Toby, (01)
I'll start with your closing statement: (02)
TC> In each of these, I see a upper layer of transition from
> formal semantics/ontology to folksonomy... (03)
That transitional layer is where serious difficulties lie
in software engineering, knowledge engineering, and, more
generally, *every* kind of engineering. Somehow, the precise
concepts and formalisms of each science have to be related
to other sciences and to the messy requirements of the real
world and the ways that people think about it and talk about it. (04)
TC> Classically, deep connections between systems are a barrier
> to interoperability, as they encourage the undisciplined to
> build deep dependencies. The necessary abstractions, even as
> they become more abstract, and perhaps even more semantic,
> must bear less and less connection to underlying praxis. (05)
I agree with that point, and I sympathize with the sentence
that precedes it: (06)
TC> When you toss such a standard over the transom to a
> different domain, loss of precision is not only inevitable,
> it may be a benefit. There are actual benefits to loss of
> informational integrity and completeness between domains. (07)
But I would suggest a change in terminology. The word "loss"
has negative connotations, and I'd recommend a positive emphasis.
I would focus on the gain in generality and flexibility. (08)
Your example is illustrates the issues very well: (09)
TC> I have been considering the communications between the
> engineered systems deep in buildings and the power grid during
> an emergency. These "engineered systems" use very precise, but
> unfortunately poorly abstracted interactions to operate within
> their domains. As these systems begin interacting amongst
> themselves, either tor create enterprise interactions within
> the building, or to support smart grid / smart building
> interactions, they are moving to formal ontologies based upon
> very large scale semantic efforts such as buildingSmart, and
> formal IDMs between adjacent semantic areas. (010)
Such interactions occur in every real-world problem. I often
use the example of the medical domain, which involves multiple
specializations: every specialized subdomain of medical
practice has a different perspective, and the physicians and
surgeons must interact and communicate with each other and
with GPs, nurses, and patients with very much more diverse
perspectives. Furthermore, all of them have to interact
with pharmacists, administrators, and the people who maintain
the facilities, ranging from computers and MRI scanners to
the air conditioning and lighting for the operating room. (011)
Somehow, they all have to communicate and collaborate. (012)
TC> That upper level ontology, that discovery and communication
> of value, often feels to me more like a foksonomy. (013)
I agree that it has the feeling of a folksonomy, or more generally
the informal semantics of natural languages. But I would not call
it the upper level, because it gets down to very specialized lower
levels as well as upper levels. (014)
I would call it more of a taxonomy than an ontology. But whatever
you call it, all the formalisms must, eventually, be related to
that taxonomy in order to communicate in NLs. (015)
We can't avoid the transitional layer or pretend that it's outside
the scope of what ontologists must deal with. (016)
John (017)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (018)
|