Typophile

Those ooogly typographic CAPTCHAS

kosal

You know those distorted abominations of letterforms you have to decipher when you register for sites? I’m surprised I couldn’t find much interest for them within the typographic community. Has anyone entertained the thought of creating the CAPTCHA puzzle with found letterforms? I think it would be quite beautiful. A tag search for “oneletter” on flickr will reveal the amount of variety that could be used to fool the bots and still make it easy on the eye. I don’t know much about the technical capabilities, but my guess is that with enough creativity in photography by a dedicated public, we can defeat technology. Then again, I should never underestimate those like Google.

‹ Font embedding in Adobe (Macromedia) Director Helvetica in the news ›
chrisgaul

No doubt.

You might also be interested in the reCaptcha project that works as a captcha system, but simultaneously uses successful decodings by humans to digitise books for online use.

Kevin Larson

I gave a talk at TypeCon 2007 about the work that I was involved in where we investigated how to make CAPTCHAs as easy to read as possible for humans and as difficult to read as possible for computer programs. It’s a difficult challenge because the computer programs can be right less than 5% of the time to make money for the spammers, while we want people to able to recognize them close to 100% of the time. We’ve written a few papers about our work:

http://research.microsoft.com/~kumarc/pubs/chellapilla_chi05.pdf
http://research.microsoft.com/~kumarc/pubs/chellapilla_hip05.pdf
http://research.microsoft.com/~kumarc/pubs/chellapilla_ceas05.pdf

Cheers, Kevin

Ricardo Cordoba

Read about reCAPTCHA here... Then read about Luis von Ahn, the man who came up with the idea, here.

Captured from web.archive.org on 8 Dec 2008. view this capture