[linux-elitists] (tmda) Re: Constraining Bogus challenges.

Andrew akohlsmith-le@benshaw.com
Wed Sep 24 06:39:23 PDT 2003


> SpamAssassin may work fine for the class of problems it was originally
> intended to solve. Yet once one has to run spamassassin as a daemon to
> make it somewhat more bearable, it's clear that it's poorly designed.

I don't know -- I can't think of any other spam filter that has the width 
and breadth of tests that SA does, the configurability to add new tests, 
optionally use network tests (blech, I have them all turned off) and 
per-user configurability that SA does.

I agree -- something that was compiled might be faster.  And I am sure that 
in time SA's algorithms can be optimized, but it comes down to the simple 
rule that you don't optimize what you don't have to.  SA runs remarkably 
well for pretty much anyone, and small tweaks like my "don't try to filter 
files > 75kb" seem to make it work very fast.

As Karsten said, hardware's cheap.  As you said, memory's cheap.  I'm a 
minimalist.  I cringe every time I need a system to do something that a P2 
or P1 can handle easily yet all I can buy are P4s.  (Yes I need the 
reliability, replacability and some of the features of (fairly) new 
systems.) -- But when the business person in me looks at it, it's cheaper 
in the long run to buy that P4 and let it sit idle 99.999% of the time.

It'd be far more work to make a compiled-language spam filter with the 
configurability and extensibility that SA has, which is probably why 
nobody's done it yet.  SA's "good enough" for practically everyone.

I'm not bashing you technically; you're absolutely right.  I'm just saying 
that it doesn't make sense from other perspectives.

Regards,
Andrew



More information about the linux-elitists mailing list