On Thu, 15 Aug 2002, Stephen Adkins wrote: > Javadoc works well because it actually examines the code and > supplements the code with comments (the "business logic" layer, > associated with determining metadata about the code). > Then it creates HTML in a very nice format (the "presentation" > layer, associated simply with presenting the metadata).
I agree, this is a cool idea. I for one would like something where I could feed docs into Mason components for rendering. Others will want XSLT, etc. > I have begun a module which I might release independently on > CPAN which extracts metadata from code, comments, and pod. > I have currently named it "Devel::Metadata". Looks like a good start. I would suggest also creating a basic "Devel::Metadata" commenting standard that would be used in preference of actually scanning code so one could simple do this: # Parameters: # # - foo # type: ARRAYREF # description: The things that will be foo'd upon # # Throws: # # - Exception::Foo::Bar::Baz # A Baz in the Bar blew the Foo Also using Module::Info as Tatsuhiko suggested would be a good thing. -dave /*================== www.urth.org we await the New Sun ==================*/
