Named Graphs (was RE: Reification - whats best practice?

Hi Patrick,

Patrick.Stickler@nokia.com writes:
 > 
 > Since named graphs are named by URIs, one can say what one
 > likes about them.

Something Steve Harris mentioned to me at foaf-galway and got me
thinking: 
If named graphs are identified by URI, what do you do when you get 2
graphs with the same URI?

I'm in the process of developing a store that handles graphs (amongst
other things), and have found this crops up quite a lot.
e.g. somebody publishes a graph, and then publishes it again slightly
ammended with the same URI. Or you get the same graph from 2 different
sources (e.g. quoted from 2 different documents), but can't guarantee
that they're the same.

AFAICS the best solution appears to be to store the graph with an
extra layer of indirection - e.g. name the graph with a bnode, and
then have properties of the bnode that identify the graph (one of
which could be a URI). This allows multiple graphs with the same URI
to be stored (but not uniquely identified). Other metadata assocated
with the graph could then be used to differentiate them (e.g. source,
datestamp).

Is there a better solution to this problem?

Cheers,

Phil

Received on Thursday, 9 September 2004 09:23:28 UTC