[Top] [All Lists]

Re: [ontolog-forum] fitness of XML for ontology

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: Kingsley Idehen <kidehen@xxxxxxxxxxxxxx>
Date: Fri, 07 Feb 2014 12:50:06 -0500
Message-id: <52F51CCE.7030802@xxxxxxxxxxxxxx>
On 2/7/14 11:13 AM, Osorno, Marcos wrote:
 That's why RDF/XML became the biggest distraction and eventual detraction to RDF , in a nutshell. Even as of today, a majority of folks assume that RDF == RDF/XML, and that any conversation about RDF ultimately boils down to issues associated with RDF/XML :-(

Kingsley – I have to respectfully disagree. In our work to use RDF/OWL in a production environment the biggest problems were usually performance and integration into existing libraries and frameworks. Dealing with the XML serialization was rather trivial and really not that hard to manually inspect or train developers on.  In the first case, we found that even though whitepaper benchmarks claimed certain levels of performance, for many of our queries and data-sets the performance wasn't the same or even useable for smallish data-sets (few 100ds GB). I'm guessing some of this was related to the nature of the test data sets versus our data sets.

As someone who's overseen the development of 300+ data transformers driven by RDF/XML, I am not implying that RDF/XML is useless. I am indicating that the manner in which its promotion was handled lead to many of today's problems with RDF. Fundamentally, the biggest problem with RDF is that it is generally misunderstood -- a problem that stems from conflating RDF/XML the notation and RDF the abstract language.

In the second case, there's only so much in our use cases that we wanted from the reasoners and to do in an isolated RDF/OWL/graph-store environment.

RDF/XML has no adverse affect on our reasoners because we reason against RDF Semantics, not the notation used to inscribe RDF statements in the document processed by the reasoner.

The barrier to converting OWL/RDF models to POJOs and was a real problem. With XML or JSON that represents Java type objects the integration of these objects into your systems is often trivial and there are many libraries to help (XJC, Jackson, protostuff, etc).

There was/is a more important barrier: actually seeing the RDF statements for what they are i.e., representations of RDF's abstract language syntax and semantics where entity relationships are expressed using subject->predicate->object triples which can be inscribed in documents using a variety of notations. Basically, the syntax and semantics were lost because too much energy went into debates about parsing and concrete syntax (notation).

It's been several years since I stopped using OWL/RDF but I do periodically look for technologies that ease the integration of RDF resources in the the still dominant software development paradigm of OO objects, maps, dictionaries, and arrays.


[1] http://bit.ly/1fluti1 -- note the *owl:differentFrom* relation I placed in this document

[2] http://bit.ly/1gRpE59 -- ditto here .


Hope that clears up my position :-)



Kingsley Idehen	      
Founder & CEO 
OpenLink Software     
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)

<Prev in Thread] Current Thread [Next in Thread>