Thursday, December 18, 2008

Have We Solved Spam?

The amount of spam in my spam folder has dropped below fifty per day down from several hundred a few months ago. The amount of spam that gets through Gmail's filters to my inbox is well under one a day. I'm not sure why. Better spam filters, better enforcement, less economic incentives for spam, maybe even the bad economy is playing a role. While not completely solved, email spam is no longer a significant problem for me.

This has happened without you having to pay me, solve a CAPTCHA or have your computer solve some hard problem to send me email. I have a mailto link on my home page (albeit generated by javascript code) and my various email addresses show up in plaintext on hundreds of pages scattered over the Internet.

For those of you who insist on making it painful to send you email (anything beyond a simple click or cut-and-paste without further editing), time to change your ways and not make your fear of spam make extra work for those of us who simply want to say hello, and maybe ask you to review a paper.


  1. Are you saying that obfuscating one's email address is a good method to avoid writing referee reports? ;-)

  2. Gmail has a huge number of false negatives though.

  3. Unfortunately, not at all. There is a drop since November 13, 2008 (you can see the reason here and here ).
    But you know what? Spam will be over on the day when you get 0 mails in your spam folder. And I don't think Gmail's filters, as smart as they can be (and they really do a great job) are the solution. The solution is something much bigger (and unfortunately seemingly distant), like adding authenticity to SMTP or finding ways to stop the spam bots to the invisible minimum.

  4. Anonymous, you mean false positives, right? (legit email going to the spam folder)

    Spam is still a huge, expensive problem even if most end users, like Lance, are largely shielded from the pain.

    And I'm not sure Lance is quite right. For one thing there's the unsettling tradeoff being made between less spam in the inbox and greater risk of false positives. False positives are often never detected, another way in which the true cost of spam is hidden.

  5. False positives are often never detected, another way in which the true cost of spam is hidden.

    Can't you reduce false positives by asking the sender to solve a CAPTCHA whenever his/her email is classified as spam?

  6. Yeah, I get virtually no spam. Less than one a day now.

    I agree about people making it easy to send them email. I have my email address on my web page with an "a href=mailto:" link to make it as easy as possible for people to contact me. The idea that spammers collect e-mail addresses by crawling the web and pattern-matching has seemed pretty dubious to me for a while now. Is there any evidence that they do that? There are easier ways to get lists of e-mail addresses. And even if they do, most of the basic workarounds you see are easily pattern-matched themselves, so what's the point?