- From: Nick Arnett <narnett@verity.com>
- Date: Wed, 2 Nov 1994 09:27:02 -0800
- To: M.Piff@sheffield.ac.uk, Multiple recipients of list <www-html@www0.cern.ch>
At 5:57 PM 11/1/94, Mike Piff wrote: >Do you mean that <section> is structural, but <theorem>, <proof>, ><implies>, <xor> are semantic? Or are only the last two symbol place- >holders semantic? It's not an either/or question, which I didn't make clear. Sorry. Relationships, such as "is a" or "relates closely to," between words or phrases are lexical. Semantics takes into account not only those lexical relationships, but also grammar and other structure, including the semiotic information communicated by layout, fonts, etc. So "semantic markup" implicitly should include structural and lexical tags. In other words, I should have said earlier that you can't have semantic markup without a lexical tagging scheme in addition to structural tagging. Using a tag like <theorem> in a document would seem to cross the fence between the two, I think, which probably isn't a good thing. What you seem to be suggesting is that you'd use this tag when there's an element of a document which "is a" theorem, a lexical relationship. So how would you use it? Writing something like this doesn't seem like a good idea: <theorem>Yadda, yadda, yadda</theorem> Instead, tagging along the following lines is much richer and flexible, since it can express relationships that are ambiguous or inexpressible in flat or relational models. Yadda, yadda, yadda<LEX element="paragraph", relationship="is a", content="theorem"><p> I have no idea if LEX is the right name, etc. Forgive me if I don't dig too deeply into the implementation issues -- I'm quite aware of my own ignorance in that regard. I know that I'd like to see this sort of thing supported; I'm hoping that some real expertise in this domain enters into the HTML development efforts. Nick
Received on Wednesday, 2 November 1994 18:26:58 UTC