>>Barry summarized very well what the goal of gatekeeping is: "if an OOR
>>is to be useful, then potential users need to have reliable
>>expectations as to what it will contain".
>
Pat Hayes:
>It seems to me that the best way - perhaps the only way - to
>determine this, is to read the ontology itself (or to put it through
>more dynamic tests, such as inputting to a reasoner or trying its
>effect on a battery of test cases.) What more does Barry expect by
>way of telling a potential user what it will contain? (01)
Barry Smith:
My idea is that the OOR gatekeeper function would include basic things like: (02)
a guarantee that the ontology has passed certain dynamic tests
a guarantee that the ontology is open source
a guarantee that the ontology has unique IDs for its terms and acceptable
versioning policies
a guarantee that the ontology is adequately labeled
a guarantee that the ontology has a plurality of users (03)
The OBO Foundry (http://obofoundry.org) is working towards a situation where
Foundry ontologies will have been peer reviewed for accuracy as representations
of the corresponding domain; in the even longer run towards a situation where
they will be one unique recommended Foundry ontology for each domain biomedical
domain. (04)
PH:
>>The question that we need to
>>answer is: In order to archive this goal of gatekeeping, do we have to
>>ban all ontologies that don't meet (i-v)?
>
>Clearly not. All this is required is open access to the actual
>ontology. IMO this is the only criterion whose necessity is worth
>serious discussion. (05)
BS:
Clearly so. If all that is required is open access to the actual ontology, then
there is no need for the ontology to be, in addition, a member of the OOR. It
can
be just somewhere on the web. (06)
Repositories built thus far with loose or strictly syntactic gatekeeping
criteria, e.g. the great http://www.daml.org/ontologies/ or the partially silly
http://www.schemaweb.info/, actually make life harder for potential users. Thus
the former has (when I last looked) 39 different ontologies for 'agent', all
covering more or less the same ground (some of them call agents 'agents', others
call them 'Agents', etc.). (07)
BS (08)
_________________________________________________________________
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/OntologySummit2008/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2008
Community Portal: http://ontolog.cim3.net/ (09)
|