1: My current plan is to fork the CLI and/or GUI, implement binning as a proof-of-concept, and also discuss the trade-offs in min-untraceability for different mixin/bin sizes closer to what might be used with Triptych since the Miller at al paper only went up to 8/4 respectively. Once implemented however, I'm not really sure what, if any, metrics should be measured/looked at. My current understanding is that binning just changes the mixin vector for the ring signature so I don't really expect it to impact performance or any other metrics (See question 3). My question is: Does it sound like I'm taking this project in the right direction? Ideally my project will be genuinely productive/useful so it would be nice to hear your opinions before I dive deep :)

2: In the event that binning alone is not sufficient for my project, would someone be able to point me towards a resource on ring reuse? I searched myself, but to no avail.

3: Moneromooo mentioned that binning may be used to decrease the space needed to describe a ring. The Miller et al paper does not mention anything like this and I also can't see how this would be the case: Binning is a different method for creating a mixin vector but the final product is still just that, right? There should be no size difference (but maybe this is because I don't know about ring signatures yet).