Which filter class is the latest? Or should we use the standard
filter, and cope with the fact that it will give false positivies?

I noticed that the current filter gives several false positives, not
only for 'special' characters, but also for apostrophes; say that the
forbidden word is Clara, I will get a positive on both Clara and
Clara's, making it impossible to use verify to check my list...

I decided to 'cheat' by cleaning up my dictionary file from all the
'forbidden' characters and now I average 280ish with a dictionary
containing 127367 words and 10 forbidden words.

//F

Btw, as a sidenote I can say that the spamfilter we have at work have
the exact same problem. It doesn't know swedish characters, so when it
looks at 'medsēĻda' it finds 'meds' which is a forbidden word...