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

Kingsley Idehen <kidehen@openlinksw.com> wrote on 08/07/2012 08:01:08 AM:

> From: Kingsley Idehen <kidehen@openlinksw.com>
> To: public-ldp-wg@w3.org, 
> Date: 08/07/2012 08:02 AM
> Subject: Re: LDP interfaces in Java (based on Jena and JAX-RS)
> 
> On 8/7/12 3:48 AM, Wilde, Erik wrote:
> > hello.
> >
> > On 2012-08-06 23:05 , "Kingsley Idehen" <kidehen@openlinksw.com> 
wrote:
> >>     We are basically trying to work out a loosely coupled 
architecture
> >>     for data object identification, representation, and access via:
> >>     1. URIs
> >>     2. Structured Data
> >>     3. RESTful interaction patterns.
> >>     It can't be more than that :-)
> > i am not sure that people agree that these are the only constraints we
> > have. because if that were true,
> > http://dret.typepad.com/dretblog/atom-landscape.html is the answer, is 
a
> > standard, provides all kinds of features beyond the use cases we have 
in
> > our document, and thus would be the way to go, instead of defining
> > something new.
> >
> > cheers,
> >
> > dret.
> 
> Of course AtomPub isn't the answer re. fine-grained data access.
> 
> You need what's outlined above which is basically the fundamental 
> architecture of the Web (AWWW) itself.
> 
> The architecture of the Web is dexterous enough to handle all of the 
> fine-grained data access needs that these conversations are hovering 
about.
> 
> BTW -- my products implemented AtomPub before the ink dried on the 
> initial specs. We still support it today and part of many protocols we 
> support. None of our protocol handling heuristics would work modulo AWWW 

> dexterity.
> 
> Decouple the letters R-D-F from Linked Data and all the real power of 
> AWWW is there to be exploited in ways only limited by our imagination.
> >

Agree with Kingsley, with one small refinement.  If we go down the route 
of defining some "simpler than R-D-F" data model, it will probably turn 
out we defining something that looks like RDF or something whose semantics 
we need to translate to RDF.  This may be worth exploring in the longer 
run.

Feels like if we can nail down our use cases and requirements, it will 
help flush out the need for such things (even when elaborated in a more 
detailed way).  Until that, we are having a bit of an abstract debate I 
believe.

As the member submission highlights, you can achieve a very powerful 
LInked Data Platform with AWWW, some clarifying rules and some concepts 
from RDF (no full SPARQL needed, etc) and using already registered media 
types based on the representations of the RDF (XML, JSON, Turtle, etc).

Thanks,
Steve Speicher
IBM Rational Software
OSLC - Lifecycle integration inspired by the web -> 
http://open-services.net

Received on Tuesday, 7 August 2012 12:20:32 UTC