On Mon, 17 Nov 2003 13:22:49 -0500 (EST), William Stearns <[EMAIL PROTECTED]> writes:

>       So if I read you correctly, adding 4800 rules essentially triples 
> the cpu time needed to process a given message or collection of messages.


>       Are there ways to improve the performance of the checks?  I ask 
> because these URI rules are tripping on about 50-60% of my current spam - 
> much more than the corresponding source domain blacklist rules.

I had a program that would be able to do this sort of test in about 15
seconds on 50,000 emails with 600 rules. I generated an automata from
the ruleset, then simulated it over the email input. Thus scaling it
up to 5000 rules wouldn't degrade performance. The thing I like about
the design is that if the write-test-debug cycle gets down to ten
seconds, you can almost interactively write rules. You can also write
a program to make guesses for speculative rules and only report if
they do fairly well.

Scott


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to