Thoughts on OpenSUSE

A few months ago, Linux Magazine columnist Jason Perlow suggested that Novell open SUSE Linux as a "public, open source project similar to Fedora." Thanks in part to Perlow's column, it looks like that's exactly what Novell is doing. Novell had planned to announce this at Linux World Conference and Expo, but it looks like it leaked out a bit early. (And they're not even suing anybody...)

Notice a trend here? Novell/SUSE was one of the last holdouts, in terms of doing "closed" development only for its distribution. Even Sun beat SUSE to the punch in going forward with OpenSolaris. But, better late than never. The writing on the wall is pretty clear: It makes no sense for a Linux vendor to follow a "closed" development model in developing a distribution. That's not to say Novell/SUSE has failed to produce a quality distribution over the years -- SUSE has consistently been one of the best distributions (at least in my opinion) for desktop and server use.

Despite its technical excellence, however, SUSE has failed to overtake Red Hat's lead in the Linux market. In truth, SUSE should have taken this step when Red Hat made the decision to have separate enterprise and "community" products. That would have been the time to catch all the users and organizations who were caught between enterprise pricing (RHEL) or a distro with a short life-cycle and little regard for stability between releases (Fedora).

Now that Novell is going forward with this, I have some suggestions. The top of the list is to make sure that people can easily contribute. Novell will need to move forward, quickly, with a public CVS or Subversion repository and a good system for granting commit access to developers outside the company. Obviously, Novell is going to take a hit to retail sales of SUSE Linux Professional, so they need to get the most bang for the buck by making SUSE the distribution that developers want to use and contribute to.

In my opinion, Red Hat did not do a very good job at community-building with Fedora Core out of the gate -- something Ubuntu has done spectacularly well in a very short time. Novell should look closely at Ubuntu as they move forward with OpenSUSE. I'd also recommend that Novell follow in Sun's footsteps in forming a Community Advisory Board (CAB) with members from outside Novell. I also like Sun's blogs, which provide some really nice insight into what's going on with OpenSolaris -- but it already looks like they have that covered.

A timed release cycle would also be a very good thing. This is working very well for Ubuntu, OpenBSD (which, I believe, was the first open source project to move to timed releases), GNOME and others. For community projects, a predictable release cycle is much better than a feature-based release cycle -- otherwise, things tend to get held up indefinitely while waiting for certain features. This also tends to motivate developers, who want to make sure they don't have to wait another six months for their pet project to make it into the stable release.

I'd also like to see Novell become part of the Debian consortium that Ian Murdock and others are working on, but I suspect that's highly unlikely.

I'm planning on being at LWCE next week, so I look forward to chatting with some of the Novell folks about OpenSUSE and getting the specifics of the project. If you had the chance to get some face-time with Novell, what would you want to know, or suggest for OpenSUSE?