X-Message-Number: 10403
Subject: Re: Electron micrographs
Date: Wed, 09 Sep 1998 13:06:00 -0500
From: Will Dye <>
Charles,
Please add me to the list of people to whom you'll send a copy of the
much-discussed electron micrographs. My reason for not asking earlier
is that I don't have the time to read them. But now that I think of
it, unless I get it into my in-box, I'll probably space it off
altogether. I told myself to do this several months ago, when Saul
Kent was wondering why more people didn't seem to care, but never got
around to it. With two jobs, two kids, and a bad case of adult
attention-deficit disorder (I'm talking to a doctor about the latter),
I'm kept pretty busy. But this is important, so if I can get it into
my in-box, I should get a chance to read it during one of my incessant
mental digressions. :-)
Here's the address:
William L. Dye
c/o Design Data
121 South 13th Street, Suite 204
Lincoln, NE 68508
FWIW... Ever since I read Ralph's paper, "The Technical Feasibility
of Cryonics", I've wanted to contribute something substantal by coding
a computer model of atom displacement from freezing. Initially, it
would be useful just for explaining what we're talking about. Later,
enough real-world measurements could be added to turn it into a
genuine tool for modelling information loss.
Imagine it this way: initially, it's just a 2-d arrangement of colored
dots ("atoms") on a screen. You start with the initial arrangement of
atoms, then put it through a transformation process that displaces the
atoms ("freezing"). Perhaps you have a crystal-like growth in the
middle of the grid that sucks certain atoms towards it, and pushes
others away. Afterwards, you have the post-freezing picture. Now the
trick is to see how well you can take the "after" picture and
reconstruct the "before" picture.
From this crude start, you add in more and more plausible real-world
constraints. Covalently bonded atoms won't push apart nearly as
easily as non-bonded atoms. Chemical reactions occur during the
freezing process as the result of chemical concentrations getting
messed up. We really oughta include that pesky third dimension
sometime. What about DNA strands and cell walls? Yadda, yadda,
yadda....
Much attention will have to be paid to the issue of "reasonable
assumptions". Molecular modelling will chew up as much CPU as you
got, and then some, so you have to spend some time deciding what's a
reasonable thing to ignore, then document why it is reasonable. As we
move to a full 3-d sim of several complete *cells*, we'll definately
need to avoid modelling each atom ab initio. The final product may be
more like FEA analysis (the stuff they use to model steel and such)
than traditional molecular modelling, but time will tell. Probably
some combination. I have some ideas on how to handle this, however.
It turns out that they're running into the same problem in nanotech,
trying to merge molecular modelling with mechanical CAD. I program a
CAD system for structural steel, wherein a lot of techniques have been
developed to handle similar problems, so it's a natural fit.
Much attention will have to be paid to the issue of what constitutes
information loss. An easy way is to just say that loss equals the sum
of the distances each restored nucleus is from their original
position, to an accuracy of some pre-chosen fraction of an atomic
radius. That doesn't, by itself, take into account which atoms are
bonded, of course. Much more importantly, it is too rigid a
measurement. We really don't care if that vitamin C molecule was
rotated, as long as its bonded. We don't care if that mRNA was at the
top of the cell or the bottom, since it's supposed to float around
anyway. We don't care if the entire body was moved ten feet to the
right, or squished by 1% in one direction, because that happens to the
cells all the time in the real world. So the crude "distance"
measurement is not acceptable as a measure of information loss. But
it's a start. My point is that information loss is a subtle concept.
We need to flesh it out more.
Such a program would be debatable and inaccurate and inadequate, of
course, but it would be a tangible model that we could "sink our
teeth" into in terms of solid theories and mathematical models. It's
much too easy for a debate to break down when all you have is an
exchange of verbal descriptions. Perhaps a computer model would
encourage us to ask the right questions; and perhaps by expressing our
ideas in the form of additional code, we'd would do a better job of
sussing out the good ideas and scaring away the bad ones. Code as
hypertext, in an abstract sort of way. :-) Anyway, that's my
hypothesis. But I've been wanting to do this for *years* now, and
still haven't found the time, with no end in sight. Sigh.
Why am I spending time writing this? I should be at work, adding an
enhancement to our software, but instead I'm digressing yet again.
AAAARRRGGGHHHHH. You see what I mean? Attention-deficit disorder.
It's so hard to keep focus on a task for more than a few minutes at a
time. Oftentimes less than a minute. Sigh. I hope the doctor comes
up with something. I'm also going to order some pills from Saul's
company that supposedly help your brain, but I'm not sure if that will
make things better or worse. I've always had lots of mental energy, I
just can't keep it harnessed. Why it's become so much worse the last
few months, I don't know, but it's affecting both my jobs. Scary.
Anyway, Charles, since I wrote all that stuff about the computer
model, I'll cc this to Cryonet. Maybe it will be of interest there.
I was VERY glad to hear that you are having fun with your job these
days. I look forward to reading more of your pro writing -- and also
the CryoNet writing. We may clash now and then, but I really do
appreciate your contributions. You've a good mind, and a genuine
desire to save lives -- a pretty good combination.
Wishing you & yours all the best,
--Will
_________________________________________________________________________
William L. Dye \ "...it would seem that our Lord finds our
\ desires not too strong, but too weak... We
\ are far too easily pleased." --C. S. Lewis
Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=10403