[Top] [All Lists]

Re: [ontology-summit] Potential Tracks for Ontology Summit 2013

To: Ontology Summit 2013 discussion <ontology-summit@xxxxxxxxxxxxxxxx>
From: Simon Spero <sesuncedu@xxxxxxxxx>
Date: Wed, 12 Dec 2012 09:47:00 -0500
Message-id: <CADE8KM51weDzXdSKB8BaSFvpX4W4tdFRi+-0dSM_2MnYpU7TBQ@xxxxxxxxxxxxxx>
On Tue, Dec 11, 2012 at 12:59 PM, Michael Gruninger <gruninger@xxxxxxxxxxxxxxx> wrote:

5. Environments for Developing and Evaluating Ontologies
- what are best practices for evaluation that we can adapt from software engineering, particularly with distributed open-source software development?

I wouldn't necessarily restrict this purely environment here, but one of the most critical requirements  for using most  agile family methods is the ability to test developed artifacts very quickly and automatically, at each stage of the development process.  This can include unit tests for individual modules in an ontology, integration tests for combinations of modules, specification/translation of requirements into tests, etc. 

Specifying tests for verification at the start of a sprint, and providing ways to validate the results at the end are key enablers.  

Areas of special difficulty in creating these tests for ontologies is that the ontology is where agreements about what various concepts "mean", so the first few iterations may be especially unstable. Also, where development is taking place across multiple independent groups, with different internal ontologies that need to map in to a common ontology, it may be difficult to test the effect of refactorings, or to determine where parts of the separate ontologies have enough commonality that they can be refactored into a separate module (and whether the ontologies should use the new module, or should map into it). 

A lot of this cuts across the other topics, but the points of emphasis are the ability to generate  tests from partial requirements, and before developing the ontology, the ability to run most tests extremely quickly, and the ability to handle iteratively refined requirements.

TDD and BDD have potential, though the need for a "ubiquitous language" is at odds with an ontological approach (or possibly is what the ontology provides).


Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2013/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2013  
Community Portal: http://ontolog.cim3.net/wiki/     (01)
<Prev in Thread] Current Thread [Next in Thread>