Debian and Ubuntu these days.
True, some Ubuntu contributors, like Mathew Garrett, do seem to have become inactive in Debian. Others, like Benjamin Mako Hill, have reduced their involvement. It is also true that relationships between Debian and Ubuntu have sometimes been acrimonious, with some Debian members regarding Ubuntu as prospering unduly at Debian's expense and without contributing its share of development. As a result, at times developers might have felt the need to choose between Debian and Ubuntu.
However, the relationship seems to have improved greatly in the last few years. Debian has set up a Derivatives FrontDesk page to increase cooperation between Debian and distributions like Ubuntu that derive it. Similarly, Ubuntu's Debian for Ubuntu Developers page stresses the importance of working with Debian developers, including the importance of reporting bugs to Debian package maintainers.
Shuttleworth's statement that "Every Debian developer is also an Ubuntu developer" is perhaps exaggerated, but with 89% of Ubuntu packages coming from Debian, it continues to have a great deal of truth. Clearly, developers do not always have to choose between Debian or Ubuntu, and are actively encouraged on both sides not to see the situation as an either-or choice.
So far as I have been able to find, no one keeps track of how much Ubuntu-centered developers contribute to Debian. However, at least for the approximately 2000 packages that make up Ubuntu, it seems safe to guess that the contribution is sometimes significant.
The new Debian release also benefits from new packages from Ubuntu, such as the rebranded version of the Ubuntu Software Center, one of the better graphical package managers currently available. Undoubtedly, there are other examples.
The Size of Debian
But if we disregard the unmeasurable contribution that Ubuntu makes to Debian, is there any sign that the number of Debian developers is declining?
At first, Christian Perrier's statistics of the number of Debian developers and active Debian developers for 2009 and 2010 seems to indicate alarming drops. Perrier lists 1461 developers in 2009, but only 1410 in 2010 -- a drop of nearly 4% in eleven months. The drop in the number of active developers is even greater, falling from 999 to 873, or a loss of 13%.
However, these numbers do not tell the whole story. In particular, the Debian project has been making an effort in the last couple of years to weed out the inactive. Since we do not know how long the inactive have not been participating in the project, the decreases may be as much an indication of the effort to remove the them as proof of a sudden decline.
Moreover, these base numbers do not indicate recent recruitment. Since Perrier's 2010 statistics, 24 new maintainers have joined the project , which reduces the loss to 11%.
Moreover, another 55 are going through the rigorous acceptance program. When these recruits are added to the number of active developers in 2010, then the loss shrinks to just over 5%. Add the 25 applicants who are on hold, and the loss is only 2%.
Of course, any change over a year is not enough to establish a trend. Nor can we assume that all new maintainers in the middle of the process will eventually become Debian developers.
However, we can say that, from the evidence available, no one can prove a pattern of decline. Moreover, we can say that the small decrease for the single year is partly voluntary, in that it is due to the Debian project's insistence on examining the fitness of new recruits closely. With laxer standards, Debian could have easily kept the number of developers at the same level.
At any rate, the number of developers, no matter how you interpret the available figures, is not the whole story. As a by-product of studying the md5sums used in Debian packages, developer Romaine Francoise concluded in August 2010 that the new release represented an increase of 30% in the number of packages in Debian over the previous release in 2009. This figure compares favorably with the 2009 release's increase of 24% in packages over the 2007 release.
Judging from these figures, Debian's growth remains as strong as it always was. Just as importantly, because the number of developers has declined at the same time, those who remain are apparently doing more with less. So far, the decline in developers is small enough that it appears to be irrelevant to the success of the distribution.
Exact figures are hard to come by. However, those that are available suggest that the Debian project is in no immediate danger of being unable to do its self-appointed tasks.
From the available information, you cannot say even say that Debian is considered less relevant to most people. The truth is (no doubt because of its reputation for being an expert's distribution), Debian has never been the most popular distribution, although always remaining solidly in the top ten. If you look at the Page Hit Ranking on Distrowatch, in 2003 -- before Ubuntu even existed -- Debian was in the fifth spot. Now, in 2011, it remains in precisely the same spot.
True, Debian dipped to seventh position in 2007 and 2008, the years when Ubuntu was first gaining recognition. Yet even that drop is slight, and was apparently only temporary.
The indicators are not complete, so they cannot be definitive. Yet they all suggest the same story: Debian is holding its own. Short of a natural disaster at DebConf, you can be reasonably sure that Debian will continue much the same as always for at least the next few years.
Now, if you excuse me, I have an apt-get dist-upgrade to run on another machine.
aptitudeGood points. I just got annoyed as usual with Vaugn whotisname's article and had a go at him in the comments (no, not using rude words or anything, that gets us nowhere).
Oh, and you should be using aptitude not apt-get ! :-P
Link FixedThanks...The link has been updated.
Broken linkThe link to http://raphaelhertzog.com/2...he-debian-ubuntu-relationship/ is broken, you have a leading parenthesis that is problematic.
Re: Debian in DistrowatchThe exact ranking can change. When I looked, it was fifth.
Debian in DistrowatchI'm not sure if we're looking at the same page, but Debian is currently in the fourth spot in Distrowatch, not the fifth.
It's also interesting to highlight the fact that of the 5 most popular distributions according to Distrowatch, 2 of them are direct derivatives (Ubuntu and Mint), which obviously weren't there before 2004, and wouldn't be there today if it weren't for Debian.
So if we take Ubuntu and Mint out of the equation, we could say that Debian is actually #2, closely behind Fedora.
Debian is successful on serversThe W3Techs statistics suggest that Debian continues to be one of the most successful GNU+Linux distros on severs. Ubuntu, for example, is not nearly as popular as Debian.
Linux Foundation's big event celebrates the 25th anniversary of Linux
Linux has evolved from a “won’t be a professional” project to one of the most professional software projects in the history of computers.
Competitors get in the game with RHEL without Red Hat
Security researchers have already notified Microsoft; some fixes are available
The company is collaborating with Google and Intel to use Kubernetes as an engine for Fuel
Customers can take a free test drive of SLES for HPC on the Azure Cloud
San Francisco-based chip company announces their first fully open source chip platform.
The whole distro gets rebuilt on glibc 2.3
Ubuntu Vendor tries to solve app packaging and distribution problem across distributions.