Ease of use does matter.
Of course it does. It even comes into play in several aspects of this case. It just doesn't come into play where you suggest it should.
The judge has (erroneously) found that the existence of applications leads to the primary appeal of an OS to the consumer.
Let's step away from the facts presented in this case for a moment to answer this question: How does one account for the popularity of MS-DOS in 1988? Let's assume that ease of use is the consumer's primary consideration when buying a computer. Why would anyone in 1988 have bought a machine running MS-DOS when Macintosh, Amiga, Atari ST and other machines were available? Do you really wish to argue that a DOS machine was easier to use?
Maybe there's some other factor like cost at play here, but even if consumers had a different and compelling reason to buy IBM-compatible machines, why did they spurn shells like Windows 1 and 2, or GEM, or GEOS in favor of the black DOS screen?
Could it be that applications like Word Perfect, Lotus 123, and dBase caused consumers to stick to a more complex UI despite the existence of more consistent alternatives? Microsoft Excel on the Mac was easier for a novice to use than was Lotus 123 on DOS, but Excel and the Mac never gained market dominance among spreadsheet users. Some factor other than ease of use must be at work here.
...MSFT's first and most important hurdle was to deal with the dearth of applications for Windows in the first place.
Exactly. And that demonstrates the power of the applications barrier to entry. Windows gained significant market share only when applications were available for it that presented a compelling alternative to DOS applications. It was a slow process. Its great advantage, however, was that Windows did not require the user to abandon the dominant OS of the day with its far richer universe of applications. A user who preferred PowerPoint on Windows over Harvard Graphics on DOS, but liked the classic clean-screen DOS-based WordPerfect 5.0 could use both on the same machine.
Consider the issue from a different perspective. If the applications barrier is relatively unimportant, then why do virtually all OS designers spend so much effort and time in assuring backwards compatibility? (Only game-console designers ignore it and the exception illustrates the point.) Apple provides all-too-many examples. When they moved to a different processor for their machines, they and their users accepted many widely-recognized compromises in speed and utility in order to maintain compatibility of the new systems with the most widely used applications. Similarly, Microsoft has accepted delays and expended massive development effort on Windows 2000 in order to achieve high compatibility with existing NT and Window 9x applications.
If, as you suggest, the judge is wrong in emphasizing the importance of the applications barrier to entry, then Microsoft is wasting its money. It shouldn't worry about existing applications, but should concern itself only with making a better OS. If there is no applications barrier to entry, then consumers would accept the new OS on its merits alone. But that doesn't happen. The applications barrier to entry is so strong and so significant that it severely constrains even the maker of the OS. |
