Anthony Atkielski <anthony@atkielski.com> writes: > Their "anti-spam" efforts cause them to reject millions of legitimate > e-mails each day, so they create problems, too. So does trying to find the legitimate mail among a pile of spam. > What process can you suggest to distinguish spam from legitimate e-mail > with 100% reliability and no human intervention? Find that, and you can > become a billionnaire. It's reality-check time. We're not going to get that, and the problem is continuing to get worse. Now what? How many people who are active and have been active for some time on the Internet are still putting e-mail addresses on web pages and then reading them with no spam filtering whatsoever, just looking through the inbox and deleting what isn't wanted? How many of you have accidentally discarded the wrong message because it got caught in a purge of spam from your inbox? Personally, I can still cope with the onslaught with no filtering other than a few simple personal rules based on observation of my incoming spam and the types of messages I personally receive, but those rules are utterly draconian from certain perspectives and still vulnerable to false positives (for example, I don't speak any Asian language, so anything I receive in that character set goes straight into the spam pile, likely never to be looked at; if someone sent me a message that contained an excerpt in Chinese, I would likely never see it). I can probably make do with Bayesian filtering for at least another two years. Forever? Unclear. It would be rather nice to think that people were working on some sort of protocol solution at the same time, though. (But another point made is very valid too, namely that someone needs to do protocol work and forming a working group is something that one does after one has some idea of what sort of protocol one is working on. There are a host of plausible-looking ideas for dealing with spam that have been tried over the years that would be a waste of time to standardize because they don't actually work.) -- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>