W3C home > Mailing lists > Public > public-prov-wg@w3.org > February 2012

resurrecting prov-o components design

From: Timothy Lebo <lebot@rpi.edu>
Date: Thu, 23 Feb 2012 20:57:37 -0500
Message-Id: <4A471518-7935-4D95-91A8-601A3C875743@rpi.edu>
To: "public-prov-wg@w3.org Group" <public-prov-wg@w3.org>
PROV-O (and -wg),

As we're juggling the variety of design factors for prov-o, I'm wondering if we should revisit the components design and consider re-adopting it.

Using components, we could easily produce "profiles" that utilize a variety of OWL axioms.
We could create an OWL-RL compliant version.
We could create a version without prov:qualified subproperties (and one with them).
We could create a version with _only_ core constructs (and not the common relations).

And each of these would be created by including the appropriate underlying subsets for the profile. That is, "prov:used a owl:ObjectProperty" is stated once and included in many profiles.

Further, we can begin the very much overdue need for a collection of __concrete examples__ that provide a systematic basis for determining how "done" we are with the owl ontology.
(This is a for loop, folks. And gives hard evidence for what we are doing wrong and right)

The component design is discussed at http://www.w3.org/2011/prov/wiki/PROV_OWL_ontology_components

The biggest challenge with adopting this approach is that PROV-O team will NOT be able to "edit" a monolithic ontology file within protege.
But I think the total time spent thus far editing the ontology indicates that the _edit_ time is not the challenge, it's the design time that is consuming us.

Received on Friday, 24 February 2012 01:58:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:58:12 UTC