> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Behalf Of Santiago > Vila
... > The details are a secret, but apparently the following general > rules apply: ... Messages that are not reported as spam are not reported at all, right? In other words, there is no "this is NOT spam" reporting outside of revocation? It seems like it would relatively simple for Razor to do a sort of double-check on any reported signature to see how many users did not consider the related e-mail to be spam. It could even do something like the Listserv anti-spam system, which looks for the same message being submitted to multiple lists and automatically temporarily blocks that sender, with a notification to the list admins. By having Razor clients each automatically submit a small number of "good" signatures, the servers could probably quickly identify bulk e-mail. I do agree with the Razor philosophy that one person's "good" mail may be another person's spam. But the natural first-order followup is to discover what the proportions are. A second-order follow-up would be to look for patterns in the viewpoints demonstrated by that behavior, so that Razor is not just learning whom to trust to identify spam, but whom to trust to identify non-spam. And I keep coming back to the problem of latency. Spam detection has to happen fast; identifying non-spam does not, which gives it a great advantage. I'm a bit disappointed with those who would just accept the system as it is, when it clearly has plenty of room for improvement. Reputation systems in general have a great deal to offer, but they are far from fully exploited. I'd expect the developers to agree, since they've created a company around it! Nick ------------------------------------------------------- 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
