bogus bogotuning, was: What is a spamicity of exactly 0.5?

[Jason A. Smith]

I'm not asking for the minimum limits to be changed or removed, I'm
simply asking for an option to be added to bogotune to disable the
enforced minimums.  The default would be to still use the minimum
message counts.  I don't see why this is such a big deal.  The option
can be accompanied by an obnoxious disclaimer or warning if you think it
is necessary.  For the user, like myself, who doesn't want to edit the
sources and recompile bogotune every time a new version is released, or
doesn't want to be forced to wait to reach this minimum, this option
would be very nice.  Luckily, I don't get a lot of spam, but this means
I will have to wait several months to accumulate enough spam to finally
be given permission to run bogotune, despite already having over a
thousand spam email saved.  My only other option is to try and find
where the limit is in the source and edit it myself and recompile.  A
much easier solution would be to just add an option.  Why is there so
much resistance to this simple idea?  Opensource software is about
freedom, right?  I know i can edit the sources myself, but I don't want
to waste my time if there is so much opposition to this option and I
don't want to maintain my own version of bogotune for the next several
months.

~Jason


On Tue, 2004-01-27 at 13:45, Greg Louis wrote:
> On 20040127 (Tue) at 1138:50 -0500, Jason A. Smith wrote:
> 
> > Also, I would just like to be able to run the bogotune command without
> > having to wait until my ham & spam folders reach some arbitrary limit. 
> 
> The reason I imposed those limits is simply that bogotune's results are
> quite worthless if you don't have significant numbers of messages in
> the training db and a reasonable set of test messages for it to chew
> on.  Rules, they say, are for the obedience of the ignorant and the
> guidance of the wise and knowing; if you can theoretically justify
> running bogotune on fewer than 2000 each of ham and spam in the
> training db and 500 of each in the test set, you're welcome to
> recompile your copy of bogofilter with those limits removed.
> 
> All I ask is that if anyone does that, they don't subsequently flame me
> on the list with "bogotune gives garbage output!" -- that complaint
> will receive a very unenthusiastic response.
-- 
Jason A. Smith <jazbo at jazbo.dyndns.org>