Scope limits

Another good meeting with my adviser today. Including lots of new avenues for exploration. Fortunately, none of those need to be included in the current iteration of "the paper" (I don't know what else to call it since the original title no longer applies). We basically put some  limits around what we're going to put in. I need to work fast so I can hit those before those boundaries get moved.

Some of the new topics are interesting. They include:

• Defining a multinomial distribution to describe the block/batch correlation. This might just be a re-parameterization of an existing distribution that I discovered in the literature last week.
• Using the Dirichlet process to drive/assess D-BESt.
• Finding more efficient ways of running the Dirichlet process (or, more promising, finding ways to pre-compute it).

All fun stuff, but stuff I'm not going to think about until we get this paper out the door.