Everyone should know that when solving a reCAPTCHA, you only have to get one of the words right.
The key is knowing which. One of them is the control word (which reCAPTCHA knows what it is and challenges you to identify it) and the other is a random word chosen from scanned books which you are supposed to OCR for them (they use us as free human labor).
In the current version, the control word is always the one which looks sharper and always made of slanted a to z characters, disposed in a wave-like fashion. If you write this correctly, the other one does not matter. I always write my favorite swear word random gibberish instead.
If the only price we pay for their “free service” is to digitize 1 word for them, big deal. I think the service they provide is well worth it. Have you ever had an open form on a website (i.e. one not protected by captcha or something similar)? You get spam. If we poison their data harvesting that only hurts the availability of this free service. I like Google books.
I did just see some Greek pop up in that second word though, so it’s good to know that I don’t have to get that right.
I agree that this is a useful service and it serves a noble cause. The problem is that the unknown words served by reCAPTCHA have become increasingly more illegible every year. About 1 in 5 cases I get a word I’m simply not sure what it is. And I recall many times when I was shown math equations, words in Hebrew or French (think accents) or simply English words upside down. When average people see these, they just give up what they were doing. Most people don’t even realize there is a refresh image button (sometimes hidden). They need to be more selective in what they show.
Have you heard of Asirra? http://research.microsoft.com/en-us/projects/asirra/security.aspx
I knew of one word thing. I had read that the other word is used for some research and experiments. Did not know of book scanning.
Interestingly enough, I think we could analyze this using the prisoner’s dilemma concept from game theory. Of course, if you are a spam bot you don’t care about digitizing books.
If you are a human, however, you have 2 options here: try to get the other word right or just write gibberish. Getting the word right gets more books digitized. Writing gibberish might be more convenient/faster for each individual. So it’s like the prisoner’s dilemma in the sense that doing whatever is better for you does not give the best outcome for the group.
This is also remarkably similar to voting/not voting in the sense that, for each person, deciding to vote or not vote does not have any impact whatsoever on the outcome of the election, but if everybody decided not to vote, the system would collapse.