- From: Joshua Allen <joshuaa@microsoft.com>
- Date: Mon, 22 Apr 2002 20:05:55 -0700
- To: "Danny Ayers" <danny666@virgilio.it>, "R.V.Guha" <guha@guha.com>, "Margaret Green" <mgreen@nextance.com>
- Cc: <www-rdf-interest@w3.org>
> mapping itself can itself be readily expressed using RDF(S) in such a way > that modifications/additions can be made without breaking existing systems. > This isn't really the case with XML Schema (IMHO) because the emphasis is > more on structure than meaning. A tree (I would hazard) is inherently more > brittle than a graph. OK, I understand the gist of the argument. At the moment, it is fairly easy to visually build a mapping between two different XSD schemas, though, which produces XSLT. There are a number of tools that do a good job of this. And my point was that *someone* has to come up with mapping, bless it, and so on. XSLT is technically expressed as XML, so the mapping between two XML instances is expressed "as XML". But I can see some value in having the mapping itself represented as assertions. > I'm pretty sure this wasn't what was being suggested. Yeah, Margaret's follow-up clarified things for me.
Received on Monday, 22 April 2002 23:05:58 UTC