SUO: InferencePSC and theory lattice
John Sowa said:
> I plan to write a tutorial about the
> lattice and how to go about using it and implementing it.
John,
I'd hope that you will include in your tutorial issues like
the #$InferencePSC (Inference Problem Solving Context).
As I understand it, this is just a name for a set of paths through
the theory lattice, that have certain interesting properties.
The OpenCyc browser states:
Everything true in Base KB knowledge is also true in Inference PSC
knowledge. #$InferencePSC has this comment: A problem solving context
in which all assertions in the entire KB are initially visible.
However, down each inference path, the chain of microtheories used
is required to have some common descendent microtheory which is
capable of inheriting all of the microtheories in the chain.
Thus, queries made in #$InferencePSC in effect compute locations
in the space of microtheories in which the query is satisfiable.
This context is used internally by the inference engine for
certain recursive calls to itself, including forward inference.
See also #$EverythingPSC.
So these paths through the theory lattice are those in which the
query is consistent, I think. But your description to date has
been focused on a single path or combination of theories.
David