Turing tests and spam

sickstan · 8th March 2004, 10:15 AM

I'm trying to figure out how these Turing tests would be used for filtering email.

Right now, whenever you want to send email via Outlook Express, you would compose a letter. This letter would be send via your machine directly to the SMTP server your ISP provides, then it would be sent through the network, finding its way onto the POP3 server, which would hold it for the receiver either to access directly using Telnet (or a web-based email reader) or download onto his/her computer for reading.

Where would the request for a Turing test occur? From the receiver's POP3 server? What form would the test take, an email back to the sender? Would the answer then be confirmed by the POP3 server before greenlighting the original email?

I guess the same could be applied to the SMTP server, especially for large ISPs, but this COULD be circumvented by spammers exploiting smaller ISPs that don't utilize these tests.

Help me out, you geeks.

BTW, Bill Gates is evil.

phildonnia · 8th March 2004, 12:49 PM

Quote:

Originally posted by sickstan

Help me out, you geeks.

You may have noticed that Spam now comes with subject lines and sender names that look more personable and friendly. These are foiling the typical dumb spam filter but they are also succeeding more often in foiling my own mental "filter"

When I can't determine any more whether something is spam anymore, I don't have much hope for semi-intelligent machines to find the difference.

Spam is like campaign finance money; it will inevitably find all the holes in the system.

8th March 2004, 10:15 AM	#1
sickstan Thinker Join Date: Jul 2002 Location: Phoenix Posts: 179	Turing tests and spam I'm trying to figure out how these Turing tests would be used for filtering email. Right now, whenever you want to send email via Outlook Express, you would compose a letter. This letter would be send via your machine directly to the SMTP server your ISP provides, then it would be sent through the network, finding its way onto the POP3 server, which would hold it for the receiver either to access directly using Telnet (or a web-based email reader) or download onto his/her computer for reading. Where would the request for a Turing test occur? From the receiver's POP3 server? What form would the test take, an email back to the sender? Would the answer then be confirmed by the POP3 server before greenlighting the original email? I guess the same could be applied to the SMTP server, especially for large ISPs, but this COULD be circumvented by spammers exploiting smaller ISPs that don't utilize these tests. Help me out, you geeks. BTW, Bill Gates is evil.
	__________________ "I'm honored to shake the hand of a brave Iraqi citizen who had his hand cut off by Saddam Hussein."—Washington, D.C., May 25, 2004

8th March 2004, 12:49 PM	#2
phildonnia Master Poster Join Date: Oct 2001 Location: Sac'to CA Posts: 2,339	Re: Turing tests and spam Quote: Originally posted by sickstan Help me out, you geeks. You may have noticed that Spam now comes with subject lines and sender names that look more personable and friendly. These are foiling the typical dumb spam filter but they are also succeeding more often in foiling my own mental "filter" When I can't determine any more whether something is spam anymore, I don't have much hope for semi-intelligent machines to find the difference. Spam is like campaign finance money; it will inevitably find all the holes in the system.