Recently I had reason to register on a work-related web forum, so I went through the usual process of entering a handle, give my email address and then I got a CAPTCHA to ascertain that I was a real human. It gave me some pause, as half the characters I was required to enter were in Greek. Had I been on a Mac, it wouldn't have been much of a problem, but on my work Linux box I had no idea how to input the required text. Fortunately requesting a new challenge generated an ISO Latin-1-only string and I could proceed.

I wonder about the implementation of the CAPTCHA—did it really generate strings of random Unicode characters? Probably not, since the first half of the string was fully in Greek, and the latter fully in Latin characters, rather than a jumble of Chinese, Indic and Runic characters. So, if there was a restriction to certain character sets, why combine two different alphabets? Or was the implementation from a distant perspective of “Since these are all European characters, they can be generated on a European keyboard, right?” Or possibly just somebody being deliberately difficult. I know the kind, I am one of them…

No comments: