Re: Transforming XML content into RDF assertions

m batsis wrote:
> 
> Michael Denny wrote:
> 
>> If one has a collection of XML documents sharing one or more XML 
>> Schemas and
>> an RDF Schema describing the knowledgebase structure, how would they 
>> proceed
>> to populate the knowledgebase? 
> 
> 
>  * Determine the parts of information to be converted to RDF, unless you 
> really want *all* of the XML to be converted to RDF.
> 
>  * Locate that information in the XML Schemas
> 
>  * Use that to write some transformer (XSLT) to produce the RDF metadata 
> according to your RDF Schema. That should be easy.
> 
> Of course this is just what comes in my mind right now. Perhaps others 
> may be able to provide information on tools, cases etc.
> 
> Manos
> 
> 

Uche Ogbuji has a series of articles on IBM Developerworks that 
describes how to do the above - well worth the read if you want to do 
this sort of stuff. There are 7 parts to the series starting here:
http://www-106.ibm.com/developerworks/xml/library/x-think4/

I think there is inbuilt support for this type of stuff in their 
(Fourthought's) 4Suite product - haven't tried it yet myself - maybe if 
he's still lurking around here he can elaborate.

Interestingly this sort of approach is being proposed by some as a way 
of getting around the current RSS impasse - if RDF is dropped, it is 
proposed that a XSLT style sheet could allow an RSS 2.0 feed to be 
easily transformed into rdf for those who need it.

Does anyone have any comments on this strategy - I'm in two minds about 
it. On the one hand I can see the point Uche makes in the above articles 
- namely that it is unrealistic to expect everyone to convert their data 
into RDF format when there is masses of existing non-RDF stuff out there 
- but that shouldn't stop us using rdf tools to manipulate/ query/ 
inference with that data if we can take an rdf "view" (using transforms 
and maybe other aproaches) over such data.

OTOH - in the case of RSS I wonder about issues like scalability if they 
drop RDF as the representation syntax - an RDF tool wishing too (for 
example) aggregate and reason over maybe hundreds of different 
news-feeds may find the transformation process is too great a bottle 
neck to allow for reasoning on the fly.

Comments welcome.


-- 
Murray Spork
Centre for Information Technology Innovation (CITI)
The Redcone Project
Queensland University of Technology, Brisbane, Australia
Phone: +61-7-3864-9488
Email: m.spork@qut.edu.au
Web: http://redcone.gbst.com/

Received on Thursday, 3 October 2002 20:40:47 UTC