bogus bogotuning

[David Relson]

On Fri, 30 Jan 2004 07:00:19 -0500
Greg Louis wrote:

> On 20040129 (Thu) at 1851:31 -0500, David Relson wrote:
> 
> > After reading Clint's comment, it appears that there are others who
> > want the opportunity to see what happens with a small message set. 
> > I'll consider adding the new option to bogofilter if someone will
> > supply a patch and if the patch includes the before _and_ after
> > warnings agreed on my Greg and Clint.
> > 
> > Jason - I believe the ball is in your court now.
> 
> In fairness to Jason, I think we need to provide the text, because
> Clint and I discussed it off the list, cc you.  

My mistake;  no intention of unfairness;  just one of distributing the
work load :-)

> Here it is (to be printed both before and after the regular output):
> 
> >  *** WARNING ***  The original authors of bogotune strongly advise
> >  against the use of the -f option, which was implemented without
> >  their approval.  The option bypasses bogotune's protective checks
> >  and allows bogotune to run without enough messages to ensure valid
> >  results.  This will often result in bogotune suggesting wrong
> >  parameter values that, if used, may significantly degrade
> >  bogofilter's accuracy.
> 
> If you decide to call it -F, as before, then s/-f/-F/ above,
> obviously.

Thanks, Greg.

-- 
David Relson                   Osage Software Systems, Inc.
relson at osagesoftware.com       Ann Arbor, MI 48103
www.osagesoftware.com          tel:  734.821.8800