Re: SSN-XG Meeting Minutes 15-July-2009

On 16/7/09 Simon.Cox@csiro.au wrote:
> However, once an ontology is 'published' then, in order to provide 
> users with a predictable way to use it, the versioning process must 
> be taken very seriously. 

I totally agree.

> Conventional version control systems were based on 'diff' - primarily 
> a lexical comparison. 
> As Oscar points out, that generally doesn't work for RDF based 
> resources, where the serialization can vary a lot without the content 
> changing. 
> Probably an API-based approach must be used. 
> e.g. Something based on SPARQL for starters - make a set of requests, 
> count the results, do some kind of checksum. 

Some time ago I had to program my own ad-hoc solution for making a 
"diff" between two OWL ontologies using Jena and Pellet.

I have been playing a bit with the OWLDiff tool and looks nice, what I 
don't know is if it provides some API to programmatically perform the 
diff. But it seems a nice solution.

Nevertheless, we have to decide on the granularity of the "diff". If we 
want to use it to notify ontology changes, we have to take care and 
inform of any change in the "axiom" level, since you can change the 
ontology and obtain a semantically-equivalent ontology.

We will have to take a deeper look into the tool and see if it is enough 
for our needs (for example, imports are currently not resolved).

Finally, besides WebProtégé, we have Collaborative Protégé 
(http://protegewiki.stanford.edu/index.php/Collaborative_Protege), which 
allows the online collaborative edition of ontologies.
This tool already allows to make annotations and track changes in the 
ontology, so maybe it solves the versioning issue.

Kind regards,

-- 

Dr. Raúl García Castro
http://delicias.dia.fi.upm.es/~rgarcia/

Ontology Engineering Group (http://www.oeg-upm.net/)
Universidad Politécnica de Madrid
Campus de Montegancedo, s/n - Boadilla del Monte - 28660 Madrid
Phone: +34 91 336 36 70 - Fax: +34 91 352 48 19

Received on Thursday, 16 July 2009 14:45:17 UTC