Ah, but which should doczar be compatible with, doxygen or javadoc? A fundamental frustration I have with both of these projects is that their syntax is designed to annotate the results of a syntax scanner in their primary language, not to describe a structure from the ground up. I could potentially have worked from the syntax of an established "generalist" project such as NaturalDocs, but most of these were designed to mimic javadoc syntax.

Regarding performance, I've fed it some nasty directories and generation time is trivial. Doczar generates docs as fast as your hard drive can move them.

Thanks for looking!