On Thu, 06 Aug 2015 22:35:54 +0200 Benny Pedersen wrote: > RW skrev den 2015-08-06 21:03: > > > "WOT also uses information from numerous > > trusted sources, such as phishing and malware blacklists, to > > provide the WOT community with real time information." > > > > It look like the part of WOT that might be useful for mail filtering > > comes from third parties. > > api is not defined yet, so a bit unfair
The object of WOT is to protect users at the time a link is clicked, not when the link is mailed, posted etc. This means it can continue to accumulate useful evidence long after a spam run is complete. I think it's pretty unlikely that it adds much to spam filtering in the short time before a URI starts to appear in URIBLs. Any attempt to evaluate WOT should take careful account of what URIBLs it might be using. > its not like it would just be if wot knows this domain its > whitelisted ... It's not feasible to take account of good reputations in the scoring, URIs in a spam are under the control of the spammer. Even if spammers didn't actively try to abuse it, it can still fail. For example, even the most spam-friendly email provider will have a good WOP reputation for their website. SA may find that domain in a default signature or on the RHS of an email address in the body. Good reputations wouldn't even be useful to mitigate URIBL FPs, unless the two results are known to be for the same domain - and that would require substantial change to URIBL handling.
