|
|
|
|
|
|
by viraptor
743 days ago
|
|
|
Have you got a source for e as the optimal branching factor? (On average it feels plausible, but haven't seen a confirmation) But I don't get the idea about writing numbers in base e - by definition integers in base e are the least efficient, because they'd be infinite. |
|
|
A non-integral number like e is obviously of limited utility in physical systems because having e adders in your ALU doesn't make much sense.
3 is closer then 2 to e, which also why ternary arithmetic was attempted for a while, but the massive advantage of binary in terms of implementation (it is much easier and far lower power to use transistors to slam a voltage high or low than keep it somewhere in the middle, and easier to arrange gates to have only on and off as inputs) won in the end.