Any number system which support a multiply will have a range over which multiply is most efficient. Multiply efficiency depends on the sample rate relative to digit length. My quest was never about physics, it was all about aligning a number system up so as to reduce the number of variables and equations that physicists and economists use.
Why did I pick Fibinacci?
50% dumb luck and a sense of ratios. And 50% because I knew they we Shannon compatible and approximately added the previous set and maximally packed. And this, I just discovered:
The most common such problem [for Fibonacci] is that of counting the number of compositions of 1s and 2s that sum to a given total n:
I knew combining packs of Nulls was important. Also I klnew the exchanges were mostly about combining the previous with the next, whihh made conservation of addition important.
This series is convergent for and its sum has a simple closed-form:
The formula above is the SNR for a three sphere packing. The x series at the bottom is noise, proportional to area of the current, minus volume of the of the layer below behind, I think. Something I have not worked.
But when I started out, I had this ides that packing could be either 2 or three quants of Nulls, that they would be a bit underpacked. So I was going with 2+3 is best approximated by 3 times Fibonacci, for the next quant up.
Anyway, group theory, maximum entropy counting and finite systems have won the day. I doubt physics can really go back to old style integration and stretchable space.
No comments:
Post a Comment