Re: LDP interfaces in Java (based on Jena and JAX-RS)

On 8/6/12 5:33 AM, Erik.Wilde@emc.com wrote:
> hello ashok.
>
> On 2012-08-06 11:08 , "Ashok Malhotra" <ashok.malhotra@oracle.com> wrote:
>> In addition to Reza's questions i would like to ask whether the WG will
>> consider
>> linked data in XML and JSON in addition to linked data in RDF.
> while i cannot speak for the group authoritatively, i had the same
> question when we started working, assuming that REST could mean negotiable
> representations and maybe giving recommendations on how to expose linked
> data for a greater variety of consumers. however, it seems that the
> majority of the group considers "linked data" to be based on RDF by
> definition, and my take-away from the initial discussions was that there
> is little interest to work on how to make linked data available to non-RDF
> clients.

RDF and Linked Data conflation serves neither well re. comprehension and 
adoption.

RDF is a framework comprised of a data model (subject-predicate-object), 
syntaxes for creating content, and content serialization formats. URIs 
are used as a denotation (naming) mechanism in RDF for subjects, 
predicates, and objects (optionally).

The data model aspect of RDF is simply the well known 
entity-attribute-value plus classes and relationships (EAV/CR) model 
enhanced via use of URIs as described above. Please note, there is 
nothing in the RDF specs about de-referencable URIs i.e., RDF doesn't 
require URIs to be de-referencable.

Linked Data is about the EAV/CR or RDF data models enhanced with 
de-referencable URIs. These URIs are functionally dual i.e., they 
combine two fundamental features:

1. entity denotation (naming)
2. web resource identification.

A de-referencable URI, as per Linked Data principles, denotes an entity 
while also using explicit or implicit indirection to resolve to a web 
resource comprised of EAV/CR or RDF based content that describes said 
URI's referent.


> when it comes to simply representing RDF in XML or JSON, there's of course
> always RDF/XML and JSON-LD, which are serializations of RDF in the
> respective formats, but i assume you were asking about scenarios where the
> data model is not based on RDF.

Best you keep the data models distinct from the syntaxes for creating 
content. Ditto the serialization formats for across-the-wire data 
transmission in response to a de-reference act.

RDF is simply an option re. Linked Data. What's important is the 
behavior of URIs, not a specific family of data representation syntaxes 
and serialization formats.

Semantics & syntax conflation remains an eternal problem re. Linked 
Data, unfortunately.


I hope this clarifies matters re. Linked Data and RDF.

>
> cheers,
>
> dret.
>
>
>
>


-- 

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