Sharp Blue: The dominance of Microsoft

In reply to my earlier article (in its incarnation as a comment on Meri’s weblog), someone called “dorkSpotter” commented:

Those losers that built that “poor production-oriented interface” called Windows are the richest people on the planet. Oh….that’s right, money is evil.

I certainly don’t think that money is evil. Indeed, my business card says “Richard Baker, physicist, entrepreneur” and my homepage says something similar. However, it certainly isn’t true that markets select technically superior products or services. Instead, they select products and services that are well adapted to the current economic environment (which includes, amongst other things, the current installed base for various other products). These two things may sometimes coincide, but in general they won’t.

This is especially true in the case of operating systems aimed at mass markets. The currently most popular operating system has a gigantic advantage, because it encourages application developers to target it, which means that there will be more and cheaper applications available for it, which in turn encourages more people to adopt the platform. This is the situation in which Microsoft currently operates. In fact, the primary competition for a new release of Windows is older releases of Windows, rather than, for example, Linux or MacOS X. All of which means that Microsoft isn’t maintaining its current position by technical superiority.

In fact, Microsoft didn’t even achieve its current position of dominance through technical excellence. Instead, it got there through outwitting IBM in strategy. Essentially, Microsoft beat its opposition in the 1980s by tying its operating system (MS-DOS, at the time) to an open hardware platform. Competition between IBM-compatible PC manufacturers meant that IBM-clone hardware was substantially cheaper than other systems, which meant that people tended to buy it, and hence to buy MS-DOS. The installed base of DOS computers by the late 1980s and early 1990s meant that Microsoft could weather the fiascos of the early versions of Windows - even Windows 3.1, the release that finally took off, was a pretty poor and confused interface as anybody else who had to suffer its use could confirm.

So why did IBM choose a CP/M-like operating system written by Microsoft rather than a more mature and sophisticated operating system like some brand of Unix? For the simple reason that Unix was developed for large, powerful computers, and desktop computers in the early 1980s were anything but powerful. What was needed was something that was incredibly simple. As desktop computers have advanced, Microsoft has, of course, converted Windows into a more Unix-like system.

It was thus only the emergence of a new niche, some canny business sense, and a fair amount of luck that allowed Microsoft OSs to win out. And there were clearly technically superior systems through this time, from the Amiga 1000 in the mid-1980s to the BeBox in the mid 1990s to machines running MacOS X today. None have been so overwhelmingly technically superior that they’ve been able to overcome the powerful effect of the Windows installed base.

Those that have won out in certain niches, such as the Amiga in video production in the years around 1990, have done so by targetting niches in which the installed base effect has been weak. Consumption-oriented interfaces might be (might have been?) another such niche. It’s certainly possible that Apple thinks so, and is targetting it with such things as iTunes and the iPod.

And, for that matter, Unix is just as much a product of the business world as any of the other systems I’ve discussed here. After all, it was produced by Bell Labs!