"Ultimately, I'm not predicting that Linux will take over the market next year. Or anytime soon, for that matter."... "I am predicting that users will switch to Linux in record numbers next year. And many will never look back."
Okay. Well, while we plot and scheme against closed source software, let's not make the same mistake that Microsoft was and assume that we can somehow usurp them: we're symbiotic.
Microsoft started targeting universities a while back, getting the CSci departments on to Windows. Microsoft needs developers, so this seems like a win. And putting computers in schools to convert people while they were young worked for Apple. But it backfired. In CSci, just like in any other research setting, you needed the source to do any useful work. Microsoft has always hired a lot of ex-Unix and ex-VMS (etc) programmers, but now the graduates don't know how to work on operating systems internals. A lot of good schools still use the wizard book and still study real working systems, but most students go to the state university and increasingly, the community colleges. Microsoft managed to poison their own well. The crap about them setting up development in Canada because it's too hard to import enough H1B's never would have happened 10 years ago (for various other reasons too).
I bring this up because open source seems overly eager to poison its own well. In fact, in a lot of ways, I think it already has done so. Past essays have been on essentially this topic. Anyway, we're highly symbiotic with the commercial sector. Just as future Microsoft systems developers needed to learn on Unix as a sandbox to be able to effectively develop closed source, there are places where we need closed.
Jobs, for one. Very few people get paid to do open source, still, though more and more get to release some of our work some of the time. A pure support model is great for going after corporations, but most money is in the consumer sector -- selling copies of Windows and Word, and various games. Places where costs don't amoritize over huge numbers of users is another -- medical information systems, such as are used by major clinics, is a battle ground for three companies, and the intricies of the business and products are volumous and yet unknown outside of those circles. The defense sector has driven some amazing development. One ongoing complaint about Linux and X is the GUI lag that doesn't exist on Windows. Linux speaks X over a Unix domain socket to the X server and all messages get queued together and get processed in order. Well, work was done, in the defense sector, for real time displays that also displayed huge amounts of information, to add a priority queue to the protocol and get a multithreaded, multi-CPU X server processing requests out of order. Not only has this solution been lost to us now, but we're ignoring the blue sky development budgets that made this kind of large scope work common in the old days. Microsoft knows about these markets and is vigerously going after them, even with the handicap of their reputation, and the handicap of Unix's reputation. The defense sector hates releasing code, and commercial Unixes gave them what they wanted, but now the commercial Unix providers are struggling in competition with Linux. Some may rejoice at DEC and the old guard going down. I'm just saying, complex interdependencies, symbiosis, etc. That's all. Before we celebrate killing anything, or even putting the squeeze on us, it's worth thinking about exactly what it means to us -- what even more insideous pest species will make its evacuated niche its strong hold.