Lists Home |
Date Index |
-----BEGIN PGP SIGNED MESSAGE-----
/ "Simon St.Laurent" <email@example.com> was heard to say:
| Jonathan Robie complains:
|> Is it wrong for XML to have ID/IDREF attributes, with associated
|> semantics, and allow attributes to be declared to be of this type?
| I'd say it was, yes. I'd have left that to an application to process,
| not built it into XML parsers.
So in this world view, we have nothing more than labelled trees.
I suppose that's fine as far as it goes, but it doesn't go very far.
I work with lots of vocabularies, some of them fairly complex, I find
it tremendously valuable to be able to describe a set of constraints
on the labelled trees that I'm willing to process. I find it a
thousand times more valuable because there's a fairly standard way of
encoding this set of constraints so that all the people who use the
tools I write can also test their documents. In particular, this means
I can punt broken documents back to authors and say "no, that's not a
bug in my tool, that's a broken document."
I suppose I could have done that all with my own application by
writing a SAX filter or something or done it all in a more
XML-friendly language like XSLT, but I'm very, very glad I didn't have
Similarly, I'm glad I can write <p> in a configuration file and <p> in
an XHTML document and have a generic mechanism for disambiguating the
two different 'p' elements. Could namespaces have been described
better? Probably. Am I willing to live with the current spec? Yes,
usually. Are namespaces being used in ways that I think are awful
design? Yes. But can you name a design that hasn't been used in a way
that someone considers awful?
Be seeing you,
Norman.Walsh@Sun.COM | "Abstraction, abstraction and abstraction."
XML Standards Architect | This is the answer to the question, "What are
Sun Microsystems, Inc. | the three most important words in
| programming?"--Paul Hudak
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.7 <http://mailcrypt.sourceforge.net/>
-----END PGP SIGNATURE-----