compile time options

[David Relson]

Greg,

As our algorithm expert, developer, and tester this might be a good time
to query the user base to see if anybody is still using Graham  or
Robinson-GM, or if people are all using Robinson-Fisher.  With the right
answer, we'd be able to retire some code, shrink the size of bogofilter,
and simplify life for the development team.

David

On Mon, 29 Sep 2003 16:48:25 -0400
Greg Louis <glouis at dynamicro.on.ca> wrote:

> On 20030929 (Mon) at 0757:05 -0400, David Relson wrote:
> > On Mon, 29 Sep 2003 10:33:13 +0200
> > Since such questions haven't been asked frequently, the FAQ
> > doesn't need much info on the subject.  Perhaps a couple of simple
> > questions like:
> > 
> > What different features can be compiled into bogofilter or left out?
> > Can I compile bogofilter without support for all the scoring
> > algorithms?
> 
> Why are so many algorithms supported anyway? is a FAQ.  I think I know
> the answer but I don't think we explain to newbies too well.

True.  Since the older code is there for historical reasons and backward
compatibility, I'm inclined to leave it alone and not promote it.  With
a willing user base, I'd be glad to delete it.


> While I'm on the subject of stuff I think obsolete:
> 
> --- robinson.c~ 2003-09-29 16:45:18.000000000 -0400
> +++ robinson.c  2003-09-29 16:45:18.000000000 -0400
> @@ -27,11 +27,6 @@
>  #define ROBINSON_MIN_DEV       0.0     /* if nonzero, use
> characteristic words */
>  #define ROBINSON_SPAM_CUTOFF   0.54    /* if it's spammier than
> this... */
>  
> -/* 11/17 - Greg Louis recommends the following values:
> -** #define ROBINSON_SPAM_CUTOFF        0.582
> -** #define ROBINSON_MIN_DEV    0.1
> -*/
> -
>  #define ROBINSON_MAX_REPEATS   1       /* cap on word frequency per
> message */
>    
>  extern int Rtable;
> 
> I stopped recommending the use of the GM about 8 months ago IIRC.

I'm certainly willing to remove the 11/17 comment.  I'd be surprised if
anyone cared.