- From: William Waites <ww@styx.org>
- Date: Thu, 20 Jun 2013 10:25:09 +0100 (BST)
- To: t-eriksson@so.taisho.co.jp
- Cc: melvincarvalho@gmail.com, kidehen@openlinksw.com, public-lod@w3.org
On Thu, 20 Jun 2013 01:28:54 +0000, エリクソン トーレ <t-eriksson@so.taisho.co.jp> said: >> ex:distance ex:earth ex:moon 381550 25150 u:km. > (Ab)using RDF I was able to (barely) document my semantics > directly in turtle. Where is the semantics and syntax of your > example described? Your data might be linked, but as a > prospective consumer of it I'm feeling a bit lost :-) I just made it up. But not out of thin air. It's basically a prolog assertion with a turtle-esque surface syntax -- that's why the predicate comes first, so you can have arbitrary arity. The definitions of ex:distance (and ex:moon, ex:earth, u:km) could be obtained by dereferencing those URIs in the same way it's done with RDF. In fact it takes the two best ideas from RDF -- URIs as identifiers for real-world things, and the mechanism of dereferencing these URIs to get more information. Pace assertions about a normative meaning of Linked Data from the RDF WG (of which I am a member), I think these two ideas are the essence of Linked Data. I'm not seriously advocating this right now, it's just an example or thought experiment to answer your question and there's too much sunk investment in RDF for such a radical change. In fact if we were making radical changes, thinking about lambda expressions might be better than doing it this way. Maybe for RDF 3.0... -w
Received on Thursday, 20 June 2013 09:25:43 UTC