[spambayes-dev] accented characters in subject

[Serge Orlov]

Hello, Paul,

> Further to my email a few weeks ago, as well as punctuation in the middle of
> words in the subject line I am also starting to get spam using umlauts,
> accents and that kind of thing in the subject line.
And they usually appear as non-converted 8 bit characters, right?
If so, I suggest generating header:8bit token. This token will be a good
spam clue since spamware quite often don't care about standards.

> 
> I am not sure what the conclusion was re the efficacy of removing
> punctuation but if it is an improvement then folding extended characters
> into plain ASCII might also help.
What is the difference between this trick and V1agra? Spambayes
coped with V1agra quite well. I keep seeing it from time to time, but
spambayes finds enough clues to overcome V1agra.