- From: Garret Wilson <garret@globalmentor.com>
- Date: Tue, 01 May 2007 16:31:58 -0300
- To: Kjetil Kjernsmo <kjetil@kjernsmo.net>
- CC: semantic-web@w3.org
Kjetil Kjernsmo wrote: > No, that will not be required if there exists a full vCard ontology in > RDF. > > >> But I think it's dangerous. >> > > Why? > ...because we might as well rename that section of the specification, "Information you are allowed to throw away." :) But ignoring the multiple namespace issue (below), this isn't a big sticking point to me. As long we don't encumber processors (like mine) that want to handle all the data appropriately. > >>> >>> ...and as long as the subsets of >>> the core and full are disjoint but the union of them makes up the >>> whole vCard, there is no variations to talk about -- there are just >>> two namespace URIs instead of one, that's all. >>> >>> >> Wait... multiple namespaces! No! That's like having two math >> ontologies: operatorPlus in the example:easy/math/concepts# >> namespace, and operatorDerivative in the >> example:difficult/math/your/processor/probably/can't/handle# >> namespace. >> > > Hehe, well, I don't see that as evil, if perhaps a little excessive in > that example. > > I see it more like the division of OWL into OWL DL and OWL Lite (OWL > Full vs OWL DL is another story), SVG into SVG Tiny, etc, and many > similar specs. There is a lot of precedence in doing it this way. > I'm not sure if I agree with the precedent. But besides, these are totally different types of subsets. We're not talking about understanding and doing something with all the information, which is what OWL/OWL Lite and SVG/SVG Tiny does. That is, it would be analogous if we said, "vCard RDF processors must be able correctly to determine which telephone number to use when asked; vCard RDF Lite processors need only be able to send email to the correct address." We're talking about hcard/vCard/RDF conversion here. OWL/OWL Lite is not about converting other semantic metadata frameworks to/from OWL, and SVG/SVG Tiny is not about converting other vector formats to/from SVG. They are about following the instructions inherent in OWL/SVG. (This is why I made the semantic understanding vs. syntactic understanding distinction above.) But again, I don't really care if you dice up vCard RDF as many ways as you want---I just don't think giving each subset a separate namespace URI is a good idea. Garret
Received on Tuesday, 1 May 2007 19:32:30 UTC