This and this made me wonder today. Now, what do Google antispam measures and Daisy have in common? Well, they both somehow came out of the cold, with just very little discussions happening in closed circles, and they both are community praised as successful moves.
Looks like open processes are unable to produce successful stuff. Compare Google’s move with e-mail antispam measures: all of a sudden ten companies gathered around a table and more or less overnight handled us a solution (as questionable as it might be) on the blog spam issue. Meanwhile, standard bodies are still frantically trying to reduce the spam load on our mailboxes. Compare the ultra-proprietary Daisy repository, courtesy of a small team of very smart guys focused to build a practical solution starting from a clean slate, even is that means don’t giving a damn about standards, with JSR170, where political issue have substantially delayed and, while they were at it, diluted quite a bit what could have been a very useful standard.
There are much more examples (JDO vs. Hibernate and Linux vs. Hurd just to name a few), but it all boils down to how difficult can be bootstrapping a successful open process, while the open system seems to work really well in picking up from an existing (and sound, mind you) foundation. Starting from scratch in the open tends to lead to bikeshedding, political fights, delays and failure, JFDI and then start the open route seems to work. Point taken.