The required property for the child groups of a group should be more consistent

Currently restrictions defined for a child group can also affect the parent group. For instance, if for the child group a minimum constraint greater than 0 is set, the parent group will implicitly become required, even if its minimum is set to 0.

This is counter-intuitive. It would be more consistent if the minimum and maximum constraints of the parent group were independent of the settings in the child groups. Then it would be possible to model constellations like an optional parent group with a child group that has required elements (i.e. the child group is optional, but if it is present, all required elements must be provided).