[EM] Election-methods digest, Vol 1 #388 - 7 msgs

[Markus Schulze]

Dear Matt,

you wrote to Mike Ossipoff (20 Dec 2003):
> To restore some credibility for yourself Mike, how about arranging
> to correct the code to make it O(N^3)?  All this time Mike is
> wasting futiley trying to beat down Markus could be better spent
> by soliciting help from someone who codes python to modify the
> code.  Then again, your attacks on Markus are so completely false
> and nasty that I really doubt there is anything you can do to
> restore your credibility with me.

Mike Ossipoff wrote (18 Dec 2003):

> repeat = 1
> while repeat = 1:
>   change = 0
>   for i = 1 to N
>      for j =  1 to N
>         for k = 1 to N
>            least = min(B(i,j), B(j,k))
>            if least > B(i,k):
>               B(i,k) = least
>               change =1
>            endif
>        endfor
>     endfor
> endfor
> if change= 0
>     repeat = 0
> endif
> endwhile

A compromise would be to write:

> repeat = 1
> while repeat = 1:
>   change = 0
>   for i = 1 to N
>      for j =  1 to N
>         for k = 1 to N
>            least = min(B(j,i), B(i,k))
>            if least > B(j,k):
>               B(j,k) = least
>               change =1
>            endif
>        endfor
>     endfor
> endfor
> if change= 0
>     repeat = 0
> endif
> endwhile

Here the professional mathematicians are satisfied because
it is garanteed that the triple-loop is passed only twice.
And those who don't have sufficient mathematical skills
to trust the Floyd algorithm have the while-loop as an
additional guaranty that the algorithm gives the correct
results.

Markus Schulze