Re: Three alternative approaches for fixing the blank node scope problem.

As my previous message is now in the agenda for tomorrow, I should point out to the WG that the current draft of Semantics actually has none of these three alternatives in it, and has reverted to something much closer to the 2004 wording. (Following PFPS' suggesiton, see http://lists.w3.org/Archives/Public/public-rdf-wg/2013Mar/0162.html ) This should keep things simpler and require a lot less debate, and removes any obligation someone might have felt to read through the large email to which this is a reply.

However, I do want to ask a quid pro quo from the other editors, so that other parts of the spec take up the slack that is now missing from Semantics. In particular:

1. NTriples and NQuads should clearly state that bnodeIDs in different documents cannot identify the same blank node. 

Can two datasets share a blank node? (I have raised this as issue 122.) 

2a. If not, Concepts should clearly state that two different datasets cannot share a blank node. (This has the consequence that a single large graph cannot be transmitted using several datasets.) 

2b. If so, then the WG needs to figure out, and Concepts needs to carefully state, how users can determine, given several datasets, which occurrences of bnodeIDs in them represent the same blank node and which do not. 

Pat

. 
------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes

Received on Wednesday, 20 March 2013 01:55:37 UTC