On Feb 6, 2004, at 1:34 AM, Patrick Cassidy wrote: (01)
>
>> I am not totally clear on the distinctions between formal logic
>> representation and symbolic reasoning systems.
>
> Adam Pease appears to have some fairly specific requirements for
> what can pass as a "representation" and doesn't think that
> Protege qualifies. I like to make a clear distinction between
> (1) a static set of data in some machine-readable format and
> (2) the procedural code which interprets the data and reasons with
> it. There are, however development systems which include both
> data-storage and some reasoning ability. Protege is one of those,
> but the reasoning available in Protege is minimal and the formal
> logical model will not support full first-order logic --
> therefore unless it is modified, a KIF (or SKIF, a variant of KIF)
> file will not necessarily be interpreted properly by the
> internal reasoning facility of Protege.
> My goal is to add **just enough** additional reasoning ability
> to Protege in order to:
> (1) provide a perspicuous and logically sound visual representation
> of the knowledge available in a SKIF file;
> (2) enable convenient editing of a knowledge base and allow it to
> be re-exported to its original SKIF format.
>
> Thus far, I think I have accomplished (1), but doing (2) is months
> away. So for now, the Protege facility can be used as a browser, but
> not as an editor -- it can't re-export into SKIF format. As a
> browser, it helps *me* to understand relations within an ontology
> in a way that the other browsers we have available don't. I hope
> it will also be useful to others.
>
> As for RDF/OWL, I haven't tried to use these representations, so
> I can't speak knowledgeably about them. There is a Protege
> plug-in under development that can import and export OWL-format
> files. So if and when I or someone else develops a way to export
> Protege files to KIF, Protege might be usable as an interlingua
> to move from SKIF to OWL format -- though a direct conversion might
> be better when it is developed. But there are probably lots
> of sticking points on specific logical structures that could
> make full translation difficult. That will become evident over
> time. For me, I only get the motivation to learn new
> file formats if I have on hand an application that requires
> it. If anyone knows of a freely-available application
> that uses OWL (and does something useful with the data),
> I would appreciate the reference.
>
> For now, in spite of spending time on the SKIF/Protege format
> conversion, I am most interested in the actual conceptual content
> of the knowledge bases we are developing, with a strong preference
> for rapidly coming to an agreement on the upper-level concepts
> that can handle a wide range of specific domains. Invoices looks
> like a good specialized domain to start with, and so we proceed.
>
> Pat
> ================================
>
> Kenneth Fields wrote:
>
>> This is optimal! This would settle the issue of browsing - Sigma
>> browser.
>> I am not totally clear on the distinctions between formal logic
>> representation and symbolic reasoning systems. I don't ask for a
>> tutorial on that matter here - I'll go do my homework on that. But I
>> assume KIF is the standard for the former and RDF / OWL for the
>> latter.
>> In thinking about deliverables, we want to the ontology representation
>> to be both logically solid for computation and semantically (Web)
>> useful.
>>
>> What does Adam mean when he says he has 'built' ontologies in Sigma -
>> I
>> don't see the editing capability like Ontolingua does.
>>
>> Ken
>> (02)
Powered by MessageSoft SMG
SPAM, virus-free and secure email
http://www.messagesoft.com (03)
_________________________________________________________________
Message Archives: http://dao.cim3.net/forum/dao-forum/
Subscribe/Unsubscribe/Config: http://dao.cim3.net/mailman/listinfo/dao-forum/
Shared Files: http://dao.cim3.net/file/
Community Wiki: http://dao.cim3.net/wiki/
Community Portal: http://dao.cim3.net/
To Post: mailto:dao-forum@dao.cim3.net (04)
|