X-Message-Number: 4640
From: Ralph Merkle <>
Subject: Response to Donaldson's post
Date: Mon, 17 Jul 1995 19:59:05 PDT

Thomas Donaldson recently said (CryoNet Message #4631) that
I have concluded that "...we need not worry about our suspension
technology."

I trust most readers will recognize that this is, shall we say,
an inaccurate representation of my views.  Various other statements
by Thomas about what I said are also significantly biased.

Thomas also said (CryoNet Message #4629): "I have not yet
examined Ralph's posting on WWW. In that sense, I'm not qualified
to say anything; however...."  Further comment on this statement
would appear superfluous :-)

The article to which Thomas is objecting is "Cryonics, cryptography,
and maximum likelihood estimation," which can be found on the web at:
http://merkle.com/merkleDir/cryptoCryo.html

The basic thrust of this article is that certain cryptanalytic methods
used to determine the (initially unknown) wiring in World War II
rotor machines could be applied to the problem of determining
neuronal structure when there is incomplete information (caused,
for example, by a poor suspension) about that structure.

Several years ago, in an article discussing the first publication of
"Molecular repair of the brain" Donaldson said "A robot of molecular
size astounds and amazes.  But exactly what computations is it supposed
to *do*?  Without answers to that question, no one can even decide whether
the approach is practical." (Cryonics, Vol. 11, No. 1, January 1990,
page 31).  More recently, in reference to "The Technical Feasibility of
Cryonics" published in Medical Hypotheses, Donaldson said "Unfortunately
his argument was then and remains utterly fallacious on this central point
of the computer power required.  Furthermore, depending on the kind of
"analysis" we ask of a data base, there is quite literally no amount of
computer time required to do the analysis." (*It's not at all so easy,*
Cryonics 3rd Quarter 1994, page 37).

Naturally, I assumed Thomas would be pleased with more detailed
descriptions of the algorithms that could be used :-)
Alas, this has not proven to be the case :-(

Those interested in my views might wish to read my cryonics page at
http://merkle.com/merkleDir/cryo.html

Those interested in articles by Tadd Hogg and colleagues on phase
transitions in constraint satisfaction search can see
ftp://ftp.parc.xerox.com/pub/dynamics/constraints.html

To quote from the cited URL: "Many studies of constraint satisfaction
problems have demonstrated, both empirically and theoretically, that
easily computed structural parameters of these problems can predict,
on average, how hard the problems are to solve by a variety of search
methods.  A major result of this work is that hard instances of
NP-complete problems are concentrated near an abrupt transition
between under- and overconstrained problems. This transition is
analogous to phase transitions seen in some physical systems."

In the current context, this implies that the computational costs
of determining the healthy state are likely to be tractable unless
the damage to the neuronal tissue has been sufficient to render
such a determination almost infeasible.


Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=4640