On 30 Mar 2009, at 16:23, Oliver Ruebenacker wrote:

     Hello Bijan, All,

On Mon, Mar 30, 2009 at 7:40 AM, Bijan Parsia
<bpar...@cs.manchester.ac.uk> wrote:
On 29 Mar 2009, at 16:48, Oliver Ruebenacker wrote:
My first advice would be to develop a clear set of requirements,

  That is probably the most used phrase in the Semantic Web community.

And least adhered to?

What's your problem with providing application requirements?

We should invent a catchy acronym for it. Maybe DACSOR?

perhaps in
terms of the ASK/TELL paradigm, for the role of the ontology in typical
Systems Biology applications you are trying to support.

  I have no idea what that paradigm is.

http://portal.acm.org/citation.cfm? id=1064.1066&coll=GUIDE&dl=GUIDE&CFID=15151515&CFTOKEN=6184618

Classic applications of (description logic based) ontologies include
terminology development (e.g., see the caImages site), form management (see
Galen), and data integration and navigation (see Tambis).

If your applications resemble these (or other standard applications) then it's easy to give a wealth of advice, methodologically and technically. If your application is radically different, then more work has to be done to
understand the application requirements.

  An ontology should be independent of a particular application.

First, there's a big difference between being tied to a particular application and being informed by intended applications. I sincerely doubt the general utility of an ontology designed without *any* thought to the class of applications for which it is intended to be used. Advice that is not at all informed by the broad characteristics of the applications for which the ontology shall be used is worthless.

  But if you are interested in what type of application I am working
on check http://vcell.org/biopax

It's not obvious to me the role of ontologies in that application. Are you using them for data integration?

As far as I can tell it's a format converter and visualizer. No ontologies needed at a first approximation.

Cheers,
Bijan.


Reply via email to