0.7.6? how about 0.8.0?
David Relson
relson at osagesoftware.com
Sat Oct 26 19:41:52 CEST 2002
At 07:41 AM 10/26/02, Boris 'pi' Piwinger wrote:
>Allyn Fratkin <allyn at fratkin.com> wrote:
>
> >the addition of the robinson algorithm is a pretty major change.
> >is it time for 0.8.0?
>
>I think with those two algorithms we should see which works
>better (more exact bogosity calculation and performance).
>Then only include that one into 0.8 to keep it clean and
>simple.
pi,
I'm inclined to agree with you, but only part way.
For 0.8, I want to keep both Graham and Robinson available. We'll change
the default to Robinson, if enought evidence supports the decision. The
configure script needs to be updated to allow configuration of bogofilter
for either algorithm or for both. If only one algorithm is selected, that
should be the only one in the binary.
> >as an aside, how do we know when we get to 1.0?
>
>0.9 should be able to do some decoding of text/* parts
>(base64 and QP). 1.0 should do Unicode internally as
>discussed recently.
Adding that decode capability is definitely worth a 0.1 increase in version
number. I don't think 1.0 needs to have that decoding. We'll add it when
we get it.
Another feature worthy of a 0.1 increase is the ability to run bogofilter
as a daemon (with memeory mapped wordlists). That should help performance
significantly for large sites and for people running bogofilter with large
test data sets. Again, I won't say "daemon" is the same as 1.0. We'll add
it in when we have it.
Like you say, "just my $.02"
David
More information about the bogofilter-dev
mailing list