Thank you Mark. This promises to be interesting.
My view may best be introduced by stating that I believe we are in the business
of creating a new science that will depend on new abstractions. These
abstractions will extend from the notion of "information" as a first class
citizen, as opposed to our default, the "particle." The latter has qualities
that can be measured and in fact the very idea of metrics is bound to this
notion of thingness.
Because we will not leave existing theoretical tools behind, we need a bridge
between the abstractions of "effect" in the particle model (fields and forces)
and the corresponding "effect" in the information model. I am fine with
extending the metaphor far enough to say that we need something like
parametrics in our new science of information. But I really balk at using the
notion from one system in another without some sort of morphism.
Much of the dialog here works with the problem of naming what that it is.
Unfortunately, the abstractions of fields and forces are a very poor formal
model, because they are defined not by their essence but by their metrics.
Having said that...
> 1.Is it necessary/useful/reasonable to make a strict
> distinction between information as a phenomenon and information measures as
> quantitative or qualitative characteristics of information?
I am rather certain that there is a very real distinction, because of how we
define the problem. After all, we are not asking how do information and
information metrics fit within the confines of rather limited abstractions. At
least I am not. But the distinction does not allow for full orthogonality from
set theory (the formalism of things), because we want to be able to model and
engineer observable phenomenon in a cleaner way. This should be the test of any
serious proposal, in my view.
This requirement is why discussion on these matters often moves into category
theory, after the fashion of Barwise and others. A spanning morphism can extend
the notion of parameters to information space, but only when considered in the
situation of that origin (meaning measurable space in the traditional sense).
> 2.Are there types or kinds of information that are not
> encompassed by the general theory of information (GTI)?
I believe so. Some types clearly have laws that affect the world, which is how
you scope the types covered by GTI. But just as particle physics finds it handy
to have virtual particles and transcendent symmetries over them, so will we
have information types that do not touch the world in an observable way; these
will be required to support clean laws of behavior, yet to be convincingly
proposed.
> 3.Is it necessary/useful/reasonable to make a distinction
> between information and an information carrier?
I suppose you will get universal agreement on this, at least here. But...
I was just at NIH at a rather introspective conference on structural biology,
which assumes that the form of the carriers collectively forms the code of the
system. They have dropped billions (quite literally) into metrics associated
with these laws of information form but are ready to abandon the concept as a
key technique. Clearly there is a system-level conveyance of information that
"carries" an organizational imperative. If these can be said to be supported
with the metaphoric virtual particle with the local interaction governed by the
form of the carrier, then the answer is both yes and no.
I am intrigued by the notion introduced here recently that suggests
"intelligence" as inhabiting this new, non-parametrizable space.
--Ted
_
Ted Goranson
tedgoran...@mac.com
http://www.sirius-beta.com
_
Ted Goranson
tedgoran...@mac.com
http://www.sirius-beta.com
___
fis mailing list
fis@listas.unizar.es
https://webmail.unizar.es/cgi-bin/mailman/listinfo/fis