[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index]
[Thread Index]
[Author Index]
Re: BioSystems - paper somewhat related to (M,R)-systems
- From: Tim Gwinn <***>
- Date: Fri, 13 May 2005 10:54:21 -0400
Hi
Ayten,
Sorry for the
delay. I agree with your remarks. It is interesting that the only reference to
Rosen is to a 1971 paper ("Some realizations of (M,R)-systems and their
interpretation", BMB 1971, 33:303-319). In that paper, Rosen discusses some
possible directions in which to attempt to move from the abstract (M,R)-system
relational model to something closer to a structural representation, which then
might be realized. Included in these directions are the formalism of state-based
finite automata and dynamical systems representations.
I suspect that
these discussions occurred prior to Rosen reaching his later
conclusions about complex models: noncomputability, analytic models not all
reducible to synthetic (structural) models, and the dilemmas this creates for
realization. In 1985 in Theoretical Biology and
Complexity, he explicitly discusses how his earlier attempts (1964
and referenced in the 1971 paper) to realize (M,R)-systems using finite automata
were unsuccessful, and uses that as a lesson to show why such
an approach was not desirable or possible for realizing complex systems.
Certainly the
results in the later Life Itself and
Essays on Life Itself sharply point out those
limitations. So it seems rather odd to me that the CJW BioSystems paper
does not take Rosen's later publications and results into
account. This doesn't mean anything negative about hybrid automata per
se, its simply a question of why they chose to use the (M,R)-system model, which
is refractory to state-based approached (whether deterministic or not), as the
launching point of their paper.
Regards,
Tim
Dear Tim,
I read the article
you mentioned. I am of the same opinion that Rosennean theory is not applied
to their work in terms modeling relations, as far as I understand,. There does
not seem to follow causality-implication, thus decoding-encoding between the
system and its model which is the main concept in the RR theory, I guess. In
effect, their work does not go beyond the hybrid automata model which may not
have required the fuller use of Rosen's M-R even it is there claimed that
highly nonlinear dynamics and nondeterminism can be captured in a formal
setting. This setting may miss the requirements of the RR's M-R meant for the
highly complex living systems. These are my impressions as non-expert (yet) on
Rosennean Bio-theory, which I believe hide much more truth than revealed in
this article. Perhaps it does not open a right door either. What do you
think?
Regards,
Ayten