I'll take a bet against MSFT.
=====================
And you'd be smart to do so.
Microsoft faces a big problem. The problem isn't Linux (which flourishes by pretending the Microsoft system doesn't exist -- by building an alternative development and distribution complex) or Sun Solaris or IBM's AS400 or Apple's OSX or it's kissing cousin FreeBSD or any of its other competitors who, right now, are all healthy niche players.
The problem is it's own software or, to be more precise, the complete everywhere-ness of past, present and, in the case of beta releases of its next OS, Longhorn, future versions of all flavors of Win32 and other MSFT apps.
People and companies are beginning to ask, 'where is the benefit to me of *upgrading*?' This problem is particularly acute with Microsoft's mega cash cow, the Office productivity suite. There are only so many ways you can type a letter or use that small subset of the spreadsheet application's functions most users know about.
Office 97, 2000, XP and now 2003, despite numerous differences between them (some interesting, some absurd, some cosmetic, some quite real), are not different enough to convince customers large and small who've already mastered one version to fork over capital for the serious initial purchase costs and perpetual licensing fees Microsoft extracts.
This problem -- this ubiquitousness of its products and, therefore, hair pullingly frustrating competition with earlier versions of itself -- is compelling Microsoft to engineer and enforce, step by step, a de facto subscription model for their software where you're never done paying.
They're also on a patent issuing binge that is probably linked to concerns about future revenue growth.
And then there is the problem with the .Net development platform which was supposed to be the foundation of an entirely new application programming interface.
MSFT is beginning to face an odd sort of highly liquid stagnation. This is still in the early phases and may branch off in unusual directions but for those of us in the trenches of network engineering and dev work it's starting to become ever more apparent.
.d.