=head1 DESCRIPTION
...
=head1 Qwertyuiop (DESCRIPTION)
...
And then someone wanting create an index where both DESCRIPTION and Qwertyuiop are recognized as "descriptions". All this makes sense of course only for applications that have defined some semantics for some particular sections.
Okay, that makes sense. Nice and straightforward.
So then the new logic for the "What's the title method" would just need to tolerate "=head1 blah blah blah whatever (NAME)" as a variant of "=head1 NAME", i.e., both meaning "the next paragraph will be the short title of this document".
Then I could make a "What are ALL the titles" method that wouldn't just stop at the first NAME/blah(NAME) thing, but would read the whole document looking for all such things.
And, in fact, I could make these just specific ('NAME'-specific) cases of what's-the-first-FOO / what're-all-the-FOOs methods, for arbitratry values of FOO (like 'NAME', 'DESCRIPTION', 'AUTHOR', etc.). Sound good?
-- Sean M. Burke http://search.cpan.org/~sburke/
