Re: [regrep] RDF Data Access WG Charter

[Farrukh Najmi <Farrukh.Najmi@Sun.COM>]

John Gillerman wrote:

>Farrukh,
>
>Can you elaborate more on your ideas here.  If the RDF API supports the
>ebXML submit/approve/deprecate/remove workflow and has a sufficiently
>powerful query mechanism, why wouldn't an RDF API completely meet our
>registry API needs?  
>
The RDF Net API provides a small subset of functionality provided by the 
ebXML Registry API. Even that functionality is specific to RDF. There is 
no support for federation, event notification, ad hoc query, 
classification, deprecation, approval etc....

ebXML Registry API supports a much broader functionality and for any 
type of content not just RDF content. What it lacking is the ability to 
manage RDF and OWL as first-class elements of its information model. 
This is what the Semantic Content Management SC (SCM SC) proposes to 
address in version 4.

While this is still half-baked thoughts and needs the team's validation, 
I envision that in version 4 we will keep and extend our API (most of 
the Registry Services spec). In the short term we will keep our existing 
RIM but extend it to add native support for OWL Ontologys. Over the long 
term we will evolve RIM so it will itself be natively expressed in OWL/RDF.

Several new capabilities will be enabled:

-Use of Ontologys for classification

-Inference based discovery that can take attribute values for ontology 
classes as well as relationships between Ontology classes into account

-Support for more general classification scheme structures such as 
Multiple Inheritence and graph

-Support for arbitrary user defined type extensibility in RIM

The important point is that we will take an evolutionary approach to 
Semantic Content Management, one that leverages the strengths of our 
existing capabilities. We will incrementally add OWL/RDF support as 
optional features. This will allow implementors and users to keep pace 
as the standard evolves.

As for the RDF Net API, it is just one possible input into the future 
RDF Data Access WG. I think that ebXML Registry has the potential for 
being a much better fit for the requirements once we add the SCM 
capabilities.

>Would we want to find out if RDF DA WG is interested in
>supporting our use cases?  
>
I do not think they want to broaden their scope to include any type of 
content and all the other stuff that we do.
What I am suggesting is that ebXML Registry 4.0 API be proposed as the 
basis for RDF Access API.

>It seems that the best thing would be the
>unification of the two API's.  
>
The two API have very different goals, origin and futures. We should 
study RDF Net API (and I am) to see what we can learn leverage but I 
suspect that this would yield modest returns.

>I fully agree that submitting requirements to
>or just following the progress of the RDF DA WG is a good thing.
>  
>
There is significant overlap in the charter of RDF DA and SCM SC. We 
need to figure out with the Semantic Web folks at w3c how we can best 
work together to avoid duplication of effort and to meet our common 
objectives synergistically.

Happy new year everyone.

--
Regards,
Farrukh


>John
>
>-----Original Message-----
>From: Farrukh Najmi [mailto:Farrukh.Najmi@Sun.COM]
>Sent: Tuesday, December 30, 2003 2:46 PM
>To: regrep@lists.oasis-open.org
>Subject: Re: [regrep] RDF Data Access WG Charter
>
>
>Chiusano Joseph wrote:
>
>  
>
>>Farrukh,
>>
>>Thanks for this information. I attended an XML 2003 session [1] given by
>>Graham Moore (co-author, RDF Net API) that covered RDF Data Access, and
>>found it very interesting.
>>
>>Joe
>>
>>[1] "Semantic Web Servers - Engineering the Semantic Web":
>>http://www.xmlconference.org/xmlusa/2003/thursday.asp#35
>>
>>
>>    
>>
>
>I did not attend that session. Do you think that if ebXML Registry
>version 4 provides first class support for publish and discovery of RDF
>and OWL content that it would essentially be providing all the
>functionality provided by the RDF Net API? If so am I correct to assume
>that it would actually provide a super-set of functionality of RDF Net API?
>
>Also, did you get any sense of whether RDF Net API was on a standards
>track anywhere yet? Thanks.
>  
>