Extend Cardinality Constraints

[SundermannC]

Currently, the cardinality constructs in UVL are limited regarding several aspects.

1. The interplay between feature cardinalities and group cardinalities may be undesired for some use cases. Currently, a feature selected k times would still only be considered as 1 for the selected between n..m features of the underlying group (defined as group cardinality here: https://dl.acm.org/doi/abs/10.1145/3634713.3634719). Having a way to specify that k is added to the number of selected features in that group instead of 1 (group instance cardinality) could be helpful.
2. Cardinalities can be specified in the tree hierarchy but not in cross-tree constraints. It may make sense to have cardinality constraints over features that are not part of the same hierarchy group. The advantage here is that the same reasoning engines as for group cardinality within the hierarchy should also be capable for group cardinality in cross-tree constraints.
3. Wildcards/infinite selections of a feature are currently not supported. It may make sense to model selecting n..* features.

Some questions to consider for this discussion:
Which use cases do we have for the presented scenarios 1.-3.?
What is the impact on scalable automated reasoning? This is particularly interesting for infinite cardinality bounds.
Would it make sense to introduce a dedicated cardinality language level?