No, for me commutators are intuitive moves. Like entire solves, they are groups of groups of moves that are in the "whole 3x3 substep". My definition is incomplete, but I knew you were going to understand it anyway.
No, you aren't

I also did, but I've learnt the optimal algorithms instead of the 2-gen one.
...
What do you think about this J-perm: R' U' R F2 R' U R F' R' U' R F' R' U R F' ?
Credit to DGraciaRubik for the idea of a J-perm made by treating pairs like pieces, and for writing this quite simple algorithm:
(r' D' r) U (r' D r) U (r' D' r) U2 (r' D R)
On
this thread.
I transformed it into the algorithm you see above using
this tool.
Maybe this one be better: R' F' R D2 R' F R D' R' F' R D' R' F R D'