Daniel Berger has just published this very article: RubyForge vs CPAN. His conclusions (previsible due to the Ruby bias) were basically summarized by one of his commenters: The differences are "not as extreme as one might assume". I did not dive into the article yet, but it looks like it is founded on many hasty estimates/comparisons in many places.
CPAN vs Rubyforge
kjones4 on 2007-09-19T19:10:54
My impression after reading the article was that the author was surprised to find that Rubyforge didn't suck as bad as he thought it would. But frankly, I don't care whether CPAN is better than Rubyforge or vice versa. I would have been more interested to learn where one repository could adopt best practices from the other, or compare how one might perform the same task with both systems.
