W3C home > Mailing lists > Public > semantic-web@w3.org > January 2006

Re: [semanticweb] how to explain to humans the term ontology

From: Pdm <editor@content-wire.com>
Date: Thu, 12 Jan 2006 16:40:58 +0000
Message-ID: <43C6869A.4010903@content-wire.com>
CC: semanticweb@yahoogroups.com, semantic-web@w3.org


In computer science,  'data' offers  a reductive representation of 'the 
world' - ie whatever we can manage to shrink into a database

I look at my pile of shoes on the one hand, and my table that contains 
the data relating to the pile of shoes on the other hand, and I
see two very different things.  The reality is the pile of shoes, the 
table is a representation of the data. A photo, would be a visual 
representaiton.

Data, as understood by a computer system does  not represent reality in 
its entirety and complexity, but a part of it.

What aspects of reality are represented by data, depends by the purpose 
that the data must serve.

Similarly, when modelling  an ontology, we aim to represent reality  as 
much as possible, and then often only develop those aspects that are 
particularly
useful to the purpose of our system development (every reality can be 
dissected into subatomic particles, may its not necessary on an everyday 
basis)

On this list, we have seen a heuristic search  discussion come up ' 
search by humming' . And someone asked: what do we model, the intervals 
between different notes, or the cultural context they belong to?  He was 
asking, what aspect of reality are we going to model? (reply will 
follow, short answer is: depends on the purpose of the system)

Interestingly, this means that our ability to represent data has become 
much more versatile, now we can store and search data according to a 
variety of parameters/

In the protege list, someone started  a 'motorbike ontology'.  I am 
particularly interested in modelling the relationships of the motorbike 
compontents to each other, because  I am thinking of a motorbike 
ontology for the purpose of building a 'self assembling' motorbike, 
where the components contain a chip that
knows where the component goes and how should be fitted with the others.

My ontology may model an aspect of reality that may be totally 
irrelevant to a supplier who is simply building an online catalog for 
his spare parts

But I would be interested to know if people thinks that there should be 
only one ontology. That would pretty much mean that there should be only
one view of the world?  I dont think so.


The bottom line is that understanding reality - ie getting the facts 
that are important to your system goals right -  is key to developing 
intelligent systems
Ontologies are tools used to support that process

Wheter computer science  has anything to do with intelligent systems,  
or with reality for that matter, depends very much what aspect of the 
computer science
we consider.

In the last ten years ' a layer of intelligence has been added even to 
the dumbest bits of Computers,  ie data and networks, whereby ' 
intelligence' I refer
to the ability of a system to understand and respond to an envrionment,  
Even data and networks are becoming increasingly capable of  context 
sensitive behaviour

Not everybody's mind is sufficiently open to understand the  breadth and 
potential of new paradigms.


Paola Di Maio

>>
>>
>>
>>
>>
>
>
>
Received on Thursday, 12 January 2006 09:41:44 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:47:11 UTC