W3C home > Mailing lists > Public > public-owl-dev@w3.org > July to September 2007

Re: declaredAs

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Sun, 5 Aug 2007 02:34:25 -0400
Message-Id: <967555CA-E2F3-46AF-8477-28D28BD2C598@gmail.com>
Cc: <public-owl-dev@w3.org>
To: Boris Motik <bmotik@cs.man.ac.uk>
On Jul 30, 2007, at 1:55 PM, Boris Motik wrote:
> Our main reason for introducing declarations was to be able to  
> detect typos.

Along the lines of "do we really need this", I don't really ever  
recall an outcry for catching typos. Many of the features of OWL 1.1  
are motivated by user needs, and this feature, prone to  
misunderstanding and maintainability issues, not to mention this  
business of dropping pieces of OWL 1.0 on the floor, just doesn't  
seem to be of the same stuff as much of the rest.

Seems to me that a parser could perfectly well become aware of the  
typo you present by a single pass read of all the ontologies in the  
import closure, before the second pass emit any *warnings* it thinks  
worth of mention, and in a second pass have all the typing  
information it needs. With disk caching, which amounts to memory  
caching on most machines these days, this doesn't seem, to me at  
least, to be too costly.

I don't really understand the problem with cyclic imports, BTW.

Also, like Michael, it's hard for me to have sympathy for the OWL  
parser writers (even though I know a few of them and like them a  
lot :) As hard as it might be, it's not like that technology isn't  
well understood.  Nothing compared to writing a disjunctive datalog  
engine - something that only a few people in the world can manage ;-)

Received on Sunday, 5 August 2007 06:34:41 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:58:15 UTC