An Aussie computer scientist has come up with the idea of using Sudoku puzzles to combat spam.
Tricky mathematical puzzles like Sudoku could be the next weapon in the fight against spam, an Australian computer scientist suggests.
Paul Gardner-Stephen from Flinders University in Adelaide is looking for ways to improve existing spam filters, which try to prevent unwanted email from getting to your inbox.
“The trouble is that they’re not entirely accurate, so you still get some spam in your inbox and some real mail gets bounced,” he says.
His idea is to supplement existing filters with an automatic system that requires mail servers to solve a mathematical puzzle if they want their message to reach you.
Mr Gardner-Stephen is due to describe the system at a meeting of the Institute of Electrical and Electronics Engineers this week.
He also publishes his paper on the arXiv website.
First, the mail server computer that receives your email would use an existing spam filter to decide how likely it is that an incoming message is spam.
If it looks suspect, the server would automatically respond to the sending server with a mathematical challenge to solve before the message is delivered.
The more certain your spam filter is that the message is junk, the harder the puzzle would be to solve.
For messages that are clearly spam, the puzzle could be set to take about an hour to solve, Mr Gardner-Stephen says.
“If someone’s trying to send spam, they end up with a lot of puzzles to solve, so they can only send relatively few messages a day,” he said.
In technical terms, the system is known as a ‘proof-of-work’ scheme. By providing a solution to the puzzle, Mr Gardner-Stephen says the sending mail server is proving that the message is important enough for it to do a certain amount of work.
“Sudoku is a good example of the kind of puzzle,” he said.
“It’s easy to verify you’ve got it right, but not easy to do.”
So far, the system only exists in theoretical form, but Mr Gardner-Stephen is planning to try a test on a single mail server within the next year.
If all goes well, it might take 5 to 10 years for a system like this to be adopted across the many thousands of mail servers across the internet.
Wouldn’t it be nice?