Now that Debian Sarge is finally out the door (insert big applause for the Debian folks for passing that milestone), what comes next? Obviously, work begins on Etch, the next release -- but what should Debian's priorities be?
Ian Murdock has a few suggestions. Two, actually:
Debian should have two overarching priorities for the next release: 1. putting a timed release cycle in place, so what happened with sarge never happens again; and 2. keeping the growing family of Debian derivatives united around a common core—namely, Debian itself. What’s at stake? Bottom line: If we don’t do something about both of these problems, actual and potential, Debian will be irrelevant by the time etch is out.
I hate to do the "me too" thing, but I agree wholeheartedly, specifically about Debian's release cycle. Release cycles have been on my mind a lot lately, as I plan upgrades and fresh installs and talk to other system admins who are doing the same. Ideally, Debian would release every 18 months and continue security support for 18 months after a release is superseded.
The only thing that makes Debian hard for me to recommend is the fact that its roadmap is a bit unclear. I've worked with Red Hat's Enterprise products and Fedora Core, SUSE Linux, and many others -- but I prefer to manage Debian systems. However, it's difficult to expect businesses to plan around release dates of "whenever."
There are also suggestions to make testing more suitable for users. I'm not sure whether that solves the problem. It certainly has a marketing difficulty -- if the Debian folks want to encourage use of testing for production, they need to rename it.
Several users suggest that Debian needs to be split between server and desktop releases. In a way, I kind of see Ubuntu as the Debian desktop release. I use Ubuntu on my laptop and main desktop, and Debian on my servers. Many folks say that Ubuntu is satisfactory for servers, and I'm sure they're right, but the 18-month lifecycle for Ubuntu isn't quite long enough to suit my needs. (Ubuntu releases are scheduled every 6 months, with 12 months of support afterwards.) Many folks also use Debian on the desktop, but I've found Ubuntu to be much more suited to desktop use -- at least at the moment.
Murdock suggests a 12 to 18 month release cycle. I'd even be happy with a 2-year release cycle, with an additional year for security updates. On the desktop, 12 months is ancient, but on the server, I'd like to see something with at least a 3-year lifespan at a minimum.
I know far too many businesses that are still running Red Hat 7.3 and even Red Hat 6.2 in production because of the general pain of moving off of those releases. Open source development moves at breakneck speed - and it's impossible for companies who build products on top of Linux (and here I'm using "Linux" in the general sense of the OS and all the tools that come with) to keep up. Judging from some of the comments to Murdock's post, I'm not the only one looking for slow -- but predictable -- releases.