Comparing lists rather than strings?

[cjmcmurtrie]

Great library!

It would be very nice to be able to compare lists (for instance token lists) rather than per-character comparisons. For instance if a token is a unit of text rather than a character,

ratio('i tried', 'he is trying')

Should be zero, as there is no way to equalize them by token replacement.

difflib.SequenceMatcher(None, 'i tried'.split(), 'he is trying'.split()) return zero when splitting the strings to tokens.

I dug around and found some intimidating c code in the back-end of Levenshtein. Have you been asked this before?