- From: Bijan Parsia <bparsia@cs.man.ac.uk>
- Date: Tue, 14 Aug 2007 16:24:57 +0100
- To: Story Henry <henry.story@bblfish.net>
- Cc: semantic-web@w3.org
On 14 Aug 2007, at 14:17, Story Henry wrote:
> There is clearly a relation between graphs and their literal
> representations.
XMLLiterals are not literal representations of graphs. The are
literal representation of XML bits (fragments? infosets? not quite
sure). Those XML bits are representations of (i.e., may be parsed to)
graphs.
This is one reason that tool kits don't handle them well as graphs.
Compare with string literals with Turtle or XML literals with SVG.
> Perhaps something like the relation between a foaf:Person and their
> foaf:mbox?
No, on several fronts (i.e., people don't have literal
representations ;)).
> [] a Graph;
> n3:literal "_:it <http://purl.org/dc/elements/1.1/title>
> 'Interesting Times'";
> = { [] dc:title "Interesting Times" };
> = "_:it <http://purl.org/dc/elements/1.1/title> 'Interesting
> Times'"^n3:literal .
Ah, I see, you meant "indirect". Sure. Or you could make a special
datatype who's lexical space was RDF/XML (or turtle) and who's value
space was graphs.
[snip]
> It is clear that saving the graphs in a triple database the way one
> saves other triples, does make it easier to work with them. It is
> easier to merging, search and querying graphs if one does not
> first have to deserialise a rdf/xml representation. So I do like
> the { } notation in N3.
This is an implementation thing, I think.
Cheers,
Bijan.
Received on Tuesday, 14 August 2007 15:23:56 UTC