This is good for the longterm

This is good for the longterm

Two things I think you're missing:

1. Market-leading products have a tendency to be destroyed by something that creates a new, lower-end market, then climbs up. IBM ignored the PC market, assuming it wouldn't damage mainframe sales, as it is a market that's largely orthogonal to mainframes. They hit pain when it became clear that powerful PCs (while not as reliable as mainframes) were Good Enough for most users. Microsoft can't ignore UMPCs, or treat them as a new market segment, as it's all too plausible that they'll become Good Enough for most users; imagine (for the sake of argument) an eeePC with 20GB flash, 1GB RAM, a processor equivalent to today's 1.6GHz Core 2 Duo, and a HDMI port for connecting to your HDTV. Suddenly, if the software on the eeePC is good enough for you, an eeePC is enough; when you need a big screen, you just plug into the TV.
2. Microsoft have existing customers, who are being told that XP will not be available for love nor money in the near future; they don't yet want to migrate to Vista, and now they're being told, "you're not important enough to sell XP to; other people can have it for less than you've ever paid, but you've got to move to Vista whether you like it or not". This is not good for relations with existing customers, and gives them new incentive to look at (e.g.) Ubuntu desktops, using WINE to run their business-specific applications.

So, in the short term, no change - but, in the long term, Microsoft are facing two competitive pressures, one from cheap hardware with free or nearly free software, one from existing customers getting grumpy at not being treated well.

I've the impression that mainframes never did better than nowadays. Perhaps you're confused
with minicomputers that were wiped away by micros? (Though it happened all before my time.)

And people, don't forget that MS is a software company that has a lot of small to medium
business customers in its grasp with (semi-)custom software.