CPAN is wonderful thing - it is a great achievement of the whole Perl community and the thing that makes it apart from other languages. But somehow it works mostly for libraries. And what about the other modes of code reuse? How about CPAN for code examples? One can say we don't need CPAN for code examples - everyone is using code examples all the time and they do that without any special infrastructure. But what I see around is that some of CPAN modules are actually code examples in disguise. Some of them would be better without any code in the package and only documentation and tests validating the documented techniques. Hmm maybe we don't really need a separate CPAN for that - maybe we can keep them at the existing one - and just mark them in a special way? And maybe we need some social acceptance for packaging such 'empty' libraries?
And what with applications?