lexer change

[Tom Anderson]

On Tue, 2003-11-11 at 19:11, David Relson wrote:
> Out of curiosity, what would _you_ classify as a 10% improvement?

Well, if there were a rule which reduced 10% false positives to only a
fraction of 1%, then I'd say it would be worth looking into.  Going from
0.3% to 0.2% is just coincidental and not worth implementing it.  And
even the significant rules should be implemented only after serious
thought into their potential side-effects and future effectiveness.  And
even then, they should be kept in a seperate pre-filter library so as to
be distinct from the primary functionality and thus easily removed later
or even turned off with a command line switch.

Mixing Bayesian filtering with rules-based filtering is like
entitlement-driven quasi-free-market "pinko" capitalism... it becomes
bloated and doesn't quite accomplish the intended effect.

Tom

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://www.bogofilter.org/pipermail/bogofilter/attachments/20031112/c5943c56/attachment.sig>