Vyas Ramasubramani

OK that sounds fine. In that case I'll try to review this again sometime soon.

Fixing this requires some nontrivial internal changes. We need to be much more consistent internally about when objects own their own boxes, and we'd probably need to create a new...

You would also need to validate that the original positions are still valid in the new box, etc. So yes, I think failing is probably the safest approach.

I think you are correct that it appears to be a definition issue. In the thermodynamic limit the two results are identical, but for the finite systems we're calculating they...

I like the idea of a normalization flag, what would you name it? I would definitely like to add an equation or two to the documentation, perhaps pointing at [Wikipedia's...

Hey @mphoward I've opened a PR #529 to fix this. Sorry it took us so long, we've been in the middle of a thorough rewrite and this issue fell by...

As part of glotzerlab/freud-examples#19 @jennyfothergill has expressed interest in resolving this issue for the general case of pairs from two arbitrary subsets of the system, i.e. `points` != `query_points`. We...

Yes, that's where the normalization is currently happening. However, I'm not sure if your suggestion is exactly right. IIRC there's a prefactor that's being computed just above this number density...

I agree that there's an important distinction here. The concept of exclusion is simply implemented differently in freud. In VMD you have a complete system, then you subset this system...

I think my preference would be to provide recipes for how a user could accomplish this rather than incorporating it into the API. There are enough ways that we _could_...