SWRL- political motivated

At 22:31 08.02.2006 +0100, Dieter Fensel wrote:

>At 17:46 08.02.2006 +0000, Ian Horrocks wrote:
>
>
>>Can you explain what you mean by "political motivated" in this context?
>>
>>Ian


First, FOL is a very cumbersome tool. It does neither provide proper 
modeling support
nor does it even cover the expressivity of simple datalog (i.e., transitive 
closure). Therefore,
every rational person would go for something like second-order logic to 
have a more
suitable formalism for modeling without artificial compromises.

Still, second-order logic is not widely used. Why? Because even unification 
is undecidable
in SOL. In conclusion, research on computational traceable logical 
sublanguages has
produced sublanguages of it:

         - FOL, where unification is decidable
         - DL-type sublanguages of FOL where various reasoning tasks become 
decidable and
         implementable.
         - Rule-type languages (that look syntactically as sub-languages of 
FOL without really
         being sublanguages, since they also extend the expressive power of 
FOL).

The future may bring further interesting subfragments, however, this is 
clearly beyond the
scope of our charter. In any case, there is a meaningful pattern. 
Restricting the expressive
power of the logical language in order to gain effectiveness and efficiency 
in computational
processing of these language. This is good and valuable science and your 
FACT reasoner
proved empirically that OWL-Lite can be handled computationaly.

With SWRL it is different. It is a syntactical restriction of FOL without 
any theoretical or
empirical justification that we would gain anything in terms of 
computational complexity.
Therefore, this restriction of FOL is not justified by science but only by 
politics or taste.


----------------------------------------------------------------
Dieter Fensel, http://www.deri.org/
Tel.: +43-512-5076485/8
Skype: dieterfensel

Received on Friday, 10 February 2006 17:52:29 UTC