An approach for implementing TUNES
RE01 Rice Brian T. EM2
Mon, 12 Apr 1999 12:40:31 +1000
> Yes, we would. But I am still wondering if we can escape writing any
> "dirty" code at all, by finding or writing a logic system to find a TUNES,
> giving the TUNES definition, and the definition of our target.
> Thanks again for your comments...
> - Brent Kerby ("Iepos")
i've been silent on this issue, surprisingly. i would just like to add that
a helpful answer should be forthcoming shortly when i release my arrow paper
to the group. delays keep coming, i know, but i've also made some real
conceptual headway in trying to formalize the tunes ideas.
i'd just like to interject that i am working on an abstraction system based
on category theory. the idea is that we form a "category of contexts": a
collection of nodes in a web that represent the kinds of representations of
information that the system understands, and various arrows leading between
the nodes representing processes that interpret information that exists in
one context from the perspective of the other.
this notion generalizes the idea of abstraction, so that no longer need
"levels" exist: the hardware is no longer a base context, and no confusion
results from multiple types of abstractions over it. the new abstraction
types are merely new nodes in the web.
also, i am looking at the idea of dynamic models of inference, which allow
inference styles that range from pure boolean logic to the logic of
procedures and functions (modalities) in a clean manner.
much more will be explained in the paper soon.