xml-dev - A Light Rant On Ontological Commitment

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

It will be well to investigate the extent to which 
the so-called ontological commitment covers the 
vocabulary declared and the system by which such 
declarations are made.  Tim Berners-Lee has 
heavily stressed the machine-readable aspects 
of semantics.  A commitment to an ontology is 
also a commitment to the system.  You can ask 
yourselves if a commitment made to the semantic 
web is a commitment of the ontologies  
to the internet/www, in other words, 
to the machine lifecycle (the XML position) or 
if the commitment covers broader aspects of the 
information environment itself over the lifecycle 
of the information (the SGML position).

As Martin Bryan points out, the issue of human 
readability is fundamental to ontology. Without it, 
we can quickly lose independence of action to the 
originating authority or the ontology can quickly 
lose value by hard coupling to the system.  

The definition of "system" is key. 

Go slowly.  This is not one you want to "feel" your 
way into as was done with HTML/WWW.  Pay particular 
attention to the concept of observable behavior 
based on the knowledge level and the problem that 
the principle of rationality only weakly governs 
human behavior, thus, may be weak over machine 
behavior as well.
  
The human is slow.  While reaction time is a resource, 
the human is not distributed and that restricts 
affective scope.  The machine is very fast and 
amplifies both knowledge and superstition.  That 
is a danger.  The means for establishing a  
record of authority is key.  The machine is 
only a means of discovering and distributing that 
record.

There are ancient stories about a Generalized OntoLogical 
Emergence Machine (GOLEM) that some might want to keep 
in mind.  It may be a pernicious comparison, but 
it may also be informative.  :-)

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

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