Computational Linguistics 1, Exercise 2: splitting "ï"

Code (run with py -X utf8):

import unicodedata
w = "naïve"
print(len(w), len(unicodedata.normalize("NFD", w)))
print([unicodedata.name(c) for c in unicodedata.normalize("NFD", "ï")])

Output from the run:

5 6
['LATIN SMALL LETTER I', 'COMBINING DIAERESIS']

The result has 6 characters. "ï" splits into the base letter "i" and a
combining diaeresis. The Unicode article describes this: a combining mark can be
added after a base character, and the two encodings of an accented letter are
treated as canonically equivalent. NFD is the decomposed form; NFC (used in
the chapter) joins the pieces back together.