Performance of 0.13

[Greg Louis]

On 20030523 (Fri) at 1026:14 +0200, Boris 'pi' Piwinger wrote:

> Great work, guys!
> 
> Spam:
>   11660 test.spam
> False negatives:
> 563
> Ham:
>   21736 test.ham
> False positives:
> 1
> 
> Then I installed the new version and did the same job:
> Spam:
>   11660 test.spam
> False negatives:
> 395
> Ham:
>   21736 test.ham
> False positives:
> 1
> 
> So this is really much better as expected.
 
Good to hear it.  It might be good to drop a note to Paul Graham
(pg_at_paulgraham.com) and tell him his refinements are working for
you.  I suspect he'd be glad of the feedback, and as you know, all
three changes in 0.13.0's lexical analysis were based on suggestions he
made (http://paulgraham.com/better.html).  More communication among the
various spam-filter projects might help us all.

-- 
| G r e g  L o u i s          | gpg public key: finger     |
|   http://www.bgl.nu/~glouis |   glouis at consultronics.com |
| http://wecanstopspam.org in signatures fights junk email |