a proposal for this week - bridging the gap between JSON-LD and RDF

In the spirit of harmony, I propose that if JSON-LD is defined in terms of RDF 
datasets as I suggest in my previous message that the RDF working group remove 
the remaining differences between JSON-LD and RDF datasets, namely to
1/ allow blank nodes for graph names (which has more-or-less been decided on 
already), and
2/ allow blank nodes for properties.

All relevant document should have warnings that blank nodes should not be used 
in these places unless there is a good reason to do so, and that the use of 
blank nodes in these places will cause interoperability problems for some time 
to come.


Why do this?  Well, as soon as JSON-LD gets any use at all, there will be RDF 
datasets with blank nodes for properties.   The alternative would be to 
prohibit blank nodes for properties in JSON-LD.

Peter F. Patel-Schneider
Nuance Communications

PS:  I can hear Pat jumping up and down with glee already.

Received on Tuesday, 18 June 2013 05:02:40 UTC