I've written a class
foo.pm that collects some data from a web service. Inside
foo.pm I have added perldoc to describe the functionality as well as a short guide on the usage. There's a reference for all its methods.
I've also written a subclass
cachedFoo.pm, that uses
foo.pm as a base class, wraps its own constructor around
new method and upgrades
foo with a database connection to cache the results. I've already added perldoc to
cachedFoo.pm for the additional stuff.
Now I only want my colleagues to use
cachedFoo.pm. Should I copy the perldoc for all the inherited methods from
cachedFoo.pm or should I just say "look at the base class's docs for the accessors"? Or is there maybe another way?
It suffices to conspiciously point to other classes once near the start of the interface description, see best practices documentation example below. Your pod coverage test then should employ Pod::Coverage::CountParents to take inheritance into account.
All methods and attributes not mentioned here are
inherited from L<foo> or mixed in from L<somerole>.
Blah blah blah, Mr. Freeman
Naming each composed method explicitly does not scale. I can't recommend this: