- From: Dan Brickley <danbri@danbri.org>
- Date: Mon, 30 May 2011 10:11:36 +0300
- To: Dave Longley <dlongley@digitalbazaar.com>
- Cc: public-linked-json@w3.org
On 28 May 2011 21:43, Dave Longley <dlongley@digitalbazaar.com> wrote: > There's been some talk about trying to canonically-name bnodes during > JSON-LD normalization. Below are some preliminary thoughts on an algorithm > for resolving bnode-equality so that JSON-LD normalizers produce the same > names for bnodes. Hi Dave, This has been discussed in the RDF community since the late '90s (eg. see http://infolab.stanford.edu/~stefan/updates.html ), and most thoroughly investigated by Jeremy Carroll in his work on signing RDF graphs. The approach you outline seems in a similar direction, I think you'll find his paper interesting (especially re 'fingerprinting' bnodes w.r.t. their surrounding graph neighbourhood). I'm having trouble getting the exact link right now but searching on: jeremy carroll signing rdf graphs ... should get you close. [...] Ah, here it is: http://www.hpl.hp.com/techreports/2003/HPL-2003-142.html -- see section 4 "Labelling Blank Nodes". Recently the new RDF WG is looking into related topics, see http://www.w3.org/2011/rdf-wg/wiki/Skolemisation and http://www.w3.org/2011/rdf-wg/track/issues/40 and nearby. It is likely we will specify a class of URIs to use to represent the output of skolemisation, even if the mechanism for generating the IDs is left unrestricted. Hope this helps, cheers, Dan
Received on Monday, 30 May 2011 07:12:03 UTC