RE: SUO: Latest Draft Ballot Questions
Dear John,
I recommend not specifying ISO 11179. You might end up deciding to use
it, but having taken a look, I consider it inappropriate unless you
are only concerned with data models.
There are several alternative metadata standards that can be considered,
perhaps the best known is the Dublin Core.
Matthew West
Principal Consultant
Shell Information Technology International Limited
Shell Centre, London SE1 7NA, United Kingdom
Tel: +44 20 7934 4490 Other Tel: +44 7796 336538
Email: matthew.west@shell.com
Internet: http://www.shell.com
> -----Original Message-----
> From: sowa [mailto:sowa@bestweb.net]
> Sent: 23 May 2003 19:00
> To: jim.s3@juno.com
> Cc: standard-upper-ontology@ieee.org
> Subject: Re: SUO: Latest Draft Ballot Questions
>
>
>
> Jim,
>
> I revised the motion to accommodate suggestions that have
> been made during the discussion period.
>
> The major addition responds to the suggestion by Frank Farance
> that we should follow the guidelines of ISO/IEC 11179 for
> metadata registries. Following is the URL for that standard:
>
> http://farance.com/metadata/iso_iec_11179-3_2003.pdf
>
> As a result, I have introduced the term "ontology registry"
> for the collection of modules in the previous motion. I kept
> the name "module", but added the clause "Each module shall
> consist of a theory together with documentation and other
> metadata as specified by IS 11179-3."
>
> This change does not introduce anything that we would not
> want to include anyway, such as the notion of "stewardship",
> which means that each module shall have an organization that
> is responsible for administering it and providing contact
> information for anybody who wants to ask a question about it.
>
> The other addition is to require the axioms and definitions
> to be stated in some language that conforms to the Common Logic
> framework. That includes KIF, CGs, and OWL. To accommodate
> OpenCyc, I added the following clause: "As part of this project,
> the CycL language or some subset of CycL suitable for specifying
> the OpenCyc contributions will be defined as a CL-conformant
> language."
>
> The other changes are clarifications of the earlier version.
>
> John Sowa
>
> ______________________________________________________________
>
>
> Should the IEEE P1600.1 Standard Upper Ontology Working Group
> (SUO) commence work on a project to develop a standard for
> ontology registration and specification, according to the
> following guidelines:
>
> 1. The standard shall be based on the contributions of three
> SUO candidate projects: IFF, OpenCyc, and SUMO. As the
> standard develops, the SUO working group may decide to
> include appropriate material from other sources or to
> modify or omit any material from the candidate projects
> that is not deemed to be suitable.
>
> 2. The standard shall specify an ontology registry that conforms
> to ISO/IEC IS 11179-3 for metadata registries, but with
> appropriate extensions that may be required for defining
> ontologies and for relating them to one another.
>
> 3. The ontology registry shall be organized as a collection
> of modules, each of which is administered under a single
> stewardship, as specified by IS 11179-3. The modules
> shall be related in a generalization/specialization
> hierarchy. Each module shall identify every module in
> the registry that is an immediate generalization; it
> may identify one or more modules as specializations.
>
> 4. Each module shall consist of a theory together with
> documentation and other metadata as specified by
> IS 11179-3. The theory shall consist of axioms and
> definitions stated in a language that conforms to the
> Common Logic (CL) standard, which is currently under
> development. CL-conformant languages currently include
> the Knowledge Interchange Language (KIF), conceptual graphs
> (CGs), and the Web Ontology Language (OWL). As part of
> this project, the CycL language or some subset of CycL
> suitable for specifying the OpenCyc contributions will be
> defined as a CL-conformant language.
>
> 5. The standard shall include the specification of a methodology
> for testing the theory part of any module for consistency,
> relating theories to one another in the generalization/
> specialization hierarchy, and combining two or more theories
> to derive a new theory that is larger and more specialized
> than the theories from which it was derived.
>
>
>