[topicmapmail] multiple Source Locators

[Lars Marius Garshol]

* Jan Algermissen
| 
| Just to make my point clear, what I intended to emphasize was the
| possibilities of improved interoperability of Topic Maps with Web
| architecture, not so much with the RDF graph.

That sounds very strange to me. How does not using subject locators
help with that?
 
| BTW....
| Speaking of the benefits of Topic Maps over RDF, I see basically two:
| 
| - Topic Maps are capable of representing the fact that several URIs
|   refer to the same resource at the level of the abstract information
|   structure, while with RDF, you need to make use of additional
|   semantcis (owl:sameAs) which propagates all the way up into the
|   queries written in your applications.
| 
| - RDF's statement reification mechanism does not 'induce' subject
|   identity, meaning that two nodes that are the origin of the same
|   combination rdf:Subject, rdf:Property and rdf:Object (that represent
|   semantically the same statement) will not merge and are explicitly
|   not intended to merge.  See [1] and [2] on this.

These I agree with, but they are by no means the only. I'd add:

 - topic maps distinguish between URI references used to actually
   reference an information resource (or a representation thereof, if
   you want to be picky) and URI references used to identify things
   that are not information resources

 - topic maps support reification (in a standard way that does not
   cause data bloat and works well with tools),

 - topic maps support provenance, context, and qualification through
   scope,

 - topic maps support n-ary relationships, and

 - there is a methodology for how to use topic maps to achieve
   findability, something that is lacking in RDF.

There is probably more that escapes me right now.
 
| IMHO, while RDF seems to (at least partly) have been intended as a
| data integration technology, there are quite some obstacles when
| querying an RDF graph that is the result of merging several other
| graphs with slightly differing semantics. Topic Maps provide a far
| more elegant mechanism I found.

Yeah.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >