W3C home > Mailing lists > Public > public-rdf-wg@w3.org > June 2013

Updated JSON-LD spec to more closely align w/ RDF data model

From: Manu Sporny <msporny@digitalbazaar.com>
Date: Sun, 30 Jun 2013 14:18:00 -0400
Message-ID: <51D07658.2010104@digitalbazaar.com>
To: RDF WG <public-rdf-wg@w3.org>
David Booth, Peter S.,

There is a new time-stamped JSON-LD editor's draft that attempts to
integrate all of the discussion related to RDF data model alignment
we've had over the past several weeks:

http://json-ld.org/spec/ED/json-ld/20130630/

Diff-marked version is here:

http://json-ld.org/spec/ED/json-ld/20130630/diff-20130411.html

Pay particular attention to the changes in these sections:

http://json-ld.org/spec/ED/json-ld/20130630/diff-20130411.html#introduction

http://json-ld.org/spec/ED/json-ld/20130630/diff-20130411.html#how-to-read-this-document

http://json-ld.org/spec/ED/json-ld/20130630/diff-20130411.html#design-goals-and-rationale

http://json-ld.org/spec/ED/json-ld/20130630/diff-20130411.html#data-model

http://json-ld.org/spec/ED/json-ld/20130630/diff-20130411.html#relationship-to-rdf

David, I think I've integrated everything we have consensus on in the
JSON-LD CG. The only outstanding issue is what to do with blank node
property skolemization.

Peter, I hope the changes I made are in the right direction. I tried to
not change the terminology that we use throughout the spec too greatly
(because it would have negative cascading effects throughout all of the
JSON-LD specs), while making it very clear that the data model in
JSON-LD is an extension to the RDF data model.

When I started editing the spec to apply each of your changes, my intent
was to keep iterating until Appendix C was removed.

The removal of the blank nodes as graph labels change went just fine.

The blank node as property remains, because it is a difference between
the two models

Gregg thought that we could remove the sets/lists difference, but sets
and lists aren't talked about at all in the RDF data model (in RDF
Concepts). There is nothing to refer to and I couldn't think of a way of
papering over this difference.

I tried to align JSON numbers and JSON booleans with XML Schema, but the
value spaces and lexical spaces don't match up. They are fundamentally
different, so it's one more thing that I couldn't get rid of in the
spec. There is now a note describing why this difference exists:

"""
NOTE
All JSON numbers and booleans can be mapped to XML Schema datatypes,
which are built-in datatypes in the RDF model. Non-decimal JSON numbers
map to xsd:integer and decimal numbers map to xsd:double. JSON numbers
are described as extensions to the RDF data model because they combine
the value space of xsd:integer and xsd:double into a single value space.
JSON booleans may be mapped to XML Schema using the xsd:boolean
datatype. JSON booleans are described as extensions to the RDF data
model because, while they have the same value space, they omit the
values of 0 and 1 from the lexical space.
"""

As I mentioned on the call last week, it'll probably take us a couple of
iterations to get something that both of you and the JSON-LD CG can live
with, so please provide feedback and we'll go from there. We will
discuss these changes on the call on Tuesday if either of you would like
to join and discuss further.

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny, G+: +Manu Sporny)
Founder/CEO - Digital Bazaar, Inc.
blog: Meritora - Web payments commercial launch
http://blog.meritora.com/launch/
Received on Sunday, 30 June 2013 18:18:29 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:02:14 UTC