gregbo - 6:02 am on Oct 25, 2006 (gmt 0)

IMHO, it's entirely *pro*active for G to *in advance* discover the IP addresses of click networks so that G can presumably ignore clicks from those networks if/when they happen without having to try and guess if they were "bad" or not.

A problem with this approach is that IP networks are not fixed usage entities. If someone knows of a click network and screens it, then the fraudsters give up and/or move somewhere else, one could very well be blocking legitimate traffic. Another problem is that IP networks can change hands, so a compromised network may no longer be compromised once it's transferred to a new organization.