[Top] [All Lists]

Re: [ontolog-forum] Quote for the day -- KR and KM

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Ed Barkmeyer <edbark@xxxxxxxx>
Date: Wed, 05 Jan 2011 16:40:32 -0500
Message-id: <4D24E550.1070605@xxxxxxxx>
John,    (01)

I don't quite know what the topic of your email was, but I offer some 
further complaints.    (02)

John F. Sowa wrote:
> I have a strong sympathy for the work on the conceptual schema
> from the 1970s to the 1980s.  In the 1990s, I participated in the
> ISO work on trying to revive the idea.
> Unfortunately, all the committees that worked on conceptual schema
> standards terminated with technical reports:  ANSI/SPARC in 1978,
> ISO in 1987, and ISO in 1999.
>       (03)

Indeed.  I first met John in the ANSI/SPARC Database Systems Study Group 
in the mid-1980s, which produced one of those reports.    (04)

> I would also add that the SUO group (in the early 2000s) failed to
> produce a standard for ontology, despite being chartered by IEEE.
>       (05)

And /because/ there were competing upper ontologies that were not 
logically equivalent.  The IEEE committee just proved to be a 
battleground for academics with little or no commercial interest in the 
outcome.  (Adam Pease still has the battle scars.)    (06)

> EB:
>> My experience is that KR technologies are still largely impractical for
>> business applications, but precisely because, apart from the work of
>> people like Martin Hepp, there is very little formalized knowledge that
>> is available off-the-shelf.  As a consequence, it takes a large
>> investment to get anything useful out.
> I agree.  This is also part of the reason why the standards projects
> for a conceptual schema or ontology failed.
>       (07)

That was not my experience.  A certain major vendor failed to support 
the standards candidacy of a conceptual schema language created by one 
its elves in the woods, even though it had some commercial following.  
Another vendor actively blocked a standards effort that threatened to 
standardize NIAM, because it had the only commercial tooling for the 
language at the time.  As late as 2004, the responsible authors of a 
widely used information modeling language (other than UML) refused to 
allow the surface language to be standardized, opting only for a 
conceptual metamodel.  In short, standards projects for conceptual 
schema languages failed because the authors of the candidate languages 
wanted to retain ownership of them.     (08)

BTW, I don't want that to sound pejorative.  It was, after all, their 
intellectual property.    (09)

> EB:
>> Unfortunately, the semantics of conceptual schema languages and
>> the semantics of FOL languages are significantly different in
>> ways that make it hard to construct an FOL ontology from a
>> conceptual schema, and impossible (I think) to construct
>> an arbitrary FOL ontology in a conceptual schema language.
> No.      (010)

John, are you disagreeing with the assertion that the translation is 
difficult/impossible?  Or that it is unfortunate?    (011)

What follows in John's email is, from my point of view, a non sequitur:    (012)

> That misses the point by almost 180 degrees.
> The two fundamental problems are (a) knowledge acquisition and
> (b) the integration of that knowledge with the software tools.
>       (013)

Yes.  We discussed the first, and you are right about the second.     (014)

But Sjir's email was asking what ontologies can do that conceptual 
schemas, a la TR9007, cannot.  My point was that the definition of 
conceptual schema in TR9007 doesn't address the semantic interpretation 
or the expressiveness of the language involved.  And the definition of 
the content of the schema suggests that it needs an extremely powerful 
and expressive language.    (015)

To be clearer, you couldn't say in SDM (Hammer/MacLeod) anything more 
than you can say in modern extended DL languages.  You could say much 
more in SSADM, but the semantic model of "dynamic aspects" in SSADM was 
really:  Create, Delete, Modify.  IDEF1-X was a thin semantic coating on 
the Relational model.  And NIAM/ORM has approximately the same power as 
OWL/DL, but it adds n-ary relations, role concepts, explicit reification 
of relations, and a stronger model of unique identification, which 
greatly affects its apparent expressiveness.  (That is why Sjir asks 
what ontologies can do better.)  The main difference, however, is that 
these conceptual schema languages have a semantic model based on a 
closed world assumption -- that the information base is a 'model', not 
just the rest of the ontology -- and in most cases they have some 
intrinsic notions that are non-monotonic.  That makes them different in 
kind from FOL models per se, and requires you to build a FOL vocabulary 
to describe the fundamentals of their semantics.    (016)

That was my point.  I was just suggesting to Sjir that he is asking us 
to compare apples to a fruit whose delimiting characteristics he didn't 
define.  And from what I know of that fruit (Sudha Ram and I surveyed 
that literature in 1989, with the objective of identifying all the 
distinct modeling concepts the conceptual schema languages contained), I 
don't think I would put ontology apples in the same salad.  Both kinds 
of languages capture knowledge, but their purposes and their 
expressiveness are different.    (017)

> The ANSI/SPARC three-schema idea was good, but they didn't
> address knowledge acquisition and integration.
>       (018)

That is because it had nothing to do with knowledge capture and very 
little to do with conceptualization.  It was a technical architecture 
for database specification.  All the rest is window dressing.    (019)

> The reason why knowledge acquisition is so important is that
> the middle and lower levels of the ontology grow exponentially
> faster than the upper level.  No matter how big the upper level
> may be, it can only contribute a small part of the total that
> is needed for any significant application.
>       (020)

I don't know whether we agree here or not.  If the application ontology 
is the "lower level", then it can be pretty small.  Most of those I have 
had to build have had about 150 classes and 300 properties.  The problem 
is that 120 of the 150 classes are primitive.  As a consequence, you 
can't do much reasoning about individuals.  To make the ontology more 
useful, you need to produce definitions of most of the classes and 
properties, and that requires a much richer and clearly founded middle 
layer.  I see my ontology as a hut with multiple levels of sub-basement.    (021)

-Ed    (022)

Edward J. Barkmeyer                        Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263                Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263                Cel: +1 240-672-5800    (023)

"The opinions expressed above do not reflect consensus of NIST, 
 and have not been reviewed by any Government authority."    (024)

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    (025)

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