Hans and Ed, (01)
Arguing about ambiguities in the word 'context' is fruitless.
But the general issues about the multiple senses of words --
or even better *microsenses* -- are fundamental. (02)
HP
> At least we can describe the institutional and domain contexts
> for the systems that interact over the network ...
> as opposed to leaving context implicit. (03)
EJB
> part of our problem was accommodating multiple viewpoints in that space:
> design engineer, engineering software supplier, plant builder, equipment
> vendor, plant operator. So, of course, the ontology supports all of them
> not equally badly, and matches none of them. (04)
I partly agree with both of you, but with qualifications. (05)
As another example, consider the issues in "medical informatics" of
a patient who is admitted to a hospital. The general practitioner,
surgeon, nurses, pharmacists, admin personnel, and the patient must
communicate. (06)
The same speaker may use different words or the same words in different
senses when talking about the same case with different people for
different purposes. Replacing words with IRIs is worse than useless,
because it gives a misleading impression of precision. (07)
EJB
> It takes months to get a sense of the differences in viewpoints, because
> the people who hold them only tell you how they see the space, do you
> (sooner or later) see the conflicts, but you have to tease out the
> "principles" involved in their views. (08)
That's a weakness of batch-mode knowledge acquisition. Even if you
could "tease out" those general principles from some "expert", the
same expert may use the same words in different senses when talking
with different people or with the same people about different cases. (09)
HP
>> When e-retailers adopt the Amazon framework and business model
>> to sell their stuff on the Internet... (010)
EJB
> They are adopting the Amazon practices that they expect will
> work well for their "unique" application. (011)
Actually, the Amazon DB schema is an extremely underspecified ontology
for any of the products they sell. The only information that schema
specifies in detail is about methods of buying, selling, payments, and
shipping. Most of that information is orthogonal to any description
of the products and their use. (012)
I believe there is a lesson in Amazon's success: To make independently
developed systems interoperable, use an underspecified ontology that
minimizes the amount of detail on which agreement is required. (013)
EJB
> In a narrow domain in which the meaning of "scope" terms is presumably
> well-defined. (014)
The only narrow domains occur in oversimplified examples. I strongly
agree with your quotation from Selden Stewart: (015)
SS
> There is nothing so pernicious to large-scale designs as the seductiveness
> of simple examples. (016)
Unfortunately, most of the examples used to develop ontologies are
oversimplified. I would have more faith in something as vague as
WordNet than in a typical formal ontology with "precise" IRIs. (017)
John (018)
_________________________________________________________________
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 (019)
|