Alternatively, is anyone astonished that a number so arbitrary (in either direction) is considered a serious point of evidence? Not that there's a great way to weight the importance of programs in isolation from each other, but [random shareware X, even if business related] doesn't match the importance of, say, a major word processor in the real world. Isn't the flexibility that a given OS offers to create new programs, and the rate of change in the number of available programs, more important than the existing number anyhow?
Updated 20:05 by timothy: PJ Doland, Webmaster of the Cato Institute writes: "The article on Microsoft that you mention is now online at http://www.cato.org/pubs/pas/pa-380es.html. I'm a slashdot addict. When I saw your link to the Institute's site, I convinced the higher-ups to let me post the report early for your readers. The full report is in PDF."