On Sunday, October 9, 2005, 5:39:33 PM, Jeff Chan wrote:
The quality of SpamCop reports is dependent on the quality of what users report, plus internal processing SpamCop does to reduce the noise and errors in those reports.
It's perhaps worth noting that with some munging on our side, the SpamCop spamvertised site data result in some of the better performing SURBLs: sc.surbl.org and ab.surbl.org.
We can compensate for the occasional errors made by SpamCop users, yet it's the fact that the reports are mostly manual that make the data unusually good. Generally speaking, some humans had to take some time and make some effort to report the spam, and the reports are mostly right. Collectively, their judgements about what constitutes spam turns out to be quite useful.
Jeff C. -- Don't harm innocent bystanders.