Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: CG: Re: SUO: CYC event vs. SUMO Process -- really different?




Len,

It's the mathematician, physicist, logician, and
philosopher, Alfred North Whitehead.

>Just for those of us (me) who aren't as well read 
>in this domain, which Whitehead is being referred 
>to here?

For my summary of some of the issues about Whitehead
and others, including Peirce and Wittgenstein, see the
paper on Signs, Processes, and Language Games:

   http://www.jfsowa.com/pubs/signproc.htm

>This may be way off, but in a different domain, real 
>time 3D modeling, as expressed in languages such 
>as X3D/VRML, time is a distinct object that controls 
>event propagation via ROUTE objects, and objects while grouped, 
>also expose in/out interfaces which can only accept 
>and emit named event types.  For a tornado to 
>emerge and be considered such, one must isolate 
>the first occurrence of the type of event which 
>gives the group of objects observable tornado 
>characteristics.  One might assert that this 
>event or combinations once routed via the first 
>discrete value of the time sensor, creates a 
>state that once attained, the group cannot 
>become anything either than a tornado.  It 
>is that state that confers tornado identity 
>(as opposed to the kind of identity one 
>gets using ID types).  I note that these are 
>simulation system conventions, not reality, 
>but I am three quarters convinced that all 
>ontologies are simulation systems as well.

Any science, whether ontology or nuclear physics, can be
used as the formalism implemented in a computer simulation.
So I would agree that the ontology you use would influence
how you simulate reality in exactly the same way that the
version of physics would.

>Identity is a murky topic in computer science. 
>XML is bedeviled by it now that well-formedness 
>is the critical criteria for XML goodness instead 
>of DTD or schema-based validation.  Current wisdom 
>is that IDness has to be defined as a type from the 
>infoset.  That escapes the time argument in much 
>the same way simulation does: it creates a property 
>type but defers for later discussion what event 
>confers or assigns it.   The argument about identity 
>being an assigned property (I think it is) vs an 
>inherent property (the dominant view if one takes 
>URIness seriously) goes on.

Identity is a very murky topic in ontology as well.

John