[Shootout-list] New benchmark?

[William Douglas Neumann]

> Oh those are tedious discussions ;-)

Unfortunately, I don't think they're likely to end soon, as the 
distinction is somewhat arbitrary, and often subtle (cf the thread on 
Felix exceptions).

> What does the 'lines of code metric' tell us when one person has
> implemented insertion-sort and someone else implemented quicksort?

Well, it tells me that the guy who submitted insertion sort either needs 
to learn a bit about sorting, or that he looked at the input and saw that 
it was already sorted...

There's a (more or less) big difference between "same algorithm" and "same 
way".  For example, you can specify mergesort, but leave it up to the 
implementor if they want to use arrays or lists.  I'm sure there are 
other, less bad, examples to use there, but it's late in the afternoon, 
and "me brain no work good".

> What do the times and memory usage tell us about the languages when
> entirely different algorithms have been used?

What do they tell us when the chosen data structure is less efficient than 
another in some given language?  As you say at the top of the home page, 
"What are we really comparing  Languages or Programs; Libraries or 
Algorithms or Programming Ability?"  Plus, I think you'd find that the 
various languages would all gravitate towards the same algorithm, more or 
less ("Why is that guy using a randomized treap?", "Why are they using 
KMP?").  After all, everybody wants to be #1.

William D. Neumann
<wdnx@unm.edu>

FWO to the Nth degree!!!
---
Dear Lord, please make me the kind of person
my dog thinks I am.