xml-dev - RE: local, global (was various ontology, RDF, topic maps)

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

If you inherit an ontological service, you may 
be right.  If you delegate it, you get to reuse 
the delegate for any method that takes an 
ontology of that type.  Delegates are mortar 
that depend on what brick they are bound to 
as to the strength of the binding.  Choose wisely.

A web that can afford ontologies is 
recouping costs of services.  Focus on 
the service and the domain of the ontology 
falls out directly.  The content of the 
ontology is a harder problem but one which 
its users have to solve and will if it makes 
a hard job easier. 

Anyone who thinks we will turn on machine 
processable negotiations for critical business 
then not test them and use the results to control them  
is smoking dope.  The golem serves; it does not rule.

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Simon St.Laurent [mailto:simonstl@simonstl.com]

Ontology as mortar fits my argument quite nicely - it's convenient glue for
sticking heavy things together, but it's damn near impossible to reuse
after it has set, has almost no flexibility, and it isn't very good about
changing position, either.

I'd suggest we stay away from masonry altogether in this business.  I find
your metaphor more convincing than your argument, at least as far as
ontology is concerned.