- From: <noah_mendelsohn@us.ibm.com>
- Date: Wed, 30 May 2007 10:08:50 -0400
- To: Norman Walsh <ndw@nwalsh.com>
- Cc: "David Orchard" <dorchard@bea.com>, www-tag@w3.org
Norm Walsh writes: > I think that's just a definition of XML. I'd expect a language to > have some extra-syntactic constraints too: Yes, I would expect it to be some subset of well-formed XML. > a grammar of some sort. No, not in general if you mean an intensional grammar, though certainly very often. I see nothing wrong with defining a language by enumeration of the instances, by suggesting that anything that meets certain Schematron constraints is in, etc. If you mean grammar in that broader sense of: some syntactic constraints we can talk about, then yes. If you mean grammar in the narrow sense of something we can write down in closed form using regexs, EBNF, regular trees (RelaxNG), content models, etc., then not necessarily. In short, I wouldn't emphasize the grammar: I'd emphasize that it's a subset of well formed XML, and typically (though not necessarily) defined so that the instances have something common in form or purpose. Aside: though I suppose you could have a language that is the union of all symphonies and shopping lists, I think we don't need to go there; then again, we can talk about the language of all instances of any sort that just happen to have a version attribute on the root, or an xml:lang attribute on the root, even if we're talking about both shopping lists and symphonies. We can design document management systems that will manage a wide variety of things, if just a few attributes are maintained in common among them. That is an important use case, IMO. -------------------------------------- Noah Mendelsohn IBM Corporation One Rogers Street Cambridge, MA 02142 1-617-693-4036 --------------------------------------
Received on Wednesday, 30 May 2007 14:09:16 UTC