use distribution functions instead of passing parameters direct

[n8thangreen]

other similar package, e.g. multinma, baggr, use things like normal(0,1) or gamma(1,1) to specify the input argument prior. At the moment I just pass the name of the argument, e.g. in the previous example mean=0, sd=1 but its probably neater and more user-friendly to switch.

Basically, what these distribution functions do is just return a list of distribution names and parameters.

[n8thangreen]

actually this is already in Issue #26