Re: Natural Selection

[ Home ][ Thread ][ Subject ][ Author ][ Date ]
Tom Ditto
Tue, 3 Mar 98 13:05:03 PST


Martin wrote:
>IBM has developed a submicron
transistor with *three* states--on, off, and then some sort of middle state
whose function I wasn't quite clear on, but apparently it helps the
processing to function a lot faster/more efficiently.<

Right. I think Intel has one-upped them with an experimental
four level device. The push is to use fewer transistors for
more memory. The usual multiply by two laws apply: double
your rice grains and fill up your chess board, if you get
my drift.

Rock bottom binary has a special Boolean appeal,
but I was referencing the reliability issue in so far as
the speed of the circuits go. One key to the computer
revolution was the certainty that could be achieved with
a bi-stable circuit.

On the other hand, we have an alphabet with its 26 plus
characters, so it is certainly possible to model large
modulo systems. Hey, DNA works well enough with 4 to allow
for life to reproduce.

In an earlir entry to t&b I was, in fact, proposing a DNA-like
self-replicating archive so as to assure for duplicates of
records with or without conservators. Since this might be
chemical rather than electronic, transistors might have
nothing to do with the process, and the saturation of a
semi-conductor gate might be totally superfluous to the
archival record.

Don't get me wrong. I'm not postulating that one could
"eat a book." Nor is my seemingly strange proposition
intended to generate some low brow sci-fi pot boiler where
a book eats you. I am addressing the very sensible concern
expressed by several in this group that the archives should
not be considered as some kind of permanent physical object
to be tucked away in a time capsule to be discovered later.
Copying has been the method by which most records are
passed on from epoch to epoch, and the dynamic of making
copies must be part of the discussion.