Re: Schism in the Semantic Web community.

Hi Olivier,

I'm not on this list anymore so didn't see your message right away.

First, Swoop is currently dead. There's been no significant work on it  
in 3 years and none at all in 2. However, there was a version that  
Clark & Parsia updated to support then OWL 1.1. (Swoop was based on  
the OWL API.)

Protege 4 fully supports OWL 2 and so will OWLSight.

I would definitely quibble with "OWLAPI does not manage semantic web  
at the statement level." Indeed it does. It does not present a triple  
level api, but "statements" in OWL are, and have always been, larger  
than single triples. OWL2 has clarified that which makes a number of  
API based services (justifications, modules, metrics) much easier to  
design and implement.

As for Jena, I can't speak for the Jena developers, but it is an open  
source project. Top Braid Composer is Jena based and has support for  
OWL 1.1.

Finally, I do hope that the forthcoming SPARQL group charter will  
allow for the specification of SPARQL/OWL, which would, explicitly,  
provide a standard (query based) triple oriented API for OWL.

The WG tried really hard to bridge between OWL and RDF on every level  
without unduly shortchanging either side. It's a difficult task.

"""Will we see a schism between RDF tools, and OWL tools?
Virtuoso vs OwlGres?"""
Well, as long as they both support SPARQL, I don't see that it's too  
bad. But I also think it'd be reasonable for the products to  
distinguish themselves on the basis of orientation.
I want to point out that the OWL/XML syntax is an example where we are  
trying to bridge to a wider set of technologies to prevent "the vendor  
lock-in nightmare". OWL/XML has an XML Schema which allows you to use  
standard XML editors (like oXygen) with nice features like  
autocompletion. It also makes for nice use of XPath, XQuery, XSLT,  
data binding, CSS, etc.

So, it's easy to imagine scenarios where a company has a heavy  
commitment to a XML database based document repository and tool suite.  
They start using OWL for schema integration (using OWL/XML). After a  
while they see an OWL QL layer for their existing relational database  
can give them sound and complete distributed queries. So they deploy  
OWLGres (or Quanto). In a separate division, they've been working with  
linked data in virtuoso with OWL RL augmentation.

When the moment comes that they realize they have to integrate all  
this stuff, they can. Pretty easily. They can do SPARQL queries over  
*all* their data.

It's been a personal, driving concern for me, since before OWL 1.1  
existed, to try to make sure that the strength of the suite of  
Semantic Web technologies synergize rather than cannabilize. I include  
XML in that suite. I also, obviously, include RDF. I also include OWL.  
It's not easy to make all of these things work together well for  
everyone in every case coming from every perspective. Interoperability  
has a price! But I think and hope we've done a reasonable job with  
OWL2 and that it will help grow the Semantic Web and the technologies  
associated with it. And I hope it will do so by being genuinely useful.

Cheers,
Bijan.

P.S. Please CC replies to me, as I'm not on this list.

Received on Tuesday, 27 January 2009 09:43:35 UTC