Some results, like e.g. the ones of MR4352662, actually only hold if the dimension vector has no zero entries. This is not currently reflected in the software, which leads for instance to wrong computations of Mukai indices.
To this end, we might consider more general approaches:
- Should we have a
reduce() function that takes a quiver moduli setup and returns the supported setup only?
- Should we pass that to other methods?
- should we add an additional assumption check at the beginning of all the methods requiring nonzero entries in the dimension vector?
Some results, like e.g. the ones of MR4352662, actually only hold if the dimension vector has no zero entries. This is not currently reflected in the software, which leads for instance to wrong computations of Mukai indices.
To this end, we might consider more general approaches:
reduce()function that takes a quiver moduli setup and returns the supported setup only?