Mozilla Named Best Anti-Spam Software in German Computer Magazine
Saturday August 9th, 2003
Simon Paquet writes: "In a recent shootout of anti-spam software in the German computer magazine c't (link is in German), Mozilla Mail won the first place against competitors like K9, Apple Mail, Mailshield Desktop, SAproxy, SpamBayes, SpamCatcher, SpamNet, SpamPal and SpamSieve.
"To quote from the magazine article (very rough translation into English): 'If you're using Windows, Mac OS X or Linux — the fast and reliable Junk Mail Controls of the quite function-rich email program Mozilla are unsurpassed. Anyone who at this time is looking for an anti-spam solution should think about switching to Mozilla.'"
#1 ars has a good overview of spam sw
Sunday August 10th, 2003 3:05 AM
Ars has a really solid article on spam on their site: <http://arstechnica.com/as…2003/anti-spam/index.html>. It's an overview of the various spam solutions for windows and covers all major players (including of course mozilla). One of the conclusions is that Bayesian Filtering is the most effective method.
Mozilla never got a first place in the c't. The c't doesn't have places. Correct is that they tested it with a very good result.
The Online Article about that (also in german) : <http://www.heise.de/newst…ker/data/em-09.08.03-002/>
The article you linked to is actually only a "teaser" for the full article in this month's c't - I think I'll have to go out and buy it.
I have to strongly disagree with the article. I use both Mozilla and the SpamBayes plugin for Outlook. SpamBayes wins hands down. SpamBayes is absolutely infallible. Mozilla definetely is not. SpamBayes has improved the original Paul Graham algorithm a lot, they have done a lot of testing to find the best possible algorithm. Mozilla still uses the original algorithm. That's why I voted for Bug 181534. Unfourtunately I can only vote since I'm no good at C++ and XPCOM (so complicated) and SpamBayes wrote their stuff in Python (simple). But someone who knows both should be able to easily port it, I think.
The SpamNet/Cloudmark "collaborative filtering" approach is superior. All of the Bayesian solutions I've tried miss way too often both in letting spam slip through and much worse, blocking legit emails.
SpamNet may be great for those using Outlook Express so how about those of us using Mozilla