Re: Proof of Concept: Identity Credentials Login

On 6/16/14 9:30 PM, Manu Sporny wrote:
>> WebID+TLS does.
>> >
>> >You seem to be betting your company on JSON LD vs Turtle, rather
>> >than allowing both.
> I don't think anyone said that both aren't allowed. You MUST support
> JSON-LD... you MAY also support TURTLE. It's the Web - you can content
> negotiate.

So emphasize the fact that its MUST for JSON-LD and MAY for other 
notations (e.g., TURTLE) via content negotiation.

Right now, at first blush, the assumption is that you are pushing one 
RDF notation over another.

>
> We (Digital Bazaar) don't plan to support/both/  JSON-LD and TURTLE in
> the beginning because the addition of TURTLE doesn't really add any
> advantage to the system. If others start deploying successful commercial
> systems that content negotiate for TURTLE, I'm sure we'd follow suit.
> Adding features increases complexity. Adding features that don't provide
> new capabilities seems like bad design.
>
>> >It's a strange bet in that given that you've already written
>> >canonicalization algorithms that change JSON into ntriples, which are
>> >a form of turtle, I didnt expect it would be a huge undertaking.
> The canonicalization algorithms don't use N-Triples, they use N-Quads.
> This highlights another reason we didn't want to support TURTLE: it
> doesn't support graph labels (while JSON-LD does).
>
> We did consider N-Triples and TURTLE for the JSON-LD graph
> canonicalization algorithms and decided not to use either because
> neither provided the flexibility and scalability necessary to do proper
> digital signatures on graphs.
>

You have NQuad [1], TriG [2] for these matters. The fact that the W3C 
has so many notations isn't really the fault of TURTLE [3], especially 
when this boils down to syntax sugar for Graph denotation, prefixes, and 
use of ";" for RDF statements (triple based relations) where the subject 
is constant while predicate and objects pairs vary.

Links:

[1] http://www.w3.org/TR/2014/PR-n-quads-20140109/ -- N-Quads Notation 
(which is basically N-Triples + Named Graph Identifier)
[2] http://www.w3.org/TR/2014/PR-trig-20140109/ -- TriG (which is 
basically TURTLE + Named Graph Identifier)
[3] http://bit.ly/LISWoI -- TURTLE Notation.

-- 

Regards,

Kingsley Idehen 
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen

Received on Wednesday, 18 June 2014 13:57:46 UTC