Ara.T.Howard wrote:
> On Wed, 19 Oct 2005, Ryan Leavengood wrote:
> [...]
> it's interesting to note their solutions are an order of magnitude 
> better that
> the next fastest and __5__ orders of magnitude better than the worst.
> defintitely a strong argument to always consider re-working algoritms 
> before
> using the c compiler and extconf.rb in anger.
> 
> regards.
> 
> -a

Hmm, it doesn't effect the winners, but i would still like to mention 
that for a random testset i get these results:

david     [11, 34, 36, 37, 0, 33, 23, 9, 16]
jeff      [34, 34, 36, 37, 36, 33, 36, 34, 0, 37, 37, 23, 9, 16, 11, 11]
jegII     [34, 36, 37, 33, 0, 23, 9, 16, 11]
markvh    [34, 36, 37, 33, 0, 23, 9, 16, 11]
martin    [16, 11, 0, 33, 23, 34, 36, 9, 37]
paolo     [34, 36, 37, 33, 0, 23, 9, 16, 11]
park      [16, 33, 0, 11, 23, 34, 36, 9, 37]
paul      ["0", "9", "11", "1", "16", "23", "33", "34", "36", "37"]
robert    [16, 33, 0, 11, 23, 34, 36, 9, 37]
samk1     [16, 33, 0, 11, 23, 34, 36, 9, 37]
samk2     [11, 34, 36, 37, 0, 33, 23, 9, 16]
simonk    [11, 34, 36, 37, 0, 33, 23, 9, 16]
simons    [34, 34, 36, 37, 36, 33, 36, 34, 0, 37, 37, 23, 9, 16, 11, 11]
zach      [34, 36, 37, 33, 0, 23, 9, 16, 11]

which are at least 3 different 'solutions' (simons & jeff 16, paul 10, 
all others 9)

ok, simons and jeff can be fixed by a simple 'uniq', pauls seems to be 
broken (or made for other data?)

cheers

Simon