[vidmaven]

Ideally, specifying a unique number of magnets per side would be possible, with the default behavior being even distribution along each side.
I wouldn't find it inappropriate to limit the polygon [max. # of sides] to 4, but the overall facility would be a huge improvement.
Further, a subgroup feature, allowing groups of magnets to be "clumped" (e.g.: 2 sets of 6 magnets, 2 sets of 2 =16 connections, subgroups, with weighted distribution, maintaining a ratio of >4:1 distance between subgroup separation and group member separation.