> From: Mathieu Bouchard
> Subject: [ruby-talk:11181] Re: Generators (was: RCR Summary 02/16/01
> -suspend)
[...]

> Despite a regular functional implementation of fib(n) being in O(fib(n)) 
> time (that is, O(1.618**n) time), the implementation with generators is in
> O(n) time like yours [*]. The only difference is callcc (which I think is
> quite slow, but I don't know how it's implemented).
> 
> I don't see how fib wouldn't be a good example. The existence of an
> O(fib(n)) solution is irrelevant. You can design a slow solution to every
> problem possible.

I was curious how the ``callcc'' solution (much more interesting
IMO) compares with a straightforward ``suspendable'' solution.

Fib is fine a example - but in my loop I reinitialized to ``a = b = 1''
after 30 iterations (you have to alter the callcc solution a bit for 
this) to get a more meaningful comparison.

[..]
Christoph