bogus bogotuning

[Joerg Over Dexia]

Hi, I'd like to chime in on this, although it's almost off-topic
now...

Am 06:46 28.01.2004 -0500 teilte David Relson mir folgendes mit:
->On Wed, 28 Jan 2004 06:40:30 -0500
->Greg Louis wrote:
->> On 20040127 (Tue) at 1504:51 -0800, Greg McCann wrote:
->> > On 1/27/2004 at 5:23 PM Greg Louis 
->> > >results with bogofilter.  Should we clutter the code with
bad
->> > >options and then say,"actually this is a bad thing, don't
use it"?

That's the Un*x way, I'd believe. Look at the disclaimers with
iptables-save and iptables-restore, for example. Look at
experimental kernel code. Take the examples Greg McCann gave.
Look anywhere where there's Linux. You find disclaimers "this is
dangerous/useless/way too much, use on your own risk" anywhere,
but, actually, I don't find an example by heart where there are
artificial restrictions on a user in the Linux world just because
the developers don't know what that should be good for.
But feel free to come up with one.

->> I argue, precisely, that anyone who understands what bogotune
does and
->> how it works would not use this option.  With that few
messages, the
->> user is safer to stick with bogofilter's default parameter
values (or
->> with values obtained elsewhere on a reasonably sized corpus).


This actually sounds like the "we know what's good for you"
attitude which drives MS users up the wall....

(I'd think the careful and Linux way would be sth like:
"CAUTION: Your corpus is much to small to provide reasonable
results. Unless you've got at least n spam and ham and x mails
for comparison, you're probably better off with the default
parameters." or something to that effect, but then, doing
whatever I asked for.)

->> Although anyone can make the one-line change that David
suggested, I
->> think it sends the wrong message altogether if we offer it
officially.

It sends the message open source has been sending from the
beginning. It says: "You're responsible".

->A hammer is a tool that works great on nails. You can also use
it to
->pound your finger.  But is that really what you want?

Sometimes, and if only out of curiosity. I was careful already
back then. Since I'm not a baby anymore, I feel nobody should
take everything from my hands just because I might get a scratch.

->Bogotune is also a tool.  Please use it as intended :-)

I can't because my corpus is too small :)

I'd still like to know how it works. I might want to know how
values change while my corpus grows. I might want to play around
with it with different corpuses. I might want to know how it
handles before I'm actually using it, so I'm not in cold water
once my corpus is esteemed big enough. There are probably a
hundred reasons to use it with a too small corpus, and you can't
know them and shouldn't try to anticipate them.

Let me make this clear: I and most people on this list can make
that oneliner hack in no time. I don't lose much if you don't
change it.
I actually believe you should rethink that, not for me or anybody
specific, but for bogofilter, and bogotune, and their place in
the community you put them. 

Bogofilter/bogotune are great tools, and thanks a bunch for them,
btw!

Best regards, jo