Longhorn, whatever its eventual name (probably something boring like Windows 2006) won't be the turkey some people hope for. They are spending a lot of money and manhours rebuilding Windows from the ground up, more or less. Many of the problems Windows has today are legacy. Longhorn is almost as drastic a change for Windows as OSX was for Apple. Actually, the OS9 to OSX transition is perhaps more akin to moving from Windows 98 to Longhorn. It's a totally different codebase. Except in the case of OSX, although a new Unix distribution, it has a long development history behind it at its core, most notably the 10 or so years of NeXTStep, but also from the whole BSD community, and all the exhaustive security work done for NetBSD which I believe has never been successfully hacked.
Microsoft isprobably reusing or rejigging chunks of existing code, but essentially it's all new (and untested). Currently with XP, they are constantly patching the leaking security holes, patches upon patches, because the underlying Windows architecture wasn't designed with security in mind, so even extensive changes from the original NT codebase won't help with the underlying problem. Longhorn is being rebuilt with security very much in mind, so their chances of getting good security for once is much improved, although they have the problem of any new unproven system.
I doubt those are the actual system requirement, however. I seem to have read those specs first as an indirect quote from Bill Gates that those are the sort of specs he EXPECTS are common by then, not what's to be minimum requirements, so Longhorn will be geared towards the near future somewhat. They're copying the concept of farming all the eyecandy rendering off to the graphics card, and considering how they've developed, and what they can do now, you shouldn't need anything much faster than the current top systems. At least I hope not. Microsoft doesn't have a good record here. The hardware industry very much rely on this bloat effect to drive upgrade cycles. A programmer friend of mine once said (a few years ago) that if Windows and MS software keeps growing at the rate it is, it will soon consume the entire solar system.
When I first saw the list of features for Longhorn, I had a definite sense of Deja Vu, because it was like a list for OSX from when it came out, plus some that seem 'borrowed' from BeOS, like the filesystem as a database concept, and the better use of Metadata. OSX could do well to implement a lot of BeOS's filesystem features, too. They employ some former BeOS filesystem engineers, after all. That's who designed the OSX journalling system.
But don't be fooled, it will be a very sweet OS, and probably better looking than XP. The 'Slate' theme looks fine to me, but the final will presumably look very different. But if the system requirements truly are that great, it will take a while to adopt this new system. Business certainly will just for the extra security (they have too much investment in Windows, and according to a recently leaked MS memo to Bill, they've been fully aware of this proprietary capture in Windows APIs for years and exploit it, they even acknowledged they would be dead without it).
But I'm sure OSX won't be standing still either, and Apple will pay close attention to Longhorn. But people will still go with what they know, like sheep, buying the MS propaganda. They'll have a release akin to the Windows 95 campaign.
Maybe Macs won't seem quite as expensive once the minimum requirements for Windows blows out.
