[Mulgara-dev] File based resolver

[Andrae Muys]

On 06/07/2007, at 7:12 PM, thomas wrote:

> --On 29. Juni 2007 17:55:11 +0200 thomas <thomas at stray.net> wrote:
>> <snip>
>> * Many people are interested in using JCR for storing and querying  
>> RDF
>> triples. This is something we may want to look into in terms of
>> performance and content modelling. One approach could be to define a
>> RDF mixin type with an rdf:about property that could be used to  
>> turn a
>> node into an RDF resource. Property names would be the predicates and
>> property values the objects. This way you could easily make your
>> existing content model available as RDF. There was interest in
>> implementing a Jackrabbit binding for the ARQ SPARQL query engine
>> (http://jena.sourceforge.net/ARQ/). Something like that could be used
>> as the "SPARQL" query language in JCR.
>> </snip>
>
<snip>
> i don't know how far the componentization of mulgara holds in  
> practice but would it be imaginable to integrate it's central  
> components in jackrabbit? and since with XA2 a good part of these  
> central components get rewritten anyway, maybe the "rewriting the  
> whole thing" in the context of a jackrabbit integration becomes  
> thinkable? well, just an idea... (and one possible way to tackle  
> the file based resolver issue)

The existing resolver-spi already provides everything you require to  
provide an iTQL interface to jackrabbit - and of course once we have  
SPARQL support SPARQL could be used as well.  If someone writes a  
JCRResolver, we're done.

Andrae

-- 
Andrae Muys
andrae at netymon.com
Mulgara Consultant
Netymon Pty Ltd