Names in ontologies (was: Re: Lexical representations)

Nick Gibbins  <nmg@ecs.soton.ac.uk> writes:
>Leo Obrst <lobrst@mitre.org> writes:
>
>>  Jeremy's new rendition is much better. However, the conflation with
>>  URIs is not. Every ontology needs "display names", for many
>>  reasons. We had ample support of this in the content
>>  interoperability use cases. And these "lexical representation" names
>>  are only the beginning, and apply in a limited (but very important
>>  context). Why not tie this in to the "tags" we already need? I.e.,
>>  "true" metadata about objects.
>
>I'd be against this. The association of human-readable labels with
>ontology terms is a different task to the attachment of provenance to
>statements, and should be represented in this doc by a separate
>requirement (not least because it seems likely that they will be
>supported by different mechanisms in the final language).

I agree with Nick. There seems to be a lot of distinct things being 
put together here, maybe it would be useful to try to list them. 
First stab:

We need ways in an ontology to indicate the following things about a 'name'.

1. It is the same name as used in some other ontology, and being used 
here also. (For ontology machine use, part of the basic web syntax 
requirement) RDF does this by using the same uriref, which is 
controversial.

2. It is being used here to refer to [part of] another ontology. 
(Ditto, eg for asserting provenance, but also for importing, 
commenting on, etc. ) DAML does this by  using URLs.

3. It refers to [part of] a document [in some media type/mime 
format]. (For metadata.) RDF has no official position, but in 
practice people use URLs plus fragIds, which is indeed fragile.

4. It is a rendering [in some media type?] [in some I18N language?] 
of the name for human use (Zilch to do with ontologies, but needs to 
be kept 'with' the concept name)

Any more??

Now all kinds of otherwise opaque issues become clearer, eg should 
there be a sense-1 representation in the ontology itself of the 
'types' of the human-readable name in sense 4 ? Should the sense-2 
usage be a special case of the sense-3, or treated differently? How 
do we distinguish sense-1 from sense-2 ?

Pat
-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes

Received on Monday, 25 February 2002 19:45:23 UTC