Re: what exactly are the encoding and decoding arms?
DON MIKULECKY (MIKULECKY%VCUVAX.BITNET@letterbox.rl.ac.uk)
Thu, 28 Sep 1995 13:40:58 -0400
Don Mikulecky, MCV/VCU, Mikulecky@gems.vcu.edu
an afterthought on use of the modeling relation:
before Rosen's modeling relation, we used to think about two distinct
kinds of models (also a very controversial issue) They were:
1) models of data......in simplest form, one or another kind of curve
fitting
2) models of systems.....more often than not, these were analog models
between formal systems, but also between models od data and formal
systems. This latter case is probably close to what Bruce was implying by
leaving the natural system out, but I'm not sure. Relational models can
allow us to use the modeling relation in a much more global way as in
the M-R (metabolism-repair) system in Life Itself and elsewhere. Then,
the induividual "model of data" and "model of system" get put into sets with
many other members and the modeling relation involves mappings on these sets as
well as on those mappings themselves. At this point, relational models invol
ve
measurement as an encoding, but the the encoding arm of a relational model
is no longer simply measurement. Back at the level of an individual
model of data the single instance modeling relation is indeed a mapping
from the natural system to a set of measurements, an encoding we can call
"measurement". This specific mapping is then "lost" as a detail in
the subsequent, more abstract, relational models into which it becomes
incorporated.
I hope this helps.
Don Mikulecky