- Find a way to make the implicit symmetrizations, GeCCo assumes, more explicit. For this, super-vertices need to be taken into account as those must not be symmetrized with one another.
- Find a way to cleverly split to-be-carried-out symmetrizations in a clever way (c.f. SR case: Only antisymmetrize one index-pair & column-symmetrize final result tensor only)
- Use approach given in https://stackoverflow.com/a/11327600 to perform common subexpression elimination (CSE). Also take https://doi.org/10.1007/11758501\_39 into account