Chris,
Chris,
Re: I'm confused by this talk of identities and values. An ontology just
*is* the set of its axioms (or so sez I :-) . (01)
This is the same as the issues in non-ontology models or even DBMS and file
systems. The general approach is that there are values and identities.
There are also value types (like the number 5) where the identity and value
are indistinguishable - you can't change five and pointing to a particular
five is kind of meaningless. However pointing to a particular customer
record or a file has meaning independent of the value of that record at any
particular time. You can only refer to identifiable things. (02)
So in one case you are equating an ontology to a value type - it IS its set
of axioms. On the other hand you talk about BiologicalOntology_ver3.4.17 -
you give that value a unique identity. If I copied your ontology and people
had references to both mine and yours - they would be the same ontology, and
then you changed yours, they are no longer the same - oops. Maintaining the
integrity of identity is crucial if we are going to have reliable
references. (03)
One school of thought is that the identity of a series is made up of a
series of "snapshots". The snapshots, while they are fixed values, can also
have identity if you are able to reference one specifically. This series of
snapshots is what you have described below. Source code control systems
basically implement this paradigm for file based resources. (04)
But, I don't think equating "an ontology" to its value (the set of axioms)
is valid because it ignores identity and change over time. Even file based
ontologies have identity (perhaps not a good identity), having these issues
worked out is something a hub must address. I also don't think this is just
a "technical issue", it has to do with the semantics of the hub-lattice. (05)
-Cory
-----Original Message-----
From: Chris Menzel [mailto:cmenzel@xxxxxxxx]
Sent: Thursday, January 19, 2006 5:51 PM
To: Cory Casanave
Cc: 'ONTAC Taxonomy-Ontology Development Discussion'
Subject: Re: [ontac-dev] What is "An Ontology"? (06)
On Thu, Jan 19, 2006 at 08:22:35AM -0500, Cory Casanave wrote:
> If we were to have an ontology in our lattice of ontologies, lets say
> biological structure, would it change? (07)
Not on the definition I proposed. There would be a series of related
ontologies reflecting changes to, or the addition or removal of, axioms. (08)
> If another ontology depends on biological structure would it then
> change? (09)
A good question, for which there seem to be a couple of good answers.
Here's one idea. We can identify two types of dependence of one
ontology upon another: static and dynamic. An ontology, O1_ver5.2 say,
is statically dependent on another O2 if O2 is a particular, fixed
version in an ontology series, say, BiologicalOntology_ver3.4.17. In
this case, BiologicalOntology_ver3.4.17 should simply be considered a
subset of O1_ver5.2. In dynamic dependence, an ontology O3_ver17
imports whatever happens to be the current version in an ontology
series. It seems to me that in the case of dynamic dependence, the
imported ontology should not be considered a part of the importing
ontology O3_ver17, as, in this case, "O3_ver17" would not have a fixed
meaning. In this case, we would instead refer to, say "O3_ver17 +
O4_current", where "O4_current" at a given time refers to the current
version in the relevant series. (010)
I'm just making this stuff up, so if there are better or more mature
approaches to these issues, folks should cough them up. (011)
> Perhaps there are 2 identities here, the "slot" for the current
> specification of biological structure within the CONSMO hub and the
identity
> of the version published on a particular date by a particular authority. (012)
Right, sort of what I suggested above, if I'm understanding you
correctly. (Though I wouldn't say there are two "identities" here, just
two distinct ontologies, versions within the same ontology series (or
whatever we want to call a sequence of ontology versions). (013)
> There is also the identity of the "value" of the ontology - the set of
> axioms. (014)
I'm confused by this talk of identities and values. An ontology just
*is* the set of its axioms (or so sez I :-) . (015)
> We should know which one we are talking about. (016)
Can't argue with you there! (017)
> Configuration management is not a new problem, but the interdependence
> of axioms makes it particularly important for ontologies. (018)
Strongly agree. (019)
> If a hub is distributed and federated (like the semantic web) there
> are some unique problems to deal with (that w3c has not addressed).
> The issue I see is that with the boundaries/scope of an ontology being
> arbitrary, managing this lattice will be difficult. If the boundaries
> are arbitrary, the "slots" will change just like the versions.
> Dependency on a particular version or fixed and single-dimensioned set
> of slots would be inflexible and brittle. (020)
Well, we just can't let that happen, can we? :-) (021)
-chris (022)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-dev/
To Post: mailto:ontac-dev@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config: http://colab.cim3.net/mailman/listinfo/ontac-dev/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG (023)
|