Just read in gnat's journal that he's thrilled about the sorts of conversations he's hearing around the conferece. Don't be fooled! He's just amused by hearing me say "crap" to Eric Raymond. :)
FWIW, the conversation was about the use of metadata tags in documentation. Eric's trying to rally support for a rich, common documentation format. (The NIH quote was his) I'm all for that, but unfortunately it has two problems:
- It's tough enough to get folks to write documentation now. Raise the bar, even by a little bit, and you'll see an exponential drop-off in the number of modules and suchlike things with any documentation to speak of.
- Metadata in text documents generally doesn't work except in reasonably controlled environments. Which we don't have. It's not as bad as, say, relying on the metadata in web documents (which is downright mad, as some people will lie left, right, and center if they think they'll get an advantage from it. That number doesn't have to be very high at all to render your system worthless)
Eric's main premise, "presentation level markup must die!", is a good one. What we need to do with pod in perl 6 is to go to an entirely semantic markup system that
looks like presentation level markup. Which is to say, when people don't pay attention and use markup like it's presentation, they really use it properly semantically. (Which is going to be a good trick if we can pull it off)
