Re: SUO: RE: Ontological Tagging
Matthew,
While "dpANS" KIF doesn't have this, the most common use of KIF in
Ontolingua does have a "package" definition statement which adds this
functionality.
This sort of feature would be helpful although in effect all it is doing
is just implementing a shorthand for prefixing every constant within the
package context with the name given to the package. I think this is quite
distinct from a need to support ambiguous names for linguistic
processing. While the ontology can have many of the features of a lexicon,
and in fact support definition and use of a lexicon, it is something
different. I would recommend that that SUO have unique names for every
thing it defines and that other efforts may map the SUO terms to elements
in a lexicon such as WordNet.
We should also distinguish this from issues of equality, intention vs.
extension and SUO content vs. instance descriptions in discourse. In the
SUO we need to be able to refer to objects which have different intention
but the same extension. One classic example of this is the morning
star/evening star distinction (apologies if you're already aware of all
this, just trying to be clear for everyone). The morning star might be
defined as the brightest start early in the morning and the evening star as
the brightest at dusk. They have a different definition even though they
are in fact the same object. However, because they have distinct
definitions and they are not simply different symbols for the same thing, I
believe it would be valid to have them both in a knowledge base. We also
need to keep this issue distinct from issues about how to provide an
ontology that supports semiotics.
In summary, I think there are four distinct issues
1. Creation of a natural language lexicon mapped to the SUO (out of scope
for the SUO effort)
2. Support for equality and equivalence of terms in the SUO
3. Creation of an ontology of semiotics as part of the SUO
4. Namespace support in the language chosen for the initial expression of
the SUO
Adam
At 07:31 PM 8/15/2000 +0200, West, Matthew MR SSI-GPEA-UK wrote:
>Dear Colleagues,
>
>I would like to pick out one thing here in particular.
>
> > Physical tags come in different shapes and colours, so an
> > ontological tag has
> > two major attributes, the namespace representing the shape
> > and colour and the
> > reference representing the written data on the physical tag.
>
>We have implemented namespaces in a way that is something like this in
>EXIST. I note that there seems to be no such facility in KIF (please correct
>me if I am wrong).
>
>The practical value of namespaces is that they allow terms to have more than
>one meaning (it was already possible for different terms to have the same
>meaning using equality). Since terms usually do have more than one meaning,
>there is an immediate application in linguistics. However, it is also
>valuable when merging or mapping existing ontologies. You cannot require
>that the two ontologies shall not have used the same name for anything
>before you start. Namespaces provide a convenient way of differentiating the
>different uses.
>
>Regards
> Matthew
>============================================
>Matthew West
>Asset Information Management
>Shell Services International
>H3229, Shell Centre, London, SE1 7NA, UK.
>Tel: +44 207 934 4490 Fax: 7929
>E-mail: Matthew.R.West@is.shell.com
>http://www.shellservices.com/
>============================================
>
> > -----Original Message-----
> > From: David Mohring [mailto:heretic@ihug.co.nz]
> > Sent: 10 August 2000 18:41
> > To: standard-upper-ontology@ieee.org
> > Subject: Ontological Tagging
> >
> >
> >
> > Heresy part 1 : Ontological Tagging
> >
> > One of the things Ford Prefect had always found hardest to
> > understand about human beings was their habit of continually
> > stating and repeating the obvious, as in It's a nice day, or
> > You're very tall, or Oh dear you seem to have fallen down a
> > thirty-foot well, are you alright? At first Ford had formed a
> > theory to account for this strange behaviour. If human beings
> > don't keep exercising their lips, he thought, their mouths
> > probably seize up. After a few months' consideration and
> > observation he abandoned this theory in favour of a new one. If
> > they don't keep on exercising their lips, he thought, their
> > brains start working. After a while he abandoned this one as well
> > as being obstructively cynical and decided he quite liked human
> > beings after all, but he always remained desperately worried
> > about the terrible number of things they didn't know about.
> >
> > Quote from The Hitch Hiker's Guide to the Galaxy
> >
> > With this in mind here are some blatantly obvious observations.
> >
> > (And maybe in a later post some solutions to large scale multi-form
> > knowledge representation.)
> >
> > The Ontological Tag
> > - Or I've just come from Denver, where is my luggage?
> >
> > To applications symbols and words in an ontology are just
> > tags which perform
> > the same function as a physical tag does on a suitcase at an airport.
> >
> > Just as a piece of luggage may have many physical tags
> > attached upon it's
> > handle so may a piece of data have many ontological tags
> > referenced to it.
> >
> > Physical tags come in different shapes and colours, so an
> > ontological tag has
> > two major attributes, the namespace representing the shape
> > and colour and the
> > reference representing the written data on the physical tag.
> >
> > Just as a physical tag of one shape and colour may be of
> > interest just to one
> > process of transferring luggage from one destination to
> > another, the namespace
> > of a tag may be of interest to one set of applications.
> >
> > It greatly helps the sorting process if physical tags from
> > any two vendors
> > that are not meant to be used in the same process do not
> > share the exact same
> > shape and colour, so it also helps if the namespace of an
> > ontological tag is
> > also unique to each process.
> >
> > You can use a URI as an ontological tag using the standard format of
> > "http://vendoraddress/namespace#reference". To an application
> > what the URI
> > references is unimportant, only that it has an unique address for each
> > "destination"
> >
> > IMHO the role of SUO should stop at this point, generate
> > basic tag ontologies
> > without having to try and resolve this next issue.
> >
> > When missing the appropriate physical tag of the correct
> > shape and colour it
> > may be possible to "guess" the appropriate destination from
> > another "known"
> > type of tag. Unless following very consistent and unambiguous rules
> > this will probably result in the bag going to incorrect destination.
> > ( Well it always seems to happen to me - dammit )
> >
> > If the SUO attempt to provide an automated standard solution
> > to the above
> > problem, then its members will probably be at best morally
> > (and at worst
> > legally ) responsible for every "correct" implementation
> > based on that
> > standard.
> >
> > Imagine going down in history as the "Denver baggage handling system"
> > designers of all the new ontological based applications.
> >
> > David Mohring - Continue [Y]es [N]o ?
> >
-----------------
Adam Pease
Teknowledge
(650) 424-0500 x571