[Spambayes] French ?

[Tim Stone]

On Wed, 31 Mar 2004 09:49:55 +0200, Didier Morandi 
<didiermorandi at nerim.net> wrote:

> Good morning Madam,
>
> I'm Doctor Didier Morandi from Nigeria :-)
>
> Will this message go to your spam folder? (I hope).

Funny!  FYI, my spambayes rated this mail with a .01% probability of being 
spam.  Not bad, huh?
>
> Seriously, I'm more than interested to work on the French port of 
> SpamBayes. I did not find in the FAQ whether there is currently such 
> work or not.

There currently is not work on a French port, but, read on....

>
> Please let me know if this may be of interest. For us French speaking 
> people around the world, it is easy to tell our filters to consider 
> English as a spam language (which is obviously stupid) but we have no 
> way to filter French spam via keywords statistical analysis and I guess 
> this French spam practice may pop up one day in our countries. So...

Spambayes really doesn't care what language the spam is in.  As long as 
the program can tease apart words by grouping characters using spaces and 
punctuation marks as separators, the mail could be written in any 
language.  The mail is separated into words, and those words are compared 
statistically to words that it has been trained (by you) to recognize as 
good mail or spam.  The meaning of the words, or even whether or not they 
are valid words in any language's dictionary, is never even considered by 
the program trained.

I believe there are Spambayes users in all major language groups.  We have 
had a few bugs from the Asiatics about how their characters are handled, 
and those problems have been corrected (I think).

A French port (mmmm... sounds good!) of Spambayes would be primarily to 
change the user interfaces to French.  That might be very useful, but thus 
far, nobody has stepped forward to do such a thing, and all of the 
developers speak English as their primary language (I think)...

-- 
Tim Stone