- From: Murray Altheim <altheim@eng.sun.com>
- Date: Mon, 04 Jun 2001 18:49:45 -0700
- To: Ron Daniel <rdaniel@interwoven.com>
- CC: Lars Marius Garshol <larsga@garshol.priv.no>, topicmapmail@infoloom.com, www-rdf-interest@w3.org
Ron Daniel wrote: > > Murray said: > [on the subject of defining a mapping from XTM to RDF] > > > And I'm still unclear what we'd accomplish by this. XTM syntax is > > very well suited to its designed task, IMO. > > Murray's note suggested something to me. There are a couple of very > different styles of mapping from XTM to RDF that could be defined. > One is a syntax to syntax conversion. The other is to define how > XTM syntax can be harvested for RDF statements. This seems more > important than the former. A similar 'harvesting' was proposed for > extracting RDF statements from XLinks. > See http://www.w3.org/TR/xlink2rdf/ We apparently think alike sometimes. I'd add to your list the ability to harvest RDF to create XTM. When in Austin, Bernard and I discussed the idea of harvesting DMOZ to create an XTM topic map of it. Then gathering the change logs to alter it (rather than rebuilding from scratch). How about a topic map of WorldCat? I think both directions are potentially interesting, just depending on the form of the existing resources. > But then Murray went on to say: > > > But I don't put the technological cart > > before the horse (as some seem to) and assume that I need a specific > > technology before I figure out what problem I'm trying to solve. It > > might pay to back up a bit and write a requirements document. > > Indeed. So here is one requirement - RDF metadata specifications > need to be able to refer to concepts defined in TopicMap files. > For example, we might have a Topic Map of subject classification > codes, and want to use those as the value of a Dublin Core > 'subject' property in an RDF-compliant metadata spec. This is > a simple matter - each Topic needs to have a URI. Interestingly enough, I'm in the middle of writing that promised spec adding the Dublin Core Metadata Element Set to XHTML. One of my examples does exactly what you suggest: hook <dc:subject> into a topic map, qualified by the topic map. (There's lots to talk about here obviously) I hope to augment plink to harvest the DC content into an XTM topic map (actually, by first creating Linear Topic Map (LTM) notation). http://www.doctypes.org/plink/ > Similarly, if the maintenance agency for a controlled vocabulary > has decided to represent it in RDF-compliant XML, Topic Maps need > to be able to use such subject codes (presumably as Topics, but > you tell me). Works either way for me. So long as everything is addressable, either should work. If there's an address for it, either the resource itself or the subject it indicates can be treated as a Topic. > Third - assume the RDF metadata spec indicates that X is a subject > code for a text. We need to be able to determine that Y is also > a subject code, if Y is a 'Broader Term' for X (using XTM's version > of 'Broader Term'). One practical test of this is to load an RDF > database with info about a lot of files, then query it to list the > files matching subject 'Y'. Having access to a suitable ontology to supply such information is what would be necessary. Some of the preliminary work I've done with Cycorp's ontology would suggest this is fairly simple in XTM, and I'm guessing someone from the DAML community will likely pop up and say the same for them. > Are those the sorts of things you are looking for, Murray? Exactamundo. Murray ........................................................................... Murray Altheim <mailto:altheim@eng.sun.com> XML Technology Center Sun Microsystems, Inc., MS MPK17-102, 1601 Willow Rd., Menlo Park, CA 94025 In the evening The rice leaves in the garden Rustle in the autumn wind That blows through my reed hut. -- Minamoto no Tsunenobu
Received on Monday, 4 June 2001 21:53:30 UTC