On Sun, 2009-12-06 at 12:02 -0700, LuKreme wrote: > On 6-Dec-2009, at 02:24, rich...@buzzhost.co.uk wrote: > > A truly clean company that always uses opt-in and never spams has > > nothing to fear from any anti-spam measure. > > Oh, that is CERTAINLY not true. It's not even true of just SpamAssassin, but > it is completely disingenuous to claim that for ANY anti-spam measure. > completely clean messages that are not spam get miss-tagged ALL THE TIME. I > dig mails out of my spam folder that I want at least on a weekly basis. > > Just this week I had someone I know quite well send me an evite invitation to > a meeting next week. It was tagged with Bayes_99, probably due to all the > fake 'marketing seminar' invitations that go out. > > Ironically enough, this one would have still squeaked under the wire had I > not tagged it +2.0 for HABEAS_ACCREDITED_SOI, but that's another thread :) > > Another was an email from a friend of mine that for some odd reason tripped > bayes_99 and a relay check and a couple of others. So there's two this week. > > OTOH, there's 410 more spam messages in that folder from this week that all > appear to be correctly tagged. > That would point more to the quality of the Bayesian db's IMHO, but it was careless to say 'any'. The point to get across is reputable businesses should not need the services of an ESP who attempts to subvert anti-spam systems. Optimising email for delivery is quite one thing, getting into bed with anti-spam vendors and coders to serious weight the deck in the favour of the bulker is not acceptable.
ESP's are to be distrusted. They are mostly the devils advocate and I don't think that is unfair. When you have people like eBay spending money with you, you are not going to say 'sorry, we are dropping you because you occasionally spam millions of your users'. Feedback reminders spring to mind. A link tells you that you can update the preference to stop them, but follow the link and the option is nowhere to be found - but you get to see the latest adds going there.
