Should I glob tokens?

[Thomas Anderson]

If they have barely occurred in years, they should not be affecting 
classifications anyway, so you can just delete all of them.

Tom


Charles Hewson wrote:
> Hi all,
> 	I am planning upgrade from 0.94.12 to 1.2.0 and have been looking
> at wordilst accumulated over 5 years. About 10% of tokens start with
> currency sing "$" followed by digits. Most of these occured once in spam.
> The highest count occured 17 times in spam in 2005. Only 4 of these tokens
> ever occure in non-spam. It would seem that a combined token could be more
> efficient?
> 
> Any thoughts,
> Charles
> 
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> pub  1024D/F88852DE 2008-06-25 Charles Hewson <cahewson at eskimo.com>
>      Key fingerprint = 0779 BBA4 CF82 0707 288B  3B37 BDB7 3DC3 F888 52DE
> sub  2048g/71B13048 2008-06-25 [expires: 2009-06-25]
> 
> 	 (For info see http://www.gnupg.org)
> 
> Public key at - HTTP://WWW.ESKIMO.COM/~cahewson/pubkey.asc
> 
> _______________________________________________
> Bogofilter mailing list
> Bogofilter at bogofilter.org
> http://www.bogofilter.org/mailman/listinfo/bogofilter
> 
>