This YTEP proposes a new framework for analyzing halos in yt 3.0. None of the proposed changes will affect yt 2.x. Most of this is new machinery which existing technology (halo finders and merger tree generators) can be plugged into.
This YTEP is the product of a breakout group at the yt developers workshop held in Santa Cruz a few weeks ago. Thanks to the group for helping craft the ideas introduced in this proposal and to the workshop organizers for making this happen.
Update 1: I added text to respond to the first round of comments. This includes discussion of file formats for the HaloCatalogTimeSeries and HaloCatalog objects and some clarification of when particle information will be written to disk and stored in memory. Thanks to everyone who commented.
This is really compelling. Thanks to everyone who contributed. I think my only comment is that it would be nice to mention interop with external tools, which I think this actually helps. I also think it might be worthwhile to mention that performing analysis during the course of halo finding is how many other halo finders (notably Rockstar) operate now.
I think this looks good!
Looks good to me, Britton. Do you think this is ready to accept?