FW: Cross references to the Ontology

[Forwarded to the UWA public list in respect of the UWA's public
status.]

 

The DDWG is having some issues regarding the mapping of their vocabulary
entries to entries in the UWA ontology. The DDWG API design calls for
properties to belong to certain aspects. The actual aspects have not
been defined yet, but these are likely to include things like "display",
"camera" and "browser" (to name just a few). A property in the DDWG
viewpoint is something generic like "vendor", which may be applied in
one or more aspects. In that way one can have a vendor of the browser,
separate from the vendor of the platform and so on.

 

Sample Question: To what part of the ontology should the DDWG's property
called "vendor" be mapped, when no particular aspect is being
identified?

 

This question arose because at the recent DDWG weekly call we noted that
the DDWG's properties were actually at a higher level than (some of) the
items in the UWA ontology. To some, on first reading, this might seem to
be counter to what one would expect of a vocabulary and ontology. (That
is, the vocabulary would be specific, while the ontology would be more
generic.) However, the DDWG's specificity is achieved through the
introduction of aspects, which do not at this point appear to have a
corresponding ontological representation.

 

To make this concrete, earlier this week I circulated a table that
showed how some of the DDWG property/aspect combinations would be mapped
to corresponding entries in the ontology. There was support during the
DDWG call for this approach. It was proposed that I circulate the email
and attachment to the UWA for consideration.

 

It is now our intention to publish a first draft of the DDWG's
vocabulary document, but the issue of cross-referencing will only be
mentioned as part of editor comments in the document, and will remain so
until we have come to a mutual understanding of how to proceed.

 

Note: it is expected that the DDR API will support the idea of querying
on the basis of a DDWG Property alone (i.e. without mentioning an
aspect) in which case it will be up to the implementation of the API to
determine what aspect should be assumed, or throw an exception if no
such determination can be made. We acknowledge that the presence of such
potential implementation differences causes issues for interoperability,
but this is inevitable and not altogether undesirable in light of
commercial differentiation. However, a more deterministic behaviour is
possible when aspect is included as a query parameter, and this one can
expect to form the basis of interoperability requirements.

 

The DDWG would like to hear from UWA on this issue, particularly in
respect of providing bindings from DD vocabulary properties to UWA
properties, bindings from DD vocabulary properties+aspects to UWA
properties, and possibly from DDWG vocabulary aspects to UWA properties
(or other information within the ontology).

 

---Rotan (DDWG chair).

 

From: Rotan Hanrahan 
Sent: 28 November 2007 14:02
To: DDWG
Subject: RE: Cross references to the Ontology

 

Attached is a start of a mapping that incorporates DD property, DD
aspect and UWA property. The next step would be to see how the UWA
"browserUsableDisplayPixelsX/Y" would fit in, and so on.

 

I selected the aspects off-the-cuff. They are certainly not formalised,
and I'm wondering 1) should an initial set be formalised, 2) what should
be in the set and 3) where do we document this set?

 

I was hoping that this approach might help shed some light on the
bringing together of the FPWD of the vocabulary (or vocabularies, if
that's what we end up doing).

 

---Rotan.

 

[... remaining thread deleted ...]