W3C home > Mailing lists > Public > semantic-web@w3.org > December 2005

Re: How will the semantic web emerge

From: Henry Story <henry.story@bblfish.net>
Date: Fri, 16 Dec 2005 20:11:27 +0100
Message-Id: <E4E22B8C-8541-465C-AA03-913ECAAB74B6@bblfish.net>
Cc: Asankhaya Sharma <asankhaya@yahoo.com>, semantic-web@w3.org
To: love26@gorge.net

Well just as the phrase you cite is outrageously racist and close  
minded, so is your idea about what constitutes a good ontology.

I would say that a good ontology has to do with a few things:
	- how easy they are to learn
  	- how well they map the domain they are trying to model
	- how well they mesh with other ontologies that are well known and  
used (makes them easier to learn, and requires less work to develop  
them)
	- how likely they are to be widely accepted

In the end if you publish valuable data you can create pretty much  
any ontology you wish to. People are going to be interested in the  
data you publish, not in your ontology. If you write a good ontology  
you will find that other services could start using the ontology you  
have developed, or ask you to help standardise it. It is at this  
point that more issues will come up that you may have initially  
understood.

Again the point is that if *you* publish *valuable* data, do it at a  
SPARQL end point. This is most likely going to be the way the largest  
amount of quality data is going to get out on the web in Semantic  
form. It does not require you to agree with anyone else about your  
ontologies to start off with. With time getting the experience of  
others will of course proove very valuable.

Henry Story

On 16 Dec 2005, at 16:51, William Loughborough wrote:

> Asankhaya Sharma wrote:
>> HI,
>> I am sorry but i dont understand the irony of "good"
>> and "dead"..
>> can you throw some light on it...
>
> There is an old pejorative phrase from Western Movies: "The only  
> good Indian is a dead Indian."
>
> One problem with ontologies is that they foster hierarchical  
> attitudes towards how things get classified. To many of us, they  
> ALWAYS have "cracks" in them through which fall the "tags" we find  
> more suitable as index/annotation bases.
>
Received on Friday, 16 December 2005 19:11:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 February 2013 14:24:52 GMT