- From: Sandro Hawke <sandro@w3.org>
- Date: Wed, 28 Sep 2011 19:56:52 -0400
- To: Arnaud Le Hors <lehors@us.ibm.com>
- Cc: public-rdf-wg@w3.org
On Wed, 2011-09-28 at 15:35 -0700, Arnaud Le Hors wrote: ... > Finally, although I don't know what actually triggered Sandro's question > about whether the file contains the complete graph or not, it seems to me > that the {} proposal makes it look like what I'm seeing is the complete > graph when it may not be. I know it's up to us to define that there is no > such implication but I'd rather select a syntax that is more intuitive and > less likely to mislead a casual reader/user who may not have read the spec > carefully enough. I agree that {} makes it look more complete, but fwiw, I think what's needed is a way to make it complete.... Maybe this should be a new ISSUE. I can see reasons one might want both, but it's easy to go from complete to incomplete with a subgraph relation, and I'm not sure how to go from incomplete to complete. I guess with pointing to the file you got it from, which seems much trickier. That is, if I need to say that G1 is exactly <s> <p> <o> and G2 contains <s> <p> <o> and maybe some other triples, I can easily say: (making up two new syntaxes for clarity...) G1 @isGraph { <s> <p> <o> }. G2 subGraph G1. but it seems trickier to say: G2 @hasTriples { <s> <p> <o> }. G1 .... something about G2 and this file...? I also like the @graph syntax as being more turtle-ish, but I worry about this aspect of it. -- Sandro
Received on Wednesday, 28 September 2011 23:57:01 UTC