I have developed a working hypothesis on software development. When the lines of code exceed 100,000 and the number of arbitrary constants approaches 50,000, no man (or woman, or group of programmers) have any idea about what the program is doing.
that is *precisely* why open source (and Linux in particular) has a huge advantage over Microsoft, and to a much lesser extent, Apple -- and as time goes on and software becomes even *more* complex, that advantage will accelerate.
Because things are in fact so complex, you need as many eyes as you can to look at things. Even with a $6B budget, 8000 pairs of eyes is not anywhere near enough...
if you analyze how the open source community is structured, you see that the effort really transcends international boundaries - not only American, British and Finnish eyes are developing and improving open source, but pairs of Hungarian, Korean, French, Brazilian eyes are looking at the same code and contributing what they can to making it work, all completely out in the open and subject to review by their colleagues.
the economics and motivations of those who contribute to open source are complex - many companies, such as IBM, use the software developed under open source framework to sell hardware. many contributors to open source are trying to establish a reputation to get a better job, others do it because they have picked up the skill at work and are paying back the open source community for times that the net and the community helped them save time on the job...
the bottom line is that software is *so* huge an complex these days is that it requires eyes and contribution from as many people as possible, no matter where they come from...