Why Ubuntu release schedules should be changed

Sam Varghese
Wednesday, 18 November 2009 04:51

Opinion and Analysis

Additionally, Ubuntu does itself no favours by trying to put out something every six months - in my opinion, just to give the impression that it is in perpetual motion - because users are under the impression that they can just replace what they are running with the latest release and expect the same behaviour.

Imagine - you have sorted out the problems with a release and started doing some productive work. You then upgrade to the next simply because it has come out - you simply can't resist and, after all, every one of your buddies has that new shiny toy. You have to get it too.

Unfortunately, this scenario is played out over and over again - and users never take responsibility for their own mistakes, they blame Canonical. This is a problem of perception which has to be corrected.

An annual release would provide more time to sort through issues, more time for consultation and coordination, and more time for the small pool of developers who work for Ubuntu to sort out issues that cause them more pain than gain when they emerge after a release.

It is important to remember that despite much talk about the deep pockets which Canonical owner Mark Shuttleworth has, the number of people who actually work on Ubuntu is relatively small.

(Shuttleworth is acutely aware that it is a long, painful battle to establish GNU/Linux on the desktop and he is in it for the long haul; he is not interested in heavy investment over a few years, but rather a smaller investment over a much, much longer period.)

Add to this the fact that software releases are _never_ on time and you have very difficult hurdles to surmount.

A third reason is Ubuntu's marriage to GNOME. As is well known, GNOME is planning what it calls a paradigm shift for version 3.0 - but has been forced to push back the release by six months to September 2010.

This version is expected to be something that will bring GNOME out of the middle ages and into contention with the new, flashy KDE4.

After an initial perception of being problematic - mostly caused by ignorant reviewers who assumed that a .0 release was meant for prime time - KDE4 has proved itself to be a snazzy desktop, with several very interesting features.

And, after having had a play with a recent GNOME snapshot, it is evident that the developers there are trying to match KDE. Given this, it is self-evident that the bug count in the new GNOME will be proportionately higher.

If Ubuntu were to stick to releasing a month after GNOME releases, then all those bugs would be very much part of its release, bugs that would certainly not help its cause. It is extremely doubtful whether a month would suffice to clean out the bug cupboard properly.

And finally, given that Canonical is going to source the next release, 10.04, from Debian testing, it is bound to be very, very stable.

Hence, even small bugs will be easily noticeable in the following release; 10.10 will include the new GNOME 3.0 and probably be extra unstable. Not that one wishes to cast aspersions on the GNOME project; one cannott expect the number of changes which are coming in from version 2.30 to 3.0 to be bedded down so soon.

For Ubuntu, this will not be good at all - when you have even a small wave washing up on a quiet beach, it is noticeable. Imagine what a bigger wave on that quiet beach would look like.

If Shuttleworth can put some thought into the idea of annual releases - and then try and get Debian to release every two years - it will be possible for him to get the best from both projects: GNOME and Debian. Ubuntu has little left to prove - apart from the fact that it can be used on the desktop without fear of any more showstopper bugs.