- From: Simon Schenk <sschenk@uni-koblenz.de>
- Date: Wed, 06 May 2009 07:21:36 +0100
- To: Holger Knublauch <holger@knublauch.com>
- CC: Semantic Web <semantic-web@w3.org>
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Holger, I think my formulation was misleading. I should have qualified that 'powerful'. NetworkedGraphs are more expressive as a rules/view mechanism. Although constraint checking could be done, it is definitely more combersome than with SPIN. Also, we do not have the nice design tools and GUI machinery around it. So, I id not want to say a is better than b. I guess both are very interesting with slightly different purposes. >> The implementation evaluates rules on the fly instead of materializing >> the results. > > This is an implementation detail and "SPIN" could do that too. Regarding > the spin:rule property, SPIN simply defines a *formalism* for capturing > such rules in a suitable place. However, all SPIN rules can be made > global by inserting class matching clauses that bind ?this into the > beginning of the WHERE clause. An alternative SPIN implementation could > take these CONSTRUCTs and push them onto the server for server-side > evaluation, similar to what you describe in your Networked Graphs > implementation. Sure. Analogously, NGs could be precomputed. > Having said this, I am very interested in your work and it seems to be > nicely complementary technology to what SPIN offers. I like SPIN as well. Thought about adopting the RDF Syntax for SPARQL. Then we could modify view definitions from within views. Spooky. Cheers, Simon -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEARECAAYFAkoBLGsACgkQQ0Lz1fXAQeMdaACeIWfc1/7y15I+Loyl6H4p+iGx THYAnjr1Nv4oD0XLxuV7njssK//TixyS =5zHj -----END PGP SIGNATURE-----
Received on Wednesday, 6 May 2009 06:22:15 UTC