ACM TechNews

Web-Security Inventor Charts a Squigglier Course

Wall Street Journal (08/13/08) P. B5; Smith, Ethan

Carnegie Mellon University professor Luis von Ahn, the primary inventor of the Captcha online security test, has updated the system to make the test more secure. The new ReCaptcha system would also have users assist in the digitalization of old books and newspapers. The new system presents users with two words containing distorted characters. Both words are taken from an old book or newspaper article that has been scanned into an online library. One of the words was recognized by the scanning software, while the other was unrecognizable to the computer, possibly because of a smudge or some other imperfection on the original document. The user tries to decipher the distorted characters of both words, and if the user matches the first word correctly, which the computer already knows, then the user's reading of the unknown word is recorded. Multiple Web users will be shown the same unknown word as part of different tests, and when three people have submitted the same answer for the unknown word, it is considered solved and added to the library database to be inserted into the digital version of the document. Deciphering these unknown words is one of the greatest challenges for the Internet Archive library digitalization effort, since scanning software generally has an accuracy rating of only about 80 percent for books published before 1900. About 40,000 Web sites now use the free ReCaptcha system, and when fully operational, von Ahn expects it to process about 160 books a day for the Internet Archive. "It's a really mind-blowing application," says Internet Archive founder Brewster Kahle.

http://online.wsj.com/article/SB121857740590934637.html

---

© Copyright 2008 Information, Inc. This service may be reproduced for internal distribution.