Mathematicians will tell you that an inflection point is a point on a curve at which the sign of the curvature - its concavity - changes. Marketeers reckon it's defined as "a changing of the rules of the game resulting in a massive shift in the way business is conducted."

In other words, every time there's a shift in the way things are done, it opens up opportunities. Remember Windows, and how all those once-dominant companies -- even those with excellent products such as WordPerfect, Novell and Borland -- all turned up their toes and either vanished, downsized massively or had to go play somewhere else?

And if Linux vendors weren't rubbing their hands with glee at the prospect of Longhorn, I'd be very surprised. That is, they would have been until the recent news that Microsoft has decided to strip out all the interesting parts of its next-generation OS and instead stick to 'incremental technologies', whatever they' are.

It's all quite simple. Before the removal of WinFS, Longhorn had all the hallmarks of becoming an inflection point. Not as big as the introduction of Windows, I'll grant you, but pretty chunky all the same: a new file system and a new UI, all in the one OS.

Had Microsoft persisted with its original plans for Longhorn, it would make any sane IT manager wonder if the time weren't right to consider all the alternatives. Imagine: in two years' time, say analysts, Linux will be ready for prime-time desktop use. In other words, we might expect that, by the end of 2006, the Linux UI will be polished, the installation routines will be slick and the experience will be -- well, let's hope -- better than that of running Windows.

Project yourself forward those two years and imagine the scene had the new technologies been retained. Longhorn launches amid much hoo-hah but, after the dust raised by millions of dollars of marketing has dissipated, what's left? A new Windows with two untried technologies and many unresolved issues, especially in the minds of those tasked with installing it on thousands of heterogeneous desktops.

Never mind the benefits, feel the pain. At that point Linux, and Windows would be competing directly.

But it is not to be. Microsoft has covered its tracks by saying that it does not believe it can get the technology ready in time, an all-too-plausible excuse, one that Microsoft watchers find easy to swallow.

Linux developers are working on something similar to WinFS, the database-founded file system too. If the two were to go head-to-head, there might even be anti-trust ramifications. Not an experience that Microsoft would care to repeat, especially after so many years' toiling over the WinFS project.

Far better then, from Microsoft's perspective, to leave WinFS to slip out as and when it's fully cooked, and be sold as an upgrade for those who need it. Corporations and individuals will then be able to upgrade to Longhorn -- assuming they wish to of course -- with relatively little pain. There are no legal issues, and there's likely to be increased revenue too: you get more for two separate products than in two in the same box.

But most importantly, Microsoft has averted an inflection point, protecting its future Longhorn revenues -- and indeed the future of Windows -- from the predations of Linux vendors.

It's arguable that the ability to think long-term, to maneouvre around the icebergs of the commercial ocean and so deny the competition opportunities to advance is one key benefit of a monolithic commercial organisation, as opposed to a decentralised collective with little guidance from the centre.

To the sound of Microsoft's bacon being pulled from the fire, open source agnostics might do well to ponder this point.