ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Ontology similarity and accurate communication

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Fri, 14 Mar 2008 11:02:11 -0500
Message-id: <47DAA183.10604@xxxxxxxxxxx>
Dear Matthew,    (01)

JFS>> The airlines have been storing such information in their
 >> databases for over 40 years without having a detailed ontology
 >> about the nature of places, times, things, and events.    (02)

MW> Oh, of course, they have been holding the data for decades,
 > and fixing up/misusing data structures to do what they need.
 > But that is not the same as saying they have been using a 4-D
 > data model to do it, or as you say something with any explicit
 > ontological foundation. And since the data is adequate it is
 > possible to map that to a 4-D model.    (03)

I agree that they have not used any explicit ontology and that
much of it is very ad hoc.   Yet after sufficient debugging,
they get the correct answers, and those answers are compatible
with a knowledge base that uses either a 4D or a 3D ontology    (04)

MW> But that is quite different from it *being* a 4D model just
 > because it holds the necessary data.    (05)

I certainly agree.    (06)

MW> The problem is knowing which data of this sort is actually
 > independent of the upper ontology (i.e. it morphs when moved
 > from one to the other).    (07)

If you start with an upper ontology, it might not be clear
what lower-level aspects are independent of the top.  But
any correct low-level description of any aspect of reality
should be compatible with any correct upper-level ontology.
(But there may need to be some conventions for mapping one
description to the other.)    (08)

That implies that any DB that maintains correct data, even
if no ontology had been used, should be compatible with any
correct uppper ontology.  An ad hoc system that agrees with
reality on all observable data might be difficult to design,
maintain, etc.  But the crucial test is whether all observable
data is correct.    (09)

MW> The problem here is the spaghetti problem of having to
 > deliver pairwise interfaces.    (010)

That is indeed a problem.  But it is not clear which problem
is harder to solve:  align the upper ontologies of all the
systems involved, or check whether their local predictions are
correct.  People have been solving the latter problem in practice
for the past 50 years, but the former is still a research issue.    (011)

MW> And you won't have to go far to find someone to tell you that
 > maintaining interfaces has a high cost, because they change when
 > the system at either end changes.    (012)

I agree that there are many issues at all levels.    (013)

MW> The evidence is not about theory, but about economics.    (014)

That is indeed a crucial concern.    (015)

John    (016)


_________________________________________________________________
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    (017)

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