Joel has some observations on the Mozilla project, now that they have shipped their first major release. Joel's really big idea: design doesn't scale to large numbers of people, and certainly doesn't scale to large numbers of volunteers flung all over the globe. (Translation: this is why open-source-as-pixie-dust doesn't work for big projects just starting out.)
In this context, Joel is talking about primarily about UI, but also about software architecture. His comments are equally true for language design.
I think in both projects (Mozilla and Perl 6) the key is that little progress was made until the projects attracted so many developers as to implode under their own weight, causing many people who didn't need to be there to get out of the way (I was one of those for Perl 6), allowing the faithful remnant to do what needed to be done without interference.
That's not a flame of the process or any people, either. I think it's a recurring theme we're going to see more and more with large, high-profile open source projects in the future.
