Re: Schism in the Semantic Web community.

Hi Olivier and all,

One additional piece of news that may be of interest here: the Ontology 
Definition Metamodel (ODM) was formally adopted as an OMG 1.0 standard 
last month at the Santa Clara OMG meeting.  The finalized specification 
will be published on the OMG web site sometime next month, but in the 
interim, the current beta specification is available at 
http://www.omg.org/spec/ODM/1.0/Beta3/.

We are launching an Eclipse Modeling Project (under the MDT heading), 
planned for the upcoming Galileo release (June 2009), that will include 
ODM metamodels, profiles, and APIs, derived directly from the standard.  
The profiles provide the ODM-based UML graphical notation for RDF, RDFS, 
OWL, and Topic Maps, and the metamodels provide the XMI support for 
model interchange, required to help connect the software engineering and 
semantic web communities.   Participation and feedback from this 
community is most welcome.  Once we've launched the project, I'll 
forward the announcement to this list.

While we are a little behind the OWL 2 work at OMG (and thus on this 
Eclipse project), both Evan Wallace (NIST) and I have been members of 
the OWL 2 working group from the outset, and plan to introduce work at 
OMG to augment the ODM to support OWL 2 once the language is 
sufficiently stable.  Among the issues we will need to address are: how 
to organize the UML/MOF models to make it relatively easy to adopt 
specific profiles of OWL 2, what kinds of visual cues would be helpful 
from a graphical notation perspective, and whether we need a standalone 
metamodel for OWL 2 that we then map to an RDF-compatible metamodel, in 
the way that the language specifications are organized (ODM 1.0 does not 
do this).  The APIs developed for the Eclipse project may or may not be 
similar to the Jena API -- although the HP Labs/Jena team has had 
significant influence on the ODM and are also participating in the 
Eclipse project.  "Stay tuned" is the only thing we can say about this 
at the moment.

Long story short, I think there is indeed an effort to continue to 
support the RDF and OWL user communities together -- at least from a 
software engineering / OMG perspective, and we look forward to working 
with this community to support your requirements.

Best regards,

Elisa

Jens Lehmann wrote:
> Hello Olivier,
>
> Jens Lehmann wrote:
>   
>> Hello,
>>
>>     
> [...]
>   
>>> OWLAPI does not manage semantic web at the statement level.
>>> Jena was one of the few libraries to be both statement-based and concept-based.
>>> Basically, i would say that Jena was the only API to glue the RDF
>>> world and the OWL world together.
>>> And because of that, it had been adopted by developpers as a
>>> "one-size-fits-all" library.
>>> And by RDF database vendors as their API for their RDF storage system.
>>>
>>> This makes me wonder:
>>> What will happen in the next future?
>>> Will we see a schism between RDF tools, and OWL tools?
>>> Virtuoso vs OwlGres?
>>>
>>> I think it is good that the Semantic Web has always keeped that
>>> internal competition between statements
>>> and boxes. But if tools support is splitted, then what? Two
>>> communities? Mass storage vs inference? A crucial
>>> tools choice to make in any IT project? Then we are back in the vendor
>>> lock-in nightmare.
>>>
>>> Well, this is quite a rough reflexion.
>>>       
>
> (Sorry, that I send my first mail too early accidently.) I wanted to add
> that you are hitting an interesting spot here. I agree that RDF and OWL
> tools and - even more important - their communities should not be too
> separated. However, I believe that the upcoming OWL 2 standard does not
> have a negative impact here. Every OWL 2 ontology can be mapped to RDF
> and vice versa. Furthermore, OWL 2 has been extended to simplify tool
> support and solve some practival problems, e.g. the specification is now
> accompanied by UML diagrams, axioms can be annotated, entities can be
> used as instances and classes through punning etc. Summed up, I believe
> that if there will be a split between OWL and RDF communities (which we
> should avoid), then - from my point of view - this is unlikely to be the
> fault of the upcoming OWL 2 standard.
>
> Kind regards,
>
> Jens
>
>

Received on Friday, 9 January 2009 22:05:41 UTC