<alphamule> "<PlanckWalk> Obviously any recursive algorithm can be implemented iteratively"  Isn't that actually a third class:  until/while loops with a sentinal?  I mean, doesn't that contradict completeness theorem?
<alphamule> I mean, a for loop should always return if it's at the lowest level (assuming classical 'summation' concept of k = a to b).
<alphamule> Which implies it's definitely not complete.
<alphamule> Now, you can cheat and use something like "break" and make k = 1 to infinity or something stupid.  :P
<alphamule> But that just turns it into a sentinal.
<alphamule> That line does literally the same thing that C-like languages do when you have a condition check.
<alphamule> Just moves it around.
<alphamule> Halting Problem/Turing completeness relationship
<okuu> alphamule: You are making even less sense than usual, I'm afraid. The translation between recursive procedures and imperative iteration (in either direction) is not only feasible - it is routine. Compilers do it.
<dude12312414> it's not necessarily clear when a for or while loop terminates
<dude12312414> so this does not contradict the Halting problem
<okuu> It is similarly not necessarily clear when a recursive procedure terminates. :-|
<dude12312414> true
<okuu> And what does termination have to do with this anyway?
<alphamule> "Compilers do it" bullshit
<alphamule> The C-style 'for' loops are a poor example syntax.
<alphamule> It is guarenteed that a loop of finite number of steps per iteration will halt if there's a finite number of loop iterations.  This is not the same thing.
<alphamule> nvm
<alphamule> It's not like they're different or something.