Fabian Zickgraf

Some additional thoughts from a discussion: There are two settings where omitting the category might make the code more readable: 1. In derivations which only deal with a single category....

The commit closing this issue (bcf382c90f19816242b8ec355ed40c8d26cb8a8e) comes from the import of existing code as `CartesianCategories` and has nothing to do with this issue -> reopen.

> In particular, we'll need to take care not to run through caches twice. Good point. I think it should be safe to create a redirect function which simply goes...

That would be fine for now, I have not found other instances. I will prepare a PR.

A further question regarding this topic: Do I have to fulfil the specification ``` IsEqualForMorphisms( KernelEmbedding( alpha ), KernelEmbeddingWithGivenKernelObject( alpha, KernelObject( alpha ) ) ) ``` if I install both...

> The answer is yes :) Okay, then for the record/future reference: this issue in particular implies that if I use `IsIdenticalObj` for the equality of morphisms and rely on...

No, I don't, I just try to understand all of this in detail and try to write it down in case I need it in the future.

> A first remedy is to provide an `IsEqualForMorphisms` operation in Rows/Cols/MatrixCategory that also checks `HasIsZero` of the homalg matrices So `IsEqualForMorphisms( phi, psi )` should return false because `HasIsZero(...

> 3) such a version of `MatricesForHomalg` does not exist to my knowledge As far as I know it exists in theory, simply by loading `LIMATEmp.gi` instead of `LIMAT.gi` in...

> and measure the effect Do you have concrete tests in mind? If yes, I/we could simply comment out the optimization(s) in `MatricesForHomalg/gap/LIMAT.gi` and run the tests. I just did...