Re: How semantic is semantic web?

Hi everybody,

first of all I'd like to thank you for the many interesting replies to  
my proposal and for the many challenging references whose reading kept  
me quite busy in the last days!

The discussion referenced by Azamat provides a deep explantion of all  
the problems involved in defining what ontologies are about, but I  
must disagree with his main thesis, namely that: " 'ontologies do  
really deal with semantics, as they give reference to things in real  
world: they deal with reality and realities. As we build ontology, we  
are explaining  the structure of the world and its domains, we are  
giving real meanings (i.e. a reference  to things) to the terms and  
concepts we use; we are explaining the relationships between  
entities.' "

 From my point of view - which may be considered, if you like it, next  
to pragmatism, as Leonid suggests - the creation of an upper ontology  
can be a challenging enterprise, but it's not definetely what we need.  
As we write ontologies, in fact, we never speak about things, we  
usually speak about data, about informations, about the knowledge we  
have of things, so ontologies should not reflect the world itself, but  
our knowledge of it. And the knowledge we have of the world is always  
prone to some purpose, to some partial point af view.
Besides, we write ontologies 'just' to improve the ability of our  
machines to give us meaningful responses, we're not going to use them  
to find 'the answer to life universe and everything': we need  
semantics for practical purposes, and I think that practical purposes  
are best managed with many partial (and maybe also contradicting)  
ontologies.

I think we could say that the web is getting more semantic if we  
follow Rich's view - if I understand it correctly - and mean by  
'semantic' something like 'full of meaningful signs': this way the web  
is more semantic, as we are using more meaningful and expressive signs  
to build it, and to express metadata in some machine-processable way.  
I wouldn't say that we are "in an environment where meaning is nothing  
more than data, or metadata expressed in such a way that a computer  
program can understand it", as Alejandro says, as meaning is  
definetely the reference of a symbol, or sign, to something else, and  
computer's world is entirely made of symbols, so there is nothing to  
whom symbols could refer, i.e. there can be no meaning, without some  
human reading data. But it's true that we can now let the machines  
process data in such a way that we can get results more meaningful to  
us.

Unfortunately, 'semantic' is usually intended as 'related to meaning',  
arising the misunderstandings we know: my intention in proposing  
'epistematics' as a new term was not to have a simpler term, but one  
not already overburdened with dense meanings, as 'semantic' is.  
Moreover semantic technologies are useful, as we all know, in many  
non-web fields and, in my opinion, binding them to the web is somehow  
restrictive.

Lastly, I really liked Andreas' interoperability-based broad vision,  
and the parallel there made between  OWL and Esperanto as common  
languages studied for interoperability let me think about the work of  
a colleague of mine who actually wrote an ontology of Esperanto. Maybe  
you'd like to take a look at it: at the page  
http://www.epistematica.org/pubblicazioni/Esperanto.html you'll find  
links to the ontologies as well as to the related paper.

Cheers and thanks,

Chiara


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

Received on Wednesday, 28 February 2007 08:34:21 UTC