Confocal logic has hard-coded 'x, y, z' cartesian axes

[latchr]

While reviewing Pull Request #375 a substantial conversation was started around the fact that confocal logic has 'x', 'y', 'z' axes hard-coded into its class attribute names (for example, image_x_range). It is better to continue that conversation here as an issue separate from the immediate needs of that pull request (which have been simplified into a quickfix at #378).

There are a few separate questions:

• Should confocal logic be able to handle arbitrary axis names, or is that something for the hardware module?

• Should confocal logic be able to handle arbitrary coordinate systems (cartesian, polar, cylindrical, spherical, hyperbolic, etc)?

• If it should handle arbitrariness of either kind, how should this be implemented (dictionary, namedtuple, base class inherited into specific daughter classes, etc)?