On Fri, Jan 11, 2002 at 05:52:10PM -0700, Sean M. Burke wrote: > So. Saying that Pod's canonical "normalization" is as XML begs the > question of what the XML will look like. > One point on which I am of two winds is this: > > Consider this input: > > =head2 Turn-Ons > > I like harpsichord music. > > > Should that produce something minimal like this? > > <head2>Turn-Ons</head> > <para>I like harpsichord music.</para> > > Or something like this? > > <section2> > <head2>Turn-Ons</head> > <para>I like harpsichord music.</para> > </section2> > > The first way /merely/ recapitulates the Pod, which is fine by me. > But would anyone like the latter? What are the practical benefits of it?
I vote for the former, even though I prefer the latter. The transformation is simple enough to become standard a Pod/SAX filter used only when required[*]. As long as we're breaking the string-of-pearls nature for =head1/=head2, why not change the nested format to be <section2>/<title>, or even a nested and DocBookish <section>/<section>/<title>? SMOP, of course. ;-) Z. *: ...assuming that the first style is actually used somewhere meaningful[**]; otherwise the transform should be inverted. **: I don't think we've mentioned round-tripping Pod through a SAX interface. The minimal SAXification certainly aids this process...