Re: Well Behaved RDF - Taming Blank Nodes, etc.

>>>>> Ivan Herman <ivan@w3.org> writes:
>>>>> On Dec 13, 2012, at 12:42 , David Booth wrote:

[…]

 >>> The restriction of "no labels" is not just about "no cycles" — it's
 >>> things that are not tree-like:

 >>> :x1 :p _:a .
 >>> :x2 :q _:a .

 >> Yes, excellent example.  I explained to Pat in
 >> http://lists.w3.org/Archives/Public/semantic-web/2012Dec/0040.html
 >> why I chose the "no labels" restriction instead, but I'm open to
 >> considering either.

 > I think it would still be better to explain these things in a syntax
 > independent way.  After all, I may want to use JSON-LD or RDFa...

 > Distilling the various mails and concentrating on bnodes only, what
 > seems to be the pattern is

 > - bnodes can appear in at most one triple as an object

 AIUI, this restriction was added merely to allow for a simpler
 WB-RDF definition.  It wouldn't be necessary should Turtle have
 included a “inverse property” syntax.  (Which I doubt it'll
 acquire in a foreseeable future.)

 However, as I've noted earlier, when this restriction is in
 effect, it's simple to assign (semi-)unique identifiers to all
 the blank nodes, based solely on their relation to the other
 nodes.  (Provided that the next criterion is also met.)

 > - there can be no cycle in the graphs involving bnodes

 I believe that the point was that there'd be no cycles
 consisting /exclusively/ of bnodes.

 > Would that suffice as a more formal definition?

-- 
FSF associate member #7257

Received on Friday, 14 December 2012 07:46:13 UTC