On Wed, March 14, 2012 11:49, Cory Casanave wrote:
> John,
> With respect to your points:
>> 4. More generally, any "universal" ontology intended for a wide range
>> of applications will have to be underspecified in order to support
>> an open-ended range of models. (01)
> [cbc] Are these "universal" ontologies indeed underspecified or are they
> in fact properly specified for their scope? (02)
I would suggest that the answer could be "both". I suggest that they are
properly underspecified, to allow different applications to merge their
applications rules and constraints with the more general ones. (03)
A general purpose "universal" ontology would properly provide a sketch
of the concepts for a field. An application would add rules, constraints,
additional classes, relations, and individuals needed for its purposes. (04)
> That is, is it necessary that
> these ontologies not say things that are still assumed true for the
> "shared concept" but that we have no way of saying this in a general way? (05)
Whether they *can* be said in a general way, is not important. What is
important is whether it is useful for stating them in a theory that everyone
downloading the ontology would receive. For example, jewelers,
circuit designers, metallurgists, and bullion dealers might all wish to take
concepts for the elements Gold and Silver from a universal ontology. But
each industry would be interested in different properties for these metals.
Even if luster, price, triple point, and electrical properties can be
stated in the
universal ontology in a general way, those in each field would not want that
general information stored in their system -- if only because it would tie
up memory and possibly slow down inference. (06)
> Or, are we simply recognizing that we must be careful about the
> commitments we make and that general ontologies have, in the past,
> tended to over comment? (07)
Do you mean "over comment" -- they state things which are not logically
encoded -- or "over commit" -- they commit to statements which are not
true in all cases? I guess you could have a combination of these two
properties if a general ontology makes too strong a statement in a
comment, but does not logically encode it. (08)
Cyc, as the most detailed "general ontology" has flaws with over
commenting, under commenting, over committing, and under committing,
in different (and some cases the same) areas. Although, Open Cyc has far
less over committing and far more under committing since it has been
stripped of its rules. (09)
> 5. Each domain of application will need to add more information
> that specializes the definitions in the general ontology. And
> the details added for one domain or one kind of application are
> likely to be inconsistent with the details needed for others. (010)
> [cbc] So are these in fact then the same concept or are they derivative
> concepts, perhaps sharing the same label? (011)
This depends on your definition of "the same concept". Some philosophers
would hold that you can't have "the same concept" in two different contexts.
However, for the purposes of shareable computer ontologies, i suggest that
it is useful to treat them as the same concept. (012)
The number "5" has different properties for five-year olds, set theorists,
and bingo players, but society has deemed it to be a single concept, not
multiple concepts with the same label. (013)
-- doug (014)
> -Cory
>
> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
> [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F. Sowa
> Sent: Wednesday, March 14, 2012 11:42 AM
> To: ontolog-forum@xxxxxxxxxxxxxxxx
> Subject: Re: [ontolog-forum] Constructs, primitives, terms
>
> On 3/13/2012 12:22 PM, Rich Cooper wrote:
>> But the Amazon solution is very much made up of unstructured text, and
>> the product search is mostly using key words like Google uses for
>> search. So in that sense, Amazon's database design is very basic, and
>> its keyword search is helpful but not new technology. It's a version
>> of SEO practices: using keywords to describe your product the way your
>> customers search for it.
>
> I used Amazon.com because their DB schema is probably the most widely used
> formal ontology in the world. Any additional info in NL text or pictures
> is not part of the ontology that supports interoperability with their
> suppliers.
>
>> Is that your view of how future ontologies will be used for
>> interchange, i.e., a skeletal database model with active users
>> participating in the database for their own purposes, and using
>> unstructured text to distinguish their products from others?
>
> That's one kind of use. But many applications can profit from much more
> detailed ontologies. However, the detailed specifications are likely to
> be limited to domain-dependent microtheories.
>
>> how is that influenced by ontological thinking?
>
> Everybody has different thoughts about ontology. As I said to Ed B, I was
> only trying to make five general points (copy below).
>
> For more, see the following papers:
>
> http://www.jfsowa.com/pubs/futures.pdf
>
> http://www.jfsowa.com/pubs/fflogic.pdf
>
> John
> _____________________________________________________________________
>
> 1. Interoperability requires much more than unique identifiers that
> point to definitions shared by all applications.
>
> 2. For example, the shared definitions that support interoperability
> for transactions through Amazon.com include only as much detail
> as necessary to support Amazon operations -- and no more.
>
> 3. In different applications, those weak definitions are probably
> inadequate for messages that use the same terms in contexts with
> more detailed specializations (i.e., axioms or constraints).
>
> 4. More generally, any "universal" ontology intended for a wide range
> of applications will have to be underspecified in order to support
> an open-ended range of models.
>
> 5. Each domain of application will need to add more information
> that specializes the definitions in the general ontology. And
> the details added for one domain or one kind of application are
> likely to be inconsistent with the details needed for others.
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/ Community Wiki:
> http://ontolog.cim3.net/wiki/ To join:
> http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
>
> (015)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (016)
|