Lists Home |
Date Index |
On May 17, 2004, at 12:49 PM, John Song wrote:
> It seems that the most appealing thing to store RDF in NXD is that i
> can utilize the index and other optimizing techniques to query
> it,but i am no very sure about that.Could anybody give me some
My understanding is that the standardized XML serialization of RDF is
not terribly interesting to query using XPath or XQuery because a) the
numerous different ways one can serialize the same RDF model, and b)
most of the interesting aspects of a query on RDF have to do with the
inferences one can get from traversing the RDF graph, not from the
nodes in isolation.
Jonathan Robie's "The Syntactic Web" piece
addresses the first point by suggesting that normalized RDF be stored
and queried, but I don't know how receptive the RDF world has been to
it. If Jonathan is still on this list, he may be able to follow up.
The current work I know about by Tamino customers tends to involve the
combination of an RDF or OWL reasoner with the NXD. There are
synergies there -- Huge ontologies don't fit efficiently into memory so
DB support is helpful, and the XML queries only give the first cut of
results, and an inferencer is needed to take them further.