Sun, 29 Jan 2006
Spam not
Jonathan A. Zdziarski, Ending Spam - Bayesian Content
Filtering and the Art of Statistical Language Classification,
No Starch Press
This book claims to be of interest to:
- People who want to write spam filters
- People who want to understand spam filters
- Spammers who want to evade spam filters
Its author is the developer of DSPAM, one of the Baysian Content
filtering programs described in the book.
The line the book takes is that spam filters are becoming good
enough that evasion is futile. There's a lot of propaganda for
statistical filtering as a better solution than other approaches
(blacklisting, whitelisting, challenge-response, legal
approaches, etc.) If you haven't seen as many successful ingenious
attempts to evade your spam filters as you want to, there's a
chapter with examples of those.
I have previously been in the category of "people who want to
implement effective spam filters without knowing a lot
about them". There is actually a chapter directed at poeple
like that: "Appendix -- Shining examples of spam filtering". It
lists a half dozen programs that can be downloaded for free and
used to filter email. I enjoyed trying them out, although I
didn't find anything clearly better than what I'd been using
before. But I didn't spend all the days and weeks training that
are recommended for optimum use.
The explanations of how the filters work (and why they sometimes
don't) are well-written. So if you want to know those things,
this is a good book.
My conclusion is that this book is not a magic bullet for fixing
your spam problem, but is an interesting read for those
interested in what programmers are doing about the spam problem.
(Laura Conrad)
[]
permanent link
| 
pre Dec 04 reviews are here