Comment

Daverz August 9, 2006

How about running replace on the string before normalizing:

title.replace(u'\xdf', 'ss')

and so on for any other special cases.