Implement lookup as ranked list

[jonthegeek]

Figure out a good rank order to use for the lookup. This should probably be a process I can repeat occasionally (although probably not EVERY time I build the lookup, 'cuz I suspect it will slow it down by ~days).

Change the lookup to a named list (names = word, values = morpheme breakdown). The list should be in frequency order (by number of times it is "hit" while building the full lookup).

To determine: How do we rank a list that's smaller than the gigantic lookup? Ideally some sort of corpus should be involved, other than wiktionary.