- From: carmen <_@whats-your.name>
- Date: Thu, 5 Nov 2009 13:34:35 -0800
- To: semantic-web@w3.org
> >inventing a system of referring to > >huge sets of those in a URI. Not doable. huge sets of triples like..named graphs? its already being done > > I don't see why naming triples is not doable URIs for everything is a foundation i stick to. as well im interested in your scheme, i just do <<Suri>,<Puri>,<Ouri>> where O(bject)uri is either a data: URI or a file:// or http:// URL to a blob (or a (b)node with type/lang/encoding-tag fields around a blob) the named-graph crowd seems to advocating creating a graph for each triple youd want to reify, i take it. then youd have ot think up a name for that graph, which hopefully doesnt clash with the name of a normal resource in the system, and somehow usefully describes the triple? id love to see the compoendium of examples that would convince me named graphs is the way to go , not that im likely to switch > collections of triples isn't either, we already have native lists in RDF. > > From the perspective of a triplestore it ought to be optimisable (There > is no need for them to represent it as triples internally), and in > serializations sugared-syntax would be sufficient to make it usable. > > My reply to the original post explains some of the reasons I think we > cannot do without meta-triples. >
Received on Thursday, 5 November 2009 21:35:17 UTC