[Doctor Stochastic]

Easy. It depends on how much circuitry you want to devote to an instruction. This is not even as complicated as IEEE floating point arithmetic that most Intel processors have. There have even been fused multiply-add units for years.

The Perkin-Elmer (if they still exist) computers had instrucitons that did things such as (add, update index register, jump on condition) as a single instruction. The old IBM650 had the address of the next instruction to be executed in the instruction itself. (This was a drum memory machine and the relative location of instructions on the drum contributed to program speed. You didn't want to drop a revolution.)

It's only a curiosity but it is amusing that a single-instruction computer can emulate a universal Turing machine.

[Alamo-Girl]

Thank you so much for the explanation!

I have to admit I am rather astonished that hard-wired macros are now considered to be an "instruction." If I am to take such as an equivalence in understanding Kolmogorov Complexity then I no longer am in agreement with tortoise.

Basically, my understanding is that the Kolmogorov Complexity of an object is the length of the shortest computer program that runs on a computer and outputs that object. A short program outputting a very long number would not be considered complex by that rule.

But change the rules so that any size of macro can be hard-wired and presented as an “instruction” and I must disagree with the Kolmogorov approach to depicting complexity in biological systems. After all, an exhaustive macro subroutine, including numerous conditionals, memory arrays and symbolizations can easily be hard-wired.

Jeepers, there is no reason an entire subroutine couldn't be hardwired into a single "instruction." I could see whole sections of the Internal Revenue Code that would lend themselves to being hard-wired but that does nothing to explain biological autonomous self-organizing complexity.