testing questions

[David Relson]

On Thu, 04 Mar 2010 18:35:05 -0900
barsalou wrote:

> Let's just say that I wanted to take an empty database and make the  
> first message I sent to bogofilter return spam?  What would the  
> settings have to be?
> 
> Additionally, what would I have to do to make it return ham?
> 
> I wasn't planning to use a corpus of messages....just wanted to make  
> bogofilter return the result I wanted through specific settings.
> 
> Is that possible?
> 
> Mike Barsalou

Mike,

That's an interesting question.  I don't know if it's possible, but
here are some ideas to try.

Create the dummy wordlist and a test configuration file.  The four
critical settings are robs, robx, ham_cutoff, and spam_cutoff.  robx is
the score given to an unknown token, i.e. one not seen before.
Bogofilter's default settings put robx between ham_cutoff and
spam_cutoff so that unknown tokens will receive an 'unknown'
classification.  When robx is larger than spam_cutoff, unknowns will be
considered spam. When robx is less than ham_cutoff, unknows will be
considered ham.

Give it a try and report back :->

Regards,

David