My foremost misgiving with Microsoft: it is not a software development company: it is a marketing company before anything else. From its 1981 "creation" of PC-DOS for IBM, in reality an already-existing OS bought from Seattle Computer Products, Microsoft adopted the marketing method to which it owes its fortune: getting their software pre-installed on many computers as possible. They did the same sort of deal with IBM-Clone manufacturers with their re-branded MS-DOS software.
Pre-installing OS' on computers has two advantages: it a) assured massive profits for Microsoft even before computers hit the shelves, as the OS shipped with the computer at no visible extra charge to the buyer; b) most importantly, assures Microsoft a captive audience that will become dependant on its product: first-time computer users will "learn" whatever they see in front of them the first time they turn the computer on.
As Microsoft managed to ensure itself a large part of the PC market from the get-go, cross-platform compatibility problems resulted in a network of users that would have to rely on Microsoft's product to communicate 'seamlessly' with each other. Their large market share created another level of dependency on their product: many software developers made the economically-sound decision to create products only for the Microsoft platform. The number of developers making this decision only increased with Microsoft's market share.
Microsoft's marketing scheme in itself is condemnable: it exploits consumer ignorance. Fortunately today there are other alternatives to Windows, but unfortunately, because of Microsoft's enormous market share, these took decades to develop to an equally performant level; only recently do Microsoft consumers have an equally tantalising 'ease-of-use' (at least to their knowledge) alternative, at the same time that Microsoft's own bumbles (Vista) driven its users to seek better, often less expensive, alternatives. Old habits die hard, and Microsoft had a decade's advance in ingraining these into its consumers.
Like I said, the majority of today's Microsoft consumers are well aware of Windows weaknesses and alternative OS'; for this it is no coincidence that Microsoft is setting its sights on developing countries. If I can permit myself a moment of ironic digression, I see a parallel with the behaviour of the Catholic church in face of its dwindling flock of Western-world faithful. Yes, it's all about indoctrination.
Had Microsoft used its lead to make itself the best product in the market, all the above would be (more) forgivable: instead they used. and still use, the majority of their enormous profits for... marketing: patent protection (any idea even not their own, ideas not even in development), advertising, and even underhanded techniques like spreading fake hype (paid bloggers and comment-leavers) and FUD. Microsoft's profits are so huge that all of the above expenditures don't even dent them, yet the company is still unable to create a functional and secure piece of software. Patents and licences themselves are part of the problem: Microsoft has painted itself into a patent-protected faulty proprietary-software corner. Another thing that Microsoft seems to forget: the end user doesn't care how things work "under the hood", but they are easily discouraged when they are made aware that their OS' engine is vulnerable (without ever really understanding completely why). Mac is that mustang with the bold racing stripes (and a cool dashboard) that just works, and that's all the end-user needs - or wants - to know. The end user wants to use a tool to get another job done; having to learn how the tool works itself in order to get a job done, or having to adapt one's work habits around the tool itself, is an unnecessary additional step that many users can do without. Moulding the user around the tool has always been one of Microsoft's major goals, and this is why many of its users are loathe to change.
But to return to the main course of my argument: software development isn't Microsoft's principle trade. This is not the case of other developer companies/communities such as Linux and openBSD; Apple, whose OS today relies on a *nix engine, is difficult to place in the grouping because they also specialise in hardware development. Although most every analyst is comparing the above through statistics based on "market share", I see this analysis as skewed; Linux is free. Linux may have less "value", but today they are quietly running, with the help of Apache, most of the world's web servers. Linux development is constant, community-based, and, because it is open-source, anyone can compile a program for it. But I digress.
I'm sure that you can see now that the companies whose future I do believe in are *nix and Mac. Mac is a case apart: their speciality is, and always has been, the "common user", and all of their development, at least in the "Jobs years", has been wowing/making a simpler/better OS experience for the same. Jobs' innovation seems to be the result of a constant study of how users work today, how they can work better (without being thwacked into a totally new set of work habits), and how the advances in technology can be used to attain that goal. Mac has recently gained an enormous market share with its iPod and iPhone innovations; this was possible largely because to the consumer eye they were platform-independant tools, although apple provided an easy "link" to many platforms with its (sometimes controversial) iTunes music program. Mac puts bettering a user experience before anything else; this in itself can be considered a marketing scheme - but an irreproachable one. Jobs is at the centre of all this, and I won't hide that when I think of an Apple without him, I worry for that company's future as well.
The above is why Microsoft has always dogged in Mac's innovation footsteps - it is a company focused on what a consumer will buy (or how he can be made to buy), before anything else. Take windows (second to adopt the mouse, "mac-like" gui, amongst other features) and the Zune for examples of coming in second (or in the latter case, almost last) innovation-wise. Because they didn't come up with the idea in the first place, they are forced to do it "differently"; because of their patent/licence restrictions, they have to (reverse-)engineer their "new" product from scratch, and a "different but the same" look design-wise usually ends up in failure if the result is not up to par with the widely-accepted product that (cough) inspired the design in the first place.
I think my earlier Catholic-church analogy was a good one: the more educated we become, the more choices we have; both Microsoft and the Vatican rely on the lesser-educated/indoctrinated (and loathe to change) for their income, and this is why both are setting their sights on developing countries today. Not only is the west better educated about its choice of tools, but its communication/community habits have changed as well: we no longer have to rely on salespeople to tell us what's good or not, and a product's real quality, no matter the magnanimity of its ad campaign, can be exposed and spread to all within a number of days. Vista was a prime example (victim) of this.
Microsoft has little chance for continued success in the environment described above. If they had any b*lls, they would switch their sights to making a computer experience better for the end user, just as Mac does. This would involve a total change of ethic, staff, and branding - but the most major change by far would be their scrapping their buggy/virus-vulnerable over-patented core - just as Mac did. I don't see this happening anytime soon, but I really hope they use their still-large lead and still-huge resources to a good end.