The applications market is heading towards an environment where a given application will run on a range of operating systems. This is good for application developers, because it expands their market, and is good for consumers, no longer forced to have a certain operating system to run a certain application. This evolution is heightening competition primarily in the application and operating system markets, as attention is refocused upon what merits each product has in its own right.
Parallel to this, also evolving are CPU-independent operating systems, where the CPU does not dictate choice of operating system. This is good for operating system developers, because it expands their market, and is good for consumers, no longer forced to have a certain CPU to run a certain operating system. This is heightening competition primarily in the CPU and operating system markets, as attention again is refocused upon what merits each product has in its own right.
Each of these threads are part of the component-based open-systems tapestry, where each element of a computer system is completely homogenous (interchangable). Given that competitive intensity in each market is increasing (as products become increasingly interchangable), incumbents must innovate if they are to retain marketshare.
Innovation, however, takes curious forms. Microsoft, for example, are more talented at marketing than at writing good code. But then, it's Microsoft's destiny to write buggy, inefficient software - the closed-source approach produces no other kind. That's why scientists bother with "peer review" - it's so they don't fuck up.
Let's dwell on this for a moment. Microsoft has been busy using applications to sell its operating systems, most memorably with Word for Windows and Excel. MS-DOS was successful not on its own merits, but more because it ran Lotus 1-2-3 and WordPerfect. MS-DOS itself was not written by Microsoft, but purchased by them and licensed to IBM and later, the clone-makers (such as Compaq and Dell). What does this tell us? It tells us that Microsoft achieved its success not through clever code, but through clever marketing. But more than that, it tells us that the probability of them developing the ability to write clever code is low. Not only is this because they didn't have it in them to start with, but because since then, they have enjoyed a steady income from their monopoly position; this has let them become fat and lazy. It's not a coincidence that their software is like that too.
However, despite Microsoft's predatory market practices, unix has flourished. The open-source genie is out of the bottle, and no amount of legal legwork by Redmond can change that. Microsoft cannot compete with the open-source model; intermittent service packs are no match for constant development, by teams of thousands, worldwide. Ultimately, the open-source community will produce all of the software required to use computers, and it will all be free, feature-packed, fast, and small. This is simply because they are improving it all the time, and there are so many vested interests, they cancel each other out.
As each day passes, Microsoft's domination of the operating system marketplace is working against the company, as future markets are highly competitive, while the corporate culture retains its monopoly mentality. This means that Microsoft is becoming less and less effective - in everything. If they can successfully evolve their culture to that more suited to a monopolistically competitive marketplace, they will get away with a drop in marketshare and market power. If they cannot transform their culture.. they will go out of business. In either case, Microsoft will never again enjoy the trappings of monopoly it enjoyed with MS-DOS and Windows.