Slices time complexity

[Rustom Mody]

On Wednesday, May 20, 2015 at 12:16:49 PM UTC+5:30, Marko Rauhamaa wrote:
> Rustom Mody :
> 
> > In short any language that is implemented on von Neumann hw will need
> > to address memory.
> 
> I don't think von Neumann hardware plays a role here. I think the data
> model is inherent in Python/Java/Lisp regardless of the underlying
> formalism (which could be SKI combinatory calculus or any other
> Turing-complete formalism).

That's backwards

Python/C/Java/Lisp/Fortran (and 700 more) are made the way they are out of
aiming for some fidelity with von Neumann architecture
They all of course make different choice from choosing different weights in
the various tradeoffs.
However they all commonly share that they are implemented reasonably satisfactorily.
'Reasonable' and 'satisfactory' have unspecified semantics!!

> 
> > And although they are equal as in '==' they are not equal as in
> > behavior, memory usage etc, a fact that can only be elucidated by
> > box-n-arrow diagrams.
> 
> That's what I meant when I asked if you can get Python without getting
> something like C first. It seems the answer is, you probably can't.

I think we agree except for the 'probably'. Consider:

Q1: Is a computer (of any sort and generation of your choice) finite or infinite?
Q2: Is a Turing machine finite or infinite?
Its ironical that a Turing machine is 'more' infinite and therefore more a math
abstraction than lambda-calculus!

IOW I regard memory questions as more fundamental than you seem to