[computer-go] Voronoi diagram revisited

[Hideki Kato]

Jason House wrote:
>Hideki Kato wrote:
>> ++++++
>> +X+ at ++
>> OXO+++
>> OOXX++
>> +OO+++
>> 
>> Corresponding Voronoi diagram with Manhattan distance is (+ is neutral point):
>> 
>> XXX+O+
>> +X+O+X
>> OXO+XX
>> OOXXXX
>> OOO+XX
>
>I think that last diagram was not generated with manhattan distances, 
>but instead euclidean distances.  I believe the following would be the 
>correct diagram for manhattan distances:

Ooops, you are right. I was confusing. That diagram is of my special distance 
not Manhattan nor Euculidean. The distance is defined to approximate the number 
of moves to connect, i.e., if there are 2 paths with 2 Manhattan distance, the 
distance is computed to 1 (= 1/(1/2 + 1/2) the same as parallel connection of 
resisters).

>+X++++
>+X++++
>OXO+++
>OOXXXX
>OOO+++

gg
--
gg (Kato)