Re: Potential Formal Object from DERI over JSON-LD

On 10/22/12 7:03 AM, Peter F. Patel-Schneider wrote:
> There has to be some notion that everyone is serving up stuff that 
> others can read.  Otherwise linked data is nothing more than a slogan.
Peter, Pat:

Here is a concise description of Linked Data, as I know it.

Linked Data describes a method for publishing documents comprised of 
structured content. The content is question is constrained by the 
entity-attribute-value (EAV) model and takes the form of an entity 
relationship graph. In addition, this form of structured data 
representation requires every entity, attribute, and attribute value 
(optionally) to denoted by a de-referencable URI thereby enabling the 
use of indirection (explicit or implicit) to associate a URI with a 
description (or descriptor) document.

RDF based Linked Data is all of the above plus the addition of explicit 
semantics that extend from the data model all the way up to fine-grained 
entity relationship semantics.

Effects of Linked Data:

1. An entity name or document address gets you to the same structured 
data (content)
2. If using HTTP, actual content format is negotiable
3. The Web becomes a functional entity relationship graph based on the 
content of Web documents.

Added effects of RDF based Linked Data:

1. Web documents are comprised of self-describing content endowed with 
explicit entity relationship semantics
2. Powerful reasoning is possible since the aforementioned entity 
relationship semantics are machine comprehensible .

Then to Sandro's g-box, g-text, g-snap. I believe this can be mapped as 
follows:

g-box -- Linked Data document (this has a URL)
g-text -- entity relationship graph markup notation for document content 
(URIs -- which may or may not be de-referencable -- used as denotation 
mechanism for eav/spo triples)
g-snap -- instance of an entity relationship graph derived from the 
content of an document.


I hope this helps. I've always believed we can distill all of this 
issues into a very simple meme that achieves the following:

1. engages a broad range of developers, integrators, and end-users 
profiles -- many of which are familiar with the entity-attribute 
value-model and entity relationship subject matter in general

2. negates distraction of those suffering from R-D-F reflux or bearers 
of unhealed scars from various past wars about syntax rather than semantics

3. make it easy for early adopters of Linked Data to produce and share 
consistent collateral for marketing, tutorials etc.

4. remove HttpRange-14 permathread from Linked Data conversations

5. easily showcase DIY Linked Data deployment without implying that 
control over a domain name, a domain name server, and a web server are 
prerequisites

All of what I've outlined is possible. If it wasn't I wouldn't burn so 
much time repeating myself.

EAV, RDF, and Linked Data are loosely associated in complimentary ways. 
I am confident we can collectively produce specs and drive narratives 
that reflect this reality :-)


Links:

1. http://lists.w3.org/Archives/Public/public-rdf-wg/2012Sep/0003.html 
-- prior thread about this matter
2. http://lists.w3.org/Archives/Public/public-rdf-wg/2012May/0571.html 
-- related conversation about entity denotation and resource identification
3. http://ontolog.cim3.net/forum/ontolog-forum/2012-07/msg00191.html --  
orthogonal ontolog forum conversation about the same thing.

-- 

Regards,

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