I recently had lunch with a friend from RedHat. Suffice it to say, it was rather revelatory. At some point, the conversation drifted to rPath and how they split off because Red Hat "couldn't afford to please the diehards" - they had to make money, and the only way to do this was by ignoring the "diehards". This Red Hatter then went on to talk about how they couldn't just continue to give stuff away, they had to charge for it, yadda yadda - pretty standard stuff we've heard from Red Hat for a while now. It then occurred to me how Red Hat continues to get it wrong in the marketplace - they still think it's about engaging in solid business practices in spite of the resistance of the GNU diehards.
This very premise is simply wrong. They seem to have neatly categorized criticism of its abandonment of the desktop as just noise from the "diehards" wanting Red Hat to serve the free software community and give stuff away for free. This line of thought ceased to be relevant about 6 years ago.
What Red Hat doesn't seem to understand is the effort that will be required to bring parity to the desktop landscape. The reason this is important is that Microsoft is continuing to use the leverage from its large desktop install base to build server-based technologies. While Red Hat is laser-focused on web servers and other back-office sales, Microsoft continues to bundle more services into its offerings, and for the most part, it just works - assuming, of course, that you use Microsoft on both ends of your transactions.
Red Hat does not now, nor has it ever, grasped how much leverage it would have with a ubiquitous, user-friendly desktop. They want so badly for everyone to equate Red Hat with Linux, and they have been very successful at this. However, Red Hat just isn't large enough to outflank, on its own, a company as large as Microsoft, and yet they continue with their market myopia. Witness their absence from LinuxWorld San Francisco. Witness the bad relationship between them and some of the larger commercial entities (there are several). They want to single-handedly drive Linux forward, and they do not have a great track record in terms of working with other companies. It's no secret that they have not been strong advocates of the LSB. One note of hope is their acquisition of JBoss, but there is a distinct lack of solid partnerships between Red Hat and other strong commercial open source players. They continue to strive for a Red Hat-only market, without engaging users on a large-enough scale to create new markets.
Given this view, a lack of engagement with the commercial open source ecosystem, and a bias towards its own technologies, it would seem that Red Hat is doomed - in terms of matching its ambitions to eventual success. Unfortunately, a doomed Red Hat spells a temporarily doomed commercial Linux space, and thus a temporarily doomed commercial open source space. Red Hat does not seem to recognize that a lot is riding on their success or failure - like, say, the entire commercial Linux ecosystem. And since Red Hat/Fedora Linux is the default development platform for open source ISV's, a lot of other software infrastructure would fail with it. Realistically, Red Hat can push out most of its competitors on the server landscape - Sun, Novell, et al. - and lose to Microsoft in the back office and the continuously growing web infrastructure. It would be a classic case of winning battles but losing the war.
The other strike against Red Hat is they seem to discount Microsoft's efforts in the commoditized web infrastructure area. They do this at their peril. Microsoft appears to be learning how to make its software cheap enough to make it compelling. Given their established user base and the fact that many admins feel at home with it, IT buyers are willing to pay more for it. Again, a ubiquitous desktop is largely responsible for this. It certainly doesn't hurt that a large amount of open source software runs really well under Windows. Conceivably, assuming Microsoft doesn't screw up and that open source .NET continues to flourish, one could imagine a day where Windows becomes the de facto open source development platform.
Enter Ubuntu/Canonical. If one takes the view that the desktop is vitally important *and* a market in need of a brash, ambitious upstart, then Ubuntu seems to be a Linux distribution that understands what is needed to clear the major hurdles. That is, many many more users (and eventually developers) are needed to bring the software market to parity and give leverage to the smaller software players. New markets need to be pushed wide open, and they need a compelling reason to use Linux. Yes, this means stuff needs to be given to them. Yes, Ubuntu is bleeding money at the moment. My point is that this is a necessary evil - for now. I don't think it's a coincidence that Red Hat is public and Canonical is not. Of these two companies, despite Red Hat's current market position, it would seem that Canonical/Ubuntu is best positioned to drive commercial open source in the future - only they seem to understand the scale of the task before them. "Community building" is not some touchie-feelie exercise in charity - it is shrewd business development.
Oh but wait, Ubuntu is just a free toy given away by a crazy South African spaceman to please the GNU/Linux diehards, right?