Kingsley and Michael, (01)
I'm responding to both notes in this thread, which includes the
same issues under a subject line that is more meaningful. (02)
KI
> I don't dispute the obliteration of TimBL's vision. I don't dispute
> the essence of your meme, I am just reiterating to you that some
> of us have implemented these standards in ways that align with what
> you seek i.e., work with existing systems. (03)
That's good. I support Tim's idea of using RDF as a way of representing
simple assertions in web pages. RDFa and Microdata are also good ways
of annotating web pages. I am also happy with the idea of having DL
logics as *one* kind of representation, when it is useful. But any
strategy that ignores mainstream IT will be ignored by the mainstream. (04)
KI
> Likewise, I don't allow my company to make or sell products that
> ignore the reality of legacy software (05)
I wish Tim B-L had stated that as a prerequisite for the DAML project. (06)
MB
> I think [Martin H's] main point is about informal underspecification
> - not formal underspecification. (07)
Nothing implemented in computer software is informal. The computer will
always do something very precise. The only question is whether what
it does has any relationship to what was intended. (08)
MB
> But as it must include the intended models, there should normally be
> a low number of axioms. (09)
Yes. And the smaller the number of axioms, the more likely that people
will produce data for which the axioms are true. The major flaw in many
formal ontologies is the assumption that people who enter data will
read, remember, and observe the definitions. (010)
JFS
>> But an underspecified ontology, by itself, can't support an application.
>> You need either (a) a richer logic or (b) a programming language. (011)
MB
> Are you talking about generating applications from axioms ? Or from
> formal application descriptions ? If yes, could you provide links ? (012)
There is more legacy software in the world than new software. So the
most important axioms are the ones that describe the typical way that
people would interpret terms like 'AutoDealer', 'FireStation', 'Time',
'GovernmentOrganization', or 'FoodEstablishment.CafeOrCoffeeShop'.
Those are typical terms from http://www.schema.org/docs/full.html (013)
Note that none of those terms have any definitions other than their
placement in the hierarchy. More detailed definitions and axioms
can be used, but only for data that is generated and/or verified
by thoroughly tested computer software. (014)
MB
> Are there other reasons for axioms other than entailing new
> facts and data cleaning ? (015)
I'm using the word 'axiom' as a generic term for any statement used
by a logic-based system other than ground-level clauses. It would
include the rules of a rule-based system or the constraints on a
database. If your procedural software is formally specified, the
preconditions and postconditions for functions and procedures
would be called axioms. (016)
John (017)
_________________________________________________________________
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 (018)
|