[Top] [All Lists]

Re: [ontolog-forum] Terminology and Knowledge Engineering

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Wed, 25 Jan 2012 13:53:03 -0500
Message-id: <4F204F8F.4030605@xxxxxxxxxxx>
John (Bottoms) and Ed,    (01)

> As I understand it, NATO uses Simplified English to machine translate
> weapons system documentation to Simplified <NATOlanguage>. These require
> the addition of one or more technical vocabularies...    (02)

There has been a lot of work on Controlled Natural Language (CNLs).
Some of them are used primarily for communication among people,
some for communication between people and computers, and some
for communications with both.    (03)

> What we don't have is a partitioning of a natural language. Many have
> proposed new languages and I tire of examining them. But, I would be
> very interested in seeing a reasonable partitioning or a set of rules
> with 100% coverage that allows partitioning.    (04)

There is no way to have a universal partitioning of everything that
can be said in any NL.  But there are many useful CNLs that have been
implemented.  Following is a paper in which I discuss these and
related issues:  http://www.jfsowa.com/pubs/futures.pdf    (05)

> And my point is that these OWL ontologies are created by people who are
> just learning to formalize concepts.  This is Mozart Symphony #1 at an
> age between 20 and 40.  Yes, they are little more than terminologies.
> Be thankful that these folk are using OWL, instead of XML Schema or UML
> or ISO 11179 or SBVR.  At least with OWL the models have well-defined
> formal semantics...    (06)

My point is that Aristotle defined the formal semantics for the subset
they're using, that subset was taught to entering freshmen at every
university from the 13th to the early 20th centuries, and they used
a very readable CNL as their notation.    (07)

I blame Bertrand Russell for the downfall of logic in the 20th century.
He considered Aristotle's logic to be a competitor to symbolic logic,
and he wanted to get universities to switch.    (08)

Bertie got half his wish.  They stopped teaching Aristotle's logic.
Now, the overwhelming majority of university graduates have no training
in logic of any kind.  (And it shows.)    (09)

> And with OWL, these models have the possibility of extended axiomatization,
> as these people, or their peers and successors, get further up the learning
> curve.  An OWL terminology glass is half full.    (010)

Joe S. mentioned Formal Concept Analysis as an automated method for
deriving consistent hierarchies from a table of instances.  FCA is
already widely used for testing and verifying OWL ontologies, but
FCA has software that has been used in mission-critical applications
with hierarchies of thousands of terms.    (011)

As a growth path for the future, I consider CNLs supplemented with
techniques such as FCA and diagrams such as UML as a very fruitful
way to go.  UML is already well integrated with software and DB
design methodologies.  CNLs are an easy to learn extension.    (012)

OWL is not easy to learn, and the jump from the tiny subset that
people actually use to all of OWL2 is much, much harder than
learning and using UML + CNLs.  Furthermore, the combination
of UML + CNLs has much greater expressive power and much better
integration with mainstream IT.    (013)

John    (014)

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)

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