> Date: Mon, 24 Feb 2003 09:02:25 -0500 > From: Shawn McMahon <[EMAIL PROTECTED]> > To: [EMAIL PROTECTED] > Subject: Re: [Razor-users] Turning Razor into a censorship tool > > Razor does not list it as spam. You have a problem understanding what > Razor is. Razor lists the OPINION of a Razor member with a high trust > level, that's all. Other Razor members then choose what to do with that > opinion.
This seems to be a common response. I think that you are missing the fact that Razor is not just a neutral piece of software doing it's assigned task. Razor was designed as a SYSTEM of determining spam-or-not based on the assumption of a large user base some of whom unintentionally or maliciously provide corrupt input. You can't just say "GIGO" when the stated intent of your system is to tolerate GI. I do agree with you on the last point; the true responsibility for losing the EFF mailings lies with the system administrators who are using a beta-level anti-spam system in a production environment. But that does not mean that it's inappropriate to say "hey, Razor got a false positive; how can we reduce the frequency of such events?" Here are some ideas I've thought of while following this incident: Require a user to build and maintain a high trust score for some period of time before their input is used to determine spam-or-not or affect others' trust scores. Come up with a procmail or other method wherein razor is reapplied to mailboxes periodically to either recover freshly-revoked false positives or weed out newly-reported spam, and include that in the docs as a suggested implementation instead of or in addition to the current check-on-receipt method. -- [EMAIL PROTECTED] John Stimson http://www.idsfa.net/~john/ HMC Physics '94 ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Razor-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/razor-users
