RE01 Rice Brian T. EM2
Thu, 10 Dec 1998 15:47:17 +0800
i apologize for starting the argument. it stops on my side. i refuse to
rebut. i also admit to losing.
now i want to talk about the ideas behind identity and reference in this
we have atoms which can be referenced: Arrows.
all atoms in this system contain two (ordered) references.
the functional paradigm forces us to think of one of the references
as THE reference for the atom. it also encourages the use of the inversion
operator in creating an arrow 'pointing' in the opposite direction.
what i'm trying to work out is that arrows do not explicitly contain lists
of arrows which point to them, which becomes very significant in resolving
the shape of graphs. i still do not want to constrain the arrow definition
to a low-level format, but i want to define some context for an 'arrow
processor' which walks over arrows and develops an internal state to keep
track of properties of the system, particularly 'shape'-related ideas. in
the system as we have developed it, we guarantee the specifiability of
arrows ( or nodes) by the 'pointer' ID-number or some equivalent. we're
also intending to move this system into a context where it can attain an
infinitary population. one of the ideas, perhaps the best so far, is to
specify arrows instead by axioms. by formal specifications.
i'll develop this in my next post...