W3C home > Mailing lists > Public > public-rdf-wg@w3.org > October 2012

Re: Potential Formal Object from DERI over JSON-LD

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Tue, 23 Oct 2012 09:35:53 -0400
Message-ID: <50869D39.9050304@openlinksw.com>
To: public-rdf-wg@w3.org
On 10/23/12 8:19 AM, Peter F. Patel-Schneider wrote:
> I'm with Pat on this one.
> It is true that I don't understand this view of linked data, but I 
> don't think that the fault here lies in me.  This set of principles is 
> too vague to be anything more than a slogan.  It appears that just 
> about any logic can be used, even truly weird ones, by simply using 
> IRIs for identifiers.

It has nothing to do with logic in an explicit sense. RDF is about logic 
in an explicit sense because from top to bottom semantics are explicit 
and clearly defined. Linked Data has a single concern: webby structured 
data that exploits the implicit ingenuity of URI abstraction combined 
with an entity-attribute-value data model.
> If Linked Data was
> - use dereferenceable IRIs for identifiers

The meme says that.

> - have a semantics where identifiers globally denote


> - have a common semantic framework
It doesn't say explicitly.

> - if you coin a new IRI use one that you control

I don't think that's relevant or feasible on the World Wide Web without 
introducing challenges such as:

1. domain name ownership
2. dns server access and control
3. web server access and control
4. HttpRange-14 confusion about de-referencable URI based entity 
denotation and ambiguity .

> - use IRIs coined by others where possible
It sorta implies that, but not explicitly, and for good reason i.e., the 
audience is too broad for this to meaningful at first blush.

> - when someone asks for the document at an IRI provide information 
> that you believe about what you believe the IRI to denote

Why? The meme really implies: If you can, return a document the 
describes (or defines) the URI's referent. Otherwise, any other 
information about the URI referent will do.

> - provide information in a well-known syntax under the common semantic 
> framework

Why? It should say and imply: make a best effort to provide a document 
comprised of structured content that describes (or defines) the URI's 
> Then I would be happier.

Well many of us have been happy since 2006-2007 and we then moved out 
attention to crowd-souring a massive Linked Open Data cloud. As of 
today, we have more than 55 Billion triples in this cloud, and that's a 
very conservative estimate. In fact, I know it's much more than that.
> I would be even happier if Linked Data utilized the RDF(S)(++)(+) 
> semantics and the syntaxes were for RDF graphs.

I doesn't need to do that at all. That doesn't stop anyone making higher 
semantic fidelity Linked Data using what you've outlined. It doesn't 
even stop the W3C from making that part of a recommendation re., how to 
publish Linked Data using RDF. There's only a problem when RDF attempts 
to own Linked Data which is utterly broken.

> peter
> On 10/23/2012 07:25 AM, Richard Cyganiak wrote:
>> Hi Pat,
>> On 23 Oct 2012, at 03:07, Pat Hayes wrote:
>>>>> I would be very interested to discover what y'all consider the be 
>>>>> the definition of Linked Data. Can you provide a pointer to where 
>>>>> this can be found? Thanks in advance.
>>>> http://en.wikipedia.org/wiki/Linked_Data
>>> This is completely vacuous, almost a textbook example of a Wikipedia 
>>> article that is free of content (there are quite a number of them.) 
>>> For example, it begins, " linked data describes a method of 
>>> publishing structured data" but it then does not tell us what this 
>>> "method" actually is. Which is what my query was asking for. What 
>>> actually IS "linked data"? If I were shown some data, or a data 
>>> storage scheme of some kind, how would I know if it were an example 
>>> of linked data? How would I tell? What criteria would I use to 
>>> detect the presence of "linkedness" in the data? (Can anyone give me 
>>> an example of data that is not linked data, and tell me why it 
>>> isn't? That would be a start.)
>> http://en.wikipedia.org/wiki/Linked_Data#Principles
>>> That is great, but how would anyone know that this is what they were 
>>> in fact doing? Does just using JSON + using URIs make it linked 
>>> data? Apparently not, according to the Wikipedia article, which says 
>>> that linked data pre-dates URIs.
>> No, it says that the general idea predates URIs (and that claim is 
>> flagged with [citation needed]), but that the term was coined by 
>> TimBL. Following the references you'll see that his original article 
>> prominently features URIs.
>>> So what is it that makes the linked magic happen?
>> See the link above.
>>> This is completely vacuous
>>> free of content
>>> empty phrase devoid of meaning
>> Translation Pat => English: “I don't understand it.”
>> Best,
>> Richard
>>> Pat
>>>> That's great. However, there's a thin line between saying “we 
>>>> enable LD with JSON” and “JSON-LD is how you do LD”. The JSON-LD 
>>>> spec really ought to say only the first thing, but slips into 
>>>> implying the second too often.
>>>> Attempting to enforce a particular implementation technology for 
>>>> Linked Data, be it RDF or JSON or Atom or Microdata or whatever, 
>>>> doesn't work. This is what Kingsley keeps repeating on a daily 
>>>> basis, and he's right.
>>>> The fact that a JSON-LD document also can be parsed to an RDF graph 
>>>> is mostly orthogonal to this.
>>>> Best,
>>>> Richard
>>> ------------------------------------------------------------
>>> IHMC                                     (850)434 8903 or (650)494 3973
>>> 40 South Alcaniz St.           (850)202 4416   office
>>> Pensacola                            (850)202 4440   fax
>>> FL 32502                              (850)291 0667   mobile
>>> phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes



Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen

Received on Tuesday, 23 October 2012 13:36:15 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:04:22 UTC