https://issues.apache.org/SpamAssassin/show_bug.cgi?id=5864
--- Comment #11 from Matt Kettler <[EMAIL PROTECTED]> 2008-03-29 16:03:23 PST --- Disabling the URIBL plugin, or skipping blacklisted domains would be detrimental because it would reduce spamassassin's accuracy. It would improve performance, but that's not a good tradeoff given the reduction of accuracy. However, the change you are pushing for both reduces accuracy and performance. That's counterproductive. Why would you want SpamAssassin to be both slower and less accurate? Or do you just not understand that this will adversely affect accuracy, and are under the misconception that it will help accuracy? It won't help SA detect spam accuratly. It will actually make it less accurate. Really. Because URIBL_GREY is improperly scored, this is a necessary result of increasing it's hit rate. Given the score and the S/O ratio, there is no benefit whatsoever to this change. You have failed to ever identify any aspect of SpamAssassin that will be improved. It will increase the false positive rate more than it reduces the false negative rate. That's reduced accuracy. That is a very bad thing. It will also increase the number of DNS lookups SpamAssassin performs, which will increase the average scan time per message. That's reduced performance. I've cited two key metrics for the spamassassin system that will be damaged by this change. Accuracy and Performance will both suffer. Those are *very* important metrics to the project, and the only ones I can think of that are more important are Stability and Security, and neither of those will be affected by this change at all. In the end, what is the benefit? Can you conclusively state any benefit to the SpamAssassin system that would result from removing these two skips? How does that benefit outweigh the reduced accuracy and performance problems it introduces? -- Configure bugmail: https://issues.apache.org/SpamAssassin/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug.
