* Russ Allbery ([EMAIL PROTECTED]) [13 Oct 2002 15:36]:
> Iain 'Spoon' Truskett <[EMAIL PROTECTED]> writes:

[...]
> So the idea is to instead have a registry of XML-style names for
> attributes and then people can map the available single characters to
> the appropriate attributes in each POD file?

I'd say so. We've got 26 letters, fewer taking out the pre-reserved
ones. And to be honest, I'd be happy redefining some of those (I never
use some of them so would want to be able to redefine them so something
more practical; I could have S<> as method-name [erm, subroutine-name]
rather than the non-mnemonic N<>).

Plus, I'm thinking of my perl books site. Reviews written in POD. Thus
it would be nice to be able to have things map to book-title,
author-name, etc. Obviously purely intended for my use, but Pod::Simple
should be able to map them appropriately.

> > Hardly a mini-language. It's more a gracefully degrading syntax.

> It's a gracefully degrading syntax if we all agree on what M<> and N<>
> mean, but if M<> and N<> may mean things that are entirely different
> in each POD file, it feels to me like we're saying that everyone
> creates their own specialization of the POD language, which does feel
> like a mini-language to me.

A specialisation which eventually maps to the generic. We get to have
POD mean what we actually mean. POD is not perfect. It can't be all
things to all people, but with =extend it can mean more. With any luck
it will result in more semantic markup rather than physical markup. See
HTML 3.2 vs XHTML.

[...]
> I'm currently talking to Sean about Pod::Simple, since I took a look
> at it today and it didn't feel like it mapped anywhere near as well as
> Pod::Parser did for me.  *wry look*  I was hoping I'd fall in love
> with it, but I guess that didn't happen with Pod::Parser either.

People think in different ways =) I found it mapped well, as compared to
Pod::Parser which I do use (since Simple wasn't around a year or two
ago) but only minimally.

> I'm just not all that fond of the XML document representation model.
> I much prefer something that's more native to POD with a paragraph as
> a primary object.

Forget about XML. It's a misleading thought really. The representation
is akin to XML, but it's not real XML (thank heavens). It's a free form
tag based hierarchical notation. No namespaces, or DTDs, or crap like
that =)

Could be interesting seeing the structure output as YAML. Hmm. /me notes
that one for later.

You get notified when a paragraph starts/end and can easily just
construct a paragraph starting from the start tag and then postprocess
the paragraph when you get the end tag. By the time you've got the end
tag, all the interior sequences have been processed, so you just have to
worry about making it look nice =)


cheers,
-- 
Iain.

Reply via email to