Naming Conventions

[THargreaves]

The new state containers/storage introduce in #7 are looking pretty elegant. Was surprised how nicely it worked with the RBPF and how it was completely generic to the inner distribution type.

If we are keeping this, we should think a bit more about the naming conventions we are using. Currently there are about 3 things we're calling a state. We need names for the following things:

1. The object that stores the entire history of the filter
2. The object storing everything for the current step including the filtered/proposal distribution and any extra stuff that can be accessed from the callback (e.g. innovation matrix)
3. The distribution-like objects representing either the filtered or proposal distribution
4. The value that a standard particle has (typically a vector)
5. The names for the components of a Rao-Blackwellised particle (that contains (4) and (3))

Some preliminary suggestions I have are

1. Storage
2. Intermediate
3. State? Distribution?
4. State?
5. X and Z?

[THargreaves]

Following on from this, I think we should probably separate the code for temporary (within step) storage and historical storage into different files.

Maybe even further separate the distribution objects and the more general intermediate storage (which contains distributions).