I think Joel has described significant features in the holy grail of
ontology repositories -- one of the challenges having been just the
'syntactical change' issues he identifies, which means change tracking
in the optimal repository has to be concept-aware (or triples-aware?)
rather than text-based. Unpleasant but apparently unavoidable. (01)
Along those lines, in the 'which Protege' department another option is
the collaborative version of Protege. (Unless that is a proper subset
of 4.0, but we'd still have decide whether to use those features.) (02)
As I also worry about endless discussions of strategy vs. actually
moving the ontology work forward, I prefer the not-so-pervasively-
cooperative model that David Price describes. But I do hope the
leaders of the activity find a way to regularly expose significant
changes and have them viewed/reviewed/commented on by this community,
rather than doing Big Bang releases with all sorts of things
different. Just my personal wish, albeit one with its own costs in
development time. (03)
I think tracking issues in a system like Trac is a powerful addition
to the process, in any case. But its value will depend heavily on the
usage patterns of the key people maintaining the ontology. (04)
John (05)
On Jul 14, 2009, at 9:30 AM, David Price wrote: (06)
> I don't think this is going to be, or should be, a "cooperative
> development" in the way that Joel seems to suggest. In the proposed
> plan
> there are only one or two editors who actually modify the ontologies.
> Everyone else develops issues which are discussed and if accepted that
> are implemented by the editors. This is a "standards-making 101"
> methodology, along the lines of what happens in OMG and ISO, which I
> fully support. So, my view is that this process really isn't very much
> like software development at all.
>
> Cheers,
> David (07)
On Jul 14, 2009, at 9:13 AM, Joel Bender wrote: (08)
> Evan,
>
>
>> There is no "canonical form" for serializing an OWL model in rdf/xml.
>
> A huge hole for cooperative development of a model.
>
>> Do people anticipate changes in the serialized form to be a problem?
>
> Yes. I would like to use tools that can track changes to a set of
> documents, knowing who's been making changes and what changes have
> been
> made.
>
> I'm offering to put up a server that is running Trac [1] with a
> Mercurial [2] repository that would allow us to cooperate on
> documents.
> There is a bit of a learning curve, but I think we can cooperatively
> work out issues in short order.
>
> Anyone can put anything in the repository they want, and key for me is
> that I can get an updated version of what everyone has contributed
> in a
> simple way.
>
> As part of this development, I would also like to use tools that can
> "build" and "test" this ontology against a set of use cases. The
> errors
> and warnings from the build/test/validate test results can feed into
> tickets to make sure issues and problems are assigned to specific
> individuals or groups and resolved.
>
> For example, if I'm running TopBraid and change something to an
> owl:InverseFunctionalProperty and post it, everyone is notified of the
> change, that I made it, and when I run Pellet [3] or some other
> reasoner, it tells me I'm an idiot. Or something more kind.
>
> If I make a change to our eventual publication (and please allow it to
> be RTF or Open Office or something other than binary Microsoft Word
> files), that should be clear as well.
>
> I see a lot of this work as very similar to programming, but then
> again,
> I'm a programmer :-).
>
>
> Joel
>
> [1] <http://trac.edgewall.org/>
> [2] <http://mercurial.selenic.com/wiki/>
> [3] <http://clarkparsia.com/pellet/>
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/uom-ontology-std/
> Subscribe: mailto:uom-ontology-std-join@xxxxxxxxxxxxxxxx
> Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/uom-ontology-std/
> Shared Files: http://ontolog.cim3.net/file/work/UoM/
> Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard
> (09)
John (010)
--------------
John Graybeal <mailto:graybeal@xxxxxxxxx> -- 831-775-1956
Monterey Bay Aquarium Research Institute
Marine Metadata Interoperability Project: http://marinemetadata.org (011)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uom-ontology-std/
Subscribe: mailto:uom-ontology-std-join@xxxxxxxxxxxxxxxx
Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/uom-ontology-std/
Shared Files: http://ontolog.cim3.net/file/work/UoM/
Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard (012)
|