When I began my degree in Biology I thought I was making a ‘clean break with the past’. Last wednesday Dr. Jenkins, who has been kind enough to employ me in her lab, suggested that I approach Biology from a computational perspective. In thinking about this, it struck me that turing machines are to the intellect what telescopes are to sight.
This came out of a ferment that included a moderate amount of reading lately on Cellular Automata (CA) and reversible functions.
With a compact definition of CA’s, you could use a tuple of
(CArule_1, CArule_2, indexing_string) to generate an arbitrary set of Merkel Hash Trees.
That way, if you ever ran into the situation where a given tree had collisions, you could add another hash tree just by using two more CA rules.
Of course, this is all predicated on finding CA rules that are reliably one – to – one. Possibly this means I could use a graph expander despite their ‘mere’ probabilistic certainty of avoiding collisions.
I’m not at all certain this thinking will stand the light of day, but at the very least it’s led me to the door of graph coloring, which might have ways to let me test candidate CA rules.
Presuming, of course, you could find said subset of CA rules, using the virtualization technique described in a paper I’m too sleepy to find the reference for at the moment, you could make them reversible. Which would get you the big fat prize of being able to reverse the process: given knowledge of a hash fragment you can recreate the original