review by
Roger Walker, July 2005

This book is directed towards anyone who may have a passion for (or against!) spam. That includes geeks and hackers (possibly intermediate to advanced spam filter writers), and spammers themselves - who are invited to see why the author thinks that statistical analysis will eventually beat them.

I am not sure the author understands his own bias which comes across in the book. However, I should state my own bias: Statistical analysis requires receiving and processing every inbound message in order to work optimally, which means wastage of bandwidth and processing resources. It is also predictive, meaning that its opinion about whether a message is ham or spam is based on what it has seen before, and there will always be room for false positives and false negatives. These, to me, are unacceptable consequences, which is why I do not use statistical analysis with EVS Mail.

That aside, know that the author is also the author/maintainer of DSPAM, a freely available statistical analysis package that is considered one of the best, and which can integrate into a number of MTAs. He has compiled a lot of information for the book from other filter writers and his own research.

The book starts with the requisite introduction and spamming history lesson. It then introduces fundamental spam filtering techniques and builds upon the information in describing statistical analysis techniques, themselves, including storage and scaling issues. Then the author goes more indepth into advanced concepts. It finishes off with interviews with authors of various statistical analysis spam filters.

My only complaint with the content is that some topics seem to have more or less content than others, and I am not sure why that is. My only complaint with the book itself is the editing. Unfortunately, finding spelling/syntax/logic errors just takes away from my reading pleasure.

Despite the minor issues I have with how the book was written (and my obvious bias on the subject matter), I highly recommend the book for anyone with an interest in spam and spam control. It has a permanent place on my bookshelf, because I find it inspirational - it tells me that I'm doing something right.