FAVORITES
Popular articles from recent issues

After Goffeney took several rare works home to check the accuracy of the scanned electronic copies, he returned them to the Denver Public Library with a little something extra inside—bedbugs, to be specific.

Goffeney was promptly banned from the library, which requested that he bag the remaining books and leave them outside to be collected. But whether Goffeney was indignant, in denial, or just stubborn, he persisted in depositing the books in the library’s book drop, which library representatives say led to a second infestation. All told, 31 books had to be destroyed, and four areas of the library were fumigated.

Sound familiar? Probably. But did you know that every time you enter one of these, you may be preserving cultural history by digitizing old volumes of The New York Times? If the sites you’ve been joining use reCAPTCHA (Craigslist, Facebook, TicketMaster, and Twitter, to name a few), then you can feel proud that you’ve helped to create a digital record of twenty years’ worth of the largest metropolitan newspaper in the U.S.

Of course, most of the digitization of these volumes is done electronically. Content is scanned and then put through an optical character recognition (OCR) process. But as anyone familiar with OCR knows, the automated software isn’t perfect; it can’t understand every word, especially when source documents are not in great shape.

Enter reCAPTCHA, developed by Carnegie Mellon University and acquired by Google in September 2009. reCAPTCHA works by cleverly co-opting the eyes of millions of (unsuspecting) online users by taking the words that cannot be understood by OCR software and turning them into CAPTCHA puzzles. Online users obediently enter the word they see in the CAPTCHA box, et voila: one more small piece of The New York Times is preserved for digital eternity.

You may be wondering: if the point of CAPTCHA is for the user to match a displayed word in order to prove that they’re human, then how can reCAPTCHA know if the user entered the right answer? According to their website, reCAPTCHA cleverly provides the user with not one, but two words; if the user enters the first one correctly, the systems assumes that the unknown word has been entered correctly as well. If enough users’ responses match up, the word—and another piece of the puzzle—is solved.
Click here to learn more

CONFERENCE BUZZ

Overheard at the S1000D Conference
S1000D User Forum Trade Show, Hilton Head, SC, October 12-15, 2009

The attendees at October's S1000D Conference in Hilton Head, South Carolina represented two main groups: military users (both domestic and international) and commercial aviation companies. Military users were in various stages of implementation regarding the S1000D specification; some had implemented the specification for part or all of their data, while others were still evaluating it.