Brad IdeasCrazy ideas, inventions, essays and links from Brad Templeton |
|
|
|
NavigationUser loginIf you like this blog, do me a favour and start your Amazon shopping (especially a kindle) from this link, and I'll get a cut. Recent comments
Top EssaysRecent blog posts
BlogrollFellow EFF Folks
Cory Doctorow Larry Lessig Ed Felten Dave Farber John Perry Barlow EFF Deep Links Dave Sifry |
I know people disagree with c/r
But autoresponding isn’t going away, it’s too useful in bounces after mx, for mailing list confirmation, confirmation of email addresses entered on web pages etc.
The answer is not blacklisting autoresponders, but working to fix the autoresponse to forgery problem. Autoresponses normally don’t multiply spam like open relays, they just reflect it. That’s not good for the person it is reflected at, of course. But the autoresponse does not advertise the product so the spammer is not interested in it. He’s just putting in a forged From to get past whitelists and detectors of invalid From lines.
For autoresponses to emails, we may need to move to a regime where those who want autoresponses sign their mail. However, long before that we could move to some standization in automated responses, so that it’s easy to detect autoresponses to messages you never sent out, and be rid of them.
C/R is worth protecting because it is the only system that can turn an anti-spam filter into a no-false-positive filter. The correct approach is to discard spam you are sure is spam, pass through what you are 99.9% sure is ham, and challenge the small quantity of stuff you can’t figure out.
Some people say, “I don’t like challenges” but if you pose the question, “would you rather have a challenge, or would you rather your mail was discarded or put into spam folder that may or may not get looked at?” — the answer is different. I sent the mail for a reason, and I want a chance to override the spam filter if it decides not to deliver it. The message might be important.
One could also consider a flag to say “Don’t bother challenging me” for those who don’t want a challenge if their mail is about to be not delivered.
Non-delivery is a serious failure of the mail system. It must not go unreported. Some would argue it should be delivered to both parties. It could make sense for the sender to decide who to deliver it to, though you can’t easily stop the recipient from superseding that.
If the algorithms can’t figure out whether to deliver, only a person can.