Spinning a New Kind of Distro

When Fedora 8 came out last week, the first thing that came to mind was this was the first release of Fedora in recent memory that didn’t have some sort of glitch associated with it.

You know the kind: early mirrored versions come out, someone gets wise and then posts it on Digg or Slashdot, people start pulling down ISOs, chaos ensues.

This time, however, there was none of that. Announcements went up, mirror sites snapped on, and Fedora 8’s first day was smoothly done.

This may seem like a weird thing to point out, but I think it offers us a glimpse of the overall efficiency that’s becoming more of a factor in the Fedora Project overall.

Since the ascension of Ubuntu, it seems like the stalwart old distros like Fedora (Red Hat) and openSUSE (SUSE) have been pigeonholed into the category of “also-ran.” Nothing was wrong with the veteran distributions–they just didn’t seem to have the intangible spark that made the new kids on the block so desirable.

That judgment may have been too hasty. You don’t read about Fedora OEM deployments on PCs or embedded devices, but I have a feeling that’s about to change. While talking with Fedora Project Leader Max Spevack yesterday, I increasingly got the sense that Fedora is positioning itself for something bigger.

The key, I believe, is the spin management technology that was implemented in Fedora 7 and has now come to maturity in Fedora 8. Spins, in Fedora-parlance, are customized versions of Fedora that users can create using Fedora’s on-board mixing/package management tools. Right now, besides the two official KDE and GNOME spins from Fedora, the Fedora Spin site is host to Electronic Lab, Games, and Development spins.

On the surface, this sounded pretty nifty, but I could not help but wonder how having spins helps Fedora’s ultimate downstream product, Red Hat Enterprise Linux (RHEL). Spaveck explained that, like these custom spins, RHEL was a subset of the 9,000-plus packages that a available within the project. Introducing the remix tools to RHEL users would give end-users a chance to customize their own RHEL-based spins even more.

That’s the solid surface reason. I also wonder if remixing tools might not also be a boon to hardware vendors who would need to be able to efficiently pick and choose exactly which packages they want to install and ship Fedora on their platform–be it PC or cell phone. This remixing tech also falls neatly in place with Red Hat’s new appliance program. Whether you are building a software or hardware appliance, remixing tools will certainly help in that endeavor.

This, then, will be the differentiator between Fedora and the *buntu family. If Fedora’s spins can be more quickly adapted to individual needs, then it will have a potentially significant advantage over the *buntu family. As good as the *buntus are, creating custom versions of them would be slower than Fedora’s spin process.

The real question is, however, is such an advantage something that will make Fedora more attractive? Time will tell.