An article yesterday on ZDNet argued that Microsoft might soften its stance towards open source, and may even write applications for its famous open source competitor, Linux.
A few years ago, I would have been more skeptical. Making product for Linux might help to popularize the alternative platform, thus cutting into Microsoft's own operating system revenue. From a simple cost-benefit analysis, the cost of writing for the fast-growing Linux operating system market, in terms of lost sales, outweighed the benefit of sales into the channel.
Today, however, I'm more inclined to agree with Stuart Cohen, head of the Open Source Development Lab (OSDL) and source of the claims which drove the aforementioned ZDNet article. It now seems obvious that open source poses less of a threat to Microsoft's core business than some of its antagonists, and even some of its proponents, previously thought.
What are the odds that France would decide to ditch the language of their forefathers in favor of German? Not much. In my opinion, the odds are only slightly better that those in the Windows technology domain will shift to the Unix technology domain.
Unix tools are different than Windows tools. The Unix way of administering a server is different than the Windows way of administering a server. Unix programming is different than Windows programming. All these add up to a low likelihood of moves between domains, partly due to cost, but mostly due to the fact that the inhabitants of either domain LIKE the way they are doing things (kind of like the French like speaking French) and aren't inclined to change. Proof isn't any harder to find than statistics showing that Linux's growth has come mostly at the expense of other Unix-family OSes.
That doesn't mean that Linux might not take some share from Windows. However, the difference wouldn't be the equivalent of a nuclear assault on Microsoft's market share. That makes catering to the Linux market much more appealing, particularly as costs are lowered as Linux replaces the fragmented Unix market.
Of course, this analysis only holds if users of Linux actually have an interest in buying proprietary software (it's safe to say that Microsoft isn't going to release GPLed versions of its flagship Windows products).
I've argued in the past that price has little to do with the material cost of a product, and everything to do with a supplier's ability to convince customers that its worth paying a certain price. That is clear to anyone who has seen the prices luxury clothing stores charge for a shirt that, at most, costs only a few dollars more to make than a comparable shirt with less media-created cultural buzz. Thorstein Veblen called it "invidious distinction," and luxury goods vendors play it to a "T."
That principle applies as much to the price of luxury goods as it does to paying any price at all. If Linux has cultivated an expectation of "free product" in its users, that will be a hard habit to break. If it can't be broken, don't expect Microsoft, or ANY proprietary vendor, to bother targeting the Linux market.
I think that there is a willingness to pay, particularly among businesses looking for software to run on Linux servers. The consumer market, however, will be the big question mark.
In my opinion, the failure to cultivate a "willingness to pay" could be the single biggest barrier to desktop Linux adoption. If those who specialize in desktop software development (namely, proprietary software vendors) avoid the desktop Linux market because its users don't like to pay for product, then most users will stick with a platform that has all the options, even if they have to pay for them.