A key Internet value is low switching costs. (Well, no switching costs.) So when the costs of switching are low or non-existent, an open approach is going to win out over time.
(Amazon sells these wooden Maxim train switches for just $9.99. For two! How do you like those low switching costs.)
Google groks this. The secrets to their success are open. Everyone who wants to knows about their use of low cost PCs, about their dark fiber purchases, about their concern over energy costs. Many of their key algorithms are also open.
No, not the "secret search source" -- but having that would just let outsiders manipulate results. Open is not a suicide pact.
Facebook also groks this. Its opportunity came when rival MySpace was bought by Fox, which wanted to tie the social network's audience to its brands. With switching costs being zero, Facebook began benefiting immediately. They learned the lesson.
Switching costs are too low for any proprietary embrace to work. Open source credibility matters more than anyone imagines. So open must be the mantra, even when introducing something no one else has.
Traditional business analysis holds that switching costs are never zero, and that they are constantly rising. This is not true on the Internet, and this benefits those companies that are most open.
This idea of no switching costs does not exist (yet) in the world of Internet clients. I admire what Jason Perlow is doing, switching entirely to Linux from Windows, but the market won't follow because that entails high switching costs. Apple would likely benefit more from a rejection of Microsoft than Linux, despite its platforms being highly proprietary. Switching costs matter in that space.
But those switching costs can be reduced. What Google is really doing with Android is providing an iPhone alternative that is not only competitive, and similar in terms of usability, but also has open written all over it.
Over the long run a combination of being open with low switching costs is going to win. The question, then is how long that run is. After all, as John Maynard Keynes said, in the long run we're all dead.