Can we be more specific about the type of returned prior?

[willwerscheid]

instead of saying the class of object (normalmix, unimix, etc) we could try to infer the family (normal, point-normal, etc) from the fitted prior