W3C home > Mailing lists > Public > public-rdf-wg@w3.org > April 2011

Re: [JSON] Modularization and JSON-LD

From: Steve Harris <steve.harris@garlik.com>
Date: Fri, 8 Apr 2011 09:49:22 +0100
Cc: Manu Sporny <msporny@digitalbazaar.com>, RDF Working Group <public-rdf-wg@w3.org>
Message-Id: <CFA4937F-A31C-4956-9BAF-213FC2D0ABCE@garlik.com>
To: Richard Cyganiak <richard@cyganiak.de>
On 2011-04-07, at 21:43, Richard Cyganiak wrote:

> Hi Steve,
> 
> On 6 Apr 2011, at 12:50, Steve Harris wrote:
>>> ... JSON-LD ...
>> 
>> So, I'm still a little confused about what the use-case is for this style of RDF serialisation.
>> 
>> The goal seems to be some kind of duality where "Javascript people" can consume it natively as JSON, and "RDF people" can consume it as triples, but I'm not clear on how either part wins over what we have now.
> 
> So here it is as I understand it. RDF, despite all its shortcomings, has several very real and tangible advantages over JSON:
> 
> 1. URIs are explicitly marked up, so you get proper hypermedia and object identifiers.
> 
> 2. Terms in your data dictionary can be globally unique and resolve to human- and machine-readable definitions. Corollary: It's easy to standardize data dictionaries.
> 
> 3. Publishers can mix their own data dictionary terms with those defined and standardized by others in a single data object.
> 
> 4. Data from several sources will automatically mix at the consumer, to the extent that the sources have aligned their data dictionaries and object identifiers.
> 
> Let's assume that these advantages appeal to some JSON publishers (who don't care much for RDF).
> 
> Now let's further assume that we can standardize some conventions for JSON that, if followed, produce the benefits above.
> 
> So the carrot for "JSON people" is that they can make their JSON more hypermedia friendly and generally more webby by following some of these conventions, in some places, wherever it makes sense for them.

All good so far, but that reads as much as a "good design in your API" guide, than a selling point for JSON as RDF.

> The side effect: Now at least a part of their JSON parses to (some subset of) RDF. So the "RDF people" win because they get more RDF, from parties that would never ever publish Turtle or RDF/XML.

Yeah, here's where I stop believing it. If it was more Linked Data, I'd say that's a win, but Twitter API results being parsable as triples? I just don't see any advantage, and I see some pretty significant disadvantages  cleanliness, complexity, do one thing well etc.

RDF adoption is not (just) a numbers game. If there's 400TT of RDF out there, but no-ones making use of the advantages of RDF, that's not really a win.

If, as Manu hopes, there's some format where you can look at it, and not think it's made the JSON harder to understand, but it still had the useful properties of RDF, then that would make sense, but I've not seen that yet,

> This story is how I can talk myself into believing that JSON-LD or something like it could be a really good idea.

It was working for me too, on and off. Now I really want to see a convincing use-case though.

- Steve

-- 
Steve Harris, CTO, Garlik Limited
1-3 Halford Road, Richmond, TW10 6AW, UK
+44 20 8439 8203  http://www.garlik.com/
Registered in England and Wales 535 7233 VAT # 849 0517 11
Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD
Received on Friday, 8 April 2011 08:49:51 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:41 GMT