[Top] [All Lists]

Re: [ontolog-forum] Quote for the day

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Sun, 02 Jan 2011 16:27:41 -0500
Message-id: <4D20EDCD.7000308@xxxxxxxxxxx>
Doug, Rich, Jack, Pavithra, and Ed,    (01)

> This suggests to me that context is important in the ontology definition,
> not just in its use.  At the most basic level, the ontology would not
> include many detailed rules.  Such rules would exist in "theory" contexts
> which could be included with the more basic ontology as desired.
> Different theory contexts could be specified for different time periods,
> jurisdictions, or purposes.  E.g., a rental car company would use theories
> of an ontology of automobile brands that a repair shop or a dealer would
> not use, and they would use other theories covering the same makes and
> models.    (02)

Yes.  That's the point I was trying to make.  And the example of
a "car ontology" emphasizes the issues:  businesses that rent, sell,
and repair cars would require very different ontologies with very
different rules (or axioms) about what to do with and about cars.
Any commonality would be a rather superficial taxonomy, not a set
of detailed axioms.    (03)

But the example I cited about economics shows that even for a similar
context, different people can make very different assumptions.  Here
is the full paragraph that includes the sentence I quoted:    (04)

> Economists argued all through 2010 about the alleged benefits of the
> $814 billion economic stimulus enacted in early 2009. It was either
> a big waste of borrowed money or the thing that staved off a real
> depression. If there was a benefit to this story, it was that it
> demonstrated once again the limited utility of economic modeling:
> Models generally return results close to the assumptions of the
> economists who write them.    (05)

 From an article in _Barron's_ by T. G. Dolan:    (06)

http://online.barrons.com/article/SB50001424052970203423004576049913136801914.html    (07)

> Yes, and its not just economists; I managed a simulation group way back
> when, and every project was designed to meet the simulation expectations of
> the customer.  The whole purpose of a simulation is to prove or disprove a
> theory by use of the simulation.  So if the simulation DOESN'T convince the
> customer that s/he was right all the time, it probably wasn't well
> implemented.    (08)

If you replace the word 'customer' with 'politician', it's exactly
the same point:  the people who ask for the information tell the
person who develops the simulation model what assumptions to make.
The answers won't tell them anything they don't want to hear.    (09)

> Perhaps we should shift from talking about ontology as a static entity
> to talking about the resultant from meshing a purpose-ful ontology with
> a contextual ontology. For example, a model of an (intended) problem
> intervention system describes the mesh of the problematic situation ontology,
> the pragmatic technologies ontology, the sponsors preferences ontology,
> the expected operators' knowledge (both sentient and encoded) ontology,
> and other semiotic sets. The resulting system model is a situated ontology.    (010)

That observation is consistent with the the points by DF, TGD, and RC.    (011)

The following quotation is from the software engineering thread:    (012)

> How about considering an ontology to be semiotically equivalent to a model
> of a system? Would the principles, rules and metrics for one inform the
> other? This is not to say that we already know the principles, rules and
> metrics for discovering systems but it does suggest that those focused
> on ontologies might gain from knowledge exchange with those focused on 
>systems.    (013)

The word 'model' has several different meanings.  I consider the
word 'model' as used by economists and software engineers to be very
similar to what people in this forum call an ontology.    (014)

But in logic, a *model* of a set of axioms is some structure or
system that makes those axioms true.  Since there are many logicians
who subscribe to this list, I usually avoid the word 'model' except
when I am talking about the logic.    (015)

> Here is a nice explanation of Ontology spectrum..
> http://techwiki.openstructs.org/index.php/A_Basic_Guide_to_Ontologies    (016)

I agree that is a good summary of various issues, written by Leo Obrst.    (017)

But in those slides, Leo did not address the issues about context,
purpose, and competing expectations.  Leo's chart of the spectrum
from "weak semantics" (few axioms) to "strong semantics" (detailed
axioms) did not mention the problems that arise when people with
different contexts or purposes make different assumptions.    (018)

> Any Economic model arrayed about an equals sign is inevitably
> socialist. - design economist means and ends are unified in the tool.    (019)

The quotation by TGD is from Barron's, which is a weekly news magazine
produced by the Wall Street Journal.  It's hard to find anything that
is closer to the capitalist heartland than the WSJ.    (020)

John    (021)

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (022)

<Prev in Thread] Current Thread [Next in Thread>