Change History

rragan Wrote:
-------------------------------------------------------
> I think he assumed a polygon on a planar earth
> projection. Farthest NSEW finds would give 4
> bounds Then picking two more gets interesting. If
> the first 4 are over 10% you could almost call it
> quits. Failing that, most NE,NW,SE,SW would be
> tried for the best two I guess. Current stats
> already list most northerly, easterly, etc. If
> those are "free", that could save time.

This is exactly what I mean. At the beginning, the tasks seems to be simple, but then more and more exceptions and additional terms, just because it does not work otherwise as was thought.

To be correct from the beginning, the challenge should start from requesting 10% area of rectangular world map in specific projection to be covered, not the surface of Earth. Then we can build a single convex hull covering all finds and then find all subsets of any arbitrary number of caches quite easily to find the largest. It is possible to cover the map from 0% to 100% with this kind of rules. The checker script needed for this kind of challenge is possible. You only have to request it for you challenge.

ps. I suggest Mercator projection because it is easiest (=only) to visualize in the checker.

rragan Wrote:
-------------------------------------------------------
> I think he assumed a polygon on a planar earth
> projection. Farthest NSEW finds would give 4
> bounds Then picking two more gets interesting. If
> the first 4 are over 10% you could almost call it
> quits. Failing that, most NE,NW,SE,SW would be
> tried for the best two I guess. Current stats
> already list most northerly, easterly, etc. If
> those are "free", that could save time.

This is exactly what I mean. At the beginning, the tasks seems to be simple, but then more and more exceptions and additional terms, just because it does not work otherwise as was thought.

To be correct from the beginning, the challenge shoud ld start from requesting 10% area of rectangular world map in specific projection to be covered, not the surface of Earth. Then we can build a single convex hull covering all finds and then find all subsets of any arbitrary number of caches quite easily to find the largest. It is possible to cover the map from 0% to 100% with this kind of rules. The checker script needed for this kind of challenge is possible. You only have to request it for you challenge.

rragan Wrote:
-------------------------------------------------------
> I think he assumed a polygon on a planar earth
> projection. Farthest NSEW finds would give 4
> bounds Then picking two more gets interesting. If
> the first 4 are over 10% you could almost call it
> quits. Failing that, most NE,NW,SE,SW would be
> tried for the best two I guess. Current stats
> already list most northerly, easterly, etc. If
> those are "free", that could save time.

This is exactly what I mean. At the beginning, the tasks seems to be simple, but then more and more exceptions and additional terms, just because it does not work otherwise as was thought.

To be correct from the beginning, the challenge shoud start from requesting 10% area of rectangular world map in specific projection to be covered, not the surface of Earth. Then we can build a single convex hull covering all finds and then find all subsets of any arbitrary number of caches quite easily to find the largest. It is possible to cover the map from 0% to 100% with this kind of rules. The checker script needed for this kind of challenge is possible. You only have to request it for you challenge.