> Any deadline?

I'll share the level 2 solution end of June.

> Do you want one or all solutions?

At least one, the more the better.

> Anything besides "closed" polygons seems trivial.

Indeed.

> Coding challenges usually come with a test suite and an API.

Yes but I posted this w/o preparation to help bliako with his request.

See it rather as a mediation on a possible challenge.

But I'm starting to believe that it's too challenging on the mathematical side (though it's part of the curriculum of my 15 year old niece :-/ )

> If you say "bigger grids", are they still equidistant and quadratic? I suppose n x m is still legit?

Equidistant: yes

Quadratic: no

Disclaimer: But I'm not aware of higher level solutions yet.

> And if you want all solutions you might exclude symmetries by rotation or mirroring.

I think it should be obvious that you need to handle symmetries when trying to find all solutions.

The complexity of a branch and bound algorithm might explode otherwise.

> I forgot: Any language restrictions? Is Perl 6 ok?

Any Perl*

> A solution in JavaScript could also draw the polygon.

You are free to generate JS or SVG, but the algorithm has to be Perl.

BTW Haukex's WebPerl would do too.

A solution in JavaScript could also draw the polygon.