- From: Stian Soiland <ssoiland@cs.man.ac.uk>
- Date: Tue, 6 Feb 2007 08:47:59 +0000
- To: Adam <adamsobieski@hotmail.com>
- Cc: semantic-web@w3.org
- Message-Id: <B3070A9E-A3F0-4C51-89E9-444761154147@cs.man.ac.uk>
On 3 Feb 2007, at 19:42, Adam wrote:
> In looking over some RDF triple store libraries, I noticed some
> take a metadata-optional approach. This notation may also take
> that approach. A pure graph representation (three columns) could
> possibly discard the additional information in such a information-
> bearing URI and view the element as a node. However, there may be
> possiblities for metadata (potentially requiring a primary key)
> beyond sourcing a triple to an author, and other conceivably
> discardable statements in graph merging. So, this idea appears to
> fall into the metadata might be useful category.
Although a bit verbose, can't you use something like in N3:
{ fish:book dc:title "Moby Dick" } a n3:falsehood .
.. and limit yourself to using triple in the {}? Of course this
wouldn't capture where you actually got that tuple from, and wouldn't
give you a nice URI scheme, but at least we would know which triple
we are talking about.
--
Stian Soiland, myGrid team
School of Computer Science
The University of Manchester
http://www.cs.man.ac.uk/~ssoiland/
Received on Tuesday, 6 February 2007 08:48:24 UTC