[Sigia-l] Topic maps and IA

[Lars Marius Garshol]

* Jonas Höglund
|
| This is a very good paper indeed. 

Thank you. :-)

(And thank you very much to Peter, Jonathan Broad, and Gunnar
Langemark as well. It's much appreciated. :-)

| My vision is to build a CMIS instead, using something like TopicMap
| or OWL.

I think your alternatives are really topic maps or RDF. OWL is a
vocabulary, a set of classes and properties. The relationship between
OWL and RDF is kind of like between XML Schema and XML.

| Does anyone have any input to this? 
| * Is it doable? 

It's in fact been done several times already. Two of the topic map
vendors (empolis and Mondeca) offer tools similar to what you
describe, and the open source ZTM also does this. The latter has been
used to build about a dozen public sector portals in Norway.

I don't know of anyone who's done this based on RDF.

| * Is the technologies and the methods behind them developed enough? 

For what? :)

| * Is there enough of skills and tools developed?

I'd say so.

| I also need to better understand the differences between OWL and
| TopicMap. 

| So far, I've not been able to find any compare the two after that
| the OWL standard has been finished.

OWL has been pretty stable all along, so older comparisons are still
accurate.

| How does OWL and TopicMap compare today?

OWL is a vocabulary. Topic maps is a data model. You could use OWL
within topic maps if you had some conventions for how to do that.  The
conventions don't yet exist, but I plan to write my next paper on that
(for either ISWC or Extreme Markup).
 
| Lars, I've seen your answer to the same question dated 11 feb
| 2003. You said:
| 
| "It helps to first understand what OWL is. It's basically the
| standardized successor of DAML+OIL, and is simply a constraint
| language for RDF. So it plays the same role with respect to RDF that
| XML Schema and DTDs do with respect to XML."
| 
| Is this still true? 

Yes.

| My concern is, can I do the same thing with RDF and OWL as I can
| with TopicMap, and vice verse. 

They are quite similar, so up to a point you can. There are
differences in how easy some things are, and how well they support
various things that you might want, but in general they have very
similar capabilities.

| Does one of the languages have any extra features, capabilities or
| constrains to the other?

Yes. (The "Living with RDF and topic maps" paper you cited has the
details on that.) RDF lacks scope, n-ary relationships, bidirectional
relationships, and a theory for how to interpret URIs. Topic maps lack
OWL (at present) and a model theory (basically a mathematical theory
for how to interpret the basic vocabulary). You can get around both of
these problems by creating your own conventions.

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