Here's the thing, I just found out, via "NOVA" on PBS that Re-Captcha essentially gets ignorant users to correct OCR (Optical Character Recognition) errors FOR FREE whenever they register for a forum or even post to website at this point.
Since I have to pay for optical character recognition software myself and get no freebies I find it offensive that some computer scientist from Carnegie Mellon gets paid for the OCR error correction that he doesn't do while simultaneously getting paid for writing the rip off software to get free error correction for probably tens of thousands of books if not more from unsuspecting computer users. I just have serious problems with people being slimy.
My suggestion is to substitute "NOFREEBIES" for the OCR correction word in ReCaptcha verifiers. Software has no idea what the erroneous OCR word is to begin with so you could substitute anything in there. As long as you get the non-OCR word correct it assumes your OCR correction is right and puts it into the book database. The distinction between the two words is that the actual verifier must be optically distorted (warped, twisted, squeezed etc.) The OCR word typically is just a bad looking photocopy.
My endgame for this is to either have ReCaptcha discontinued and go back to the the original Captcha (which only includes optically distorted text for verification.) OR Force the makers of ReCaptcha to pay individual users $1 per error correction made. OR Force the makers of ReCaptcha to freely distribute a robust, multiplatform network enabled OCR application for individual users that also feeds into the ReCaptcha system.
Since I have to pay for optical character recognition software myself and get no freebies I find it offensive that some computer scientist from Carnegie Mellon gets paid for the OCR error correction that he doesn't do while simultaneously getting paid for writing the rip off software to get free error correction for probably tens of thousands of books if not more from unsuspecting computer users. I just have serious problems with people being slimy.
My suggestion is to substitute "NOFREEBIES" for the OCR correction word in ReCaptcha verifiers. Software has no idea what the erroneous OCR word is to begin with so you could substitute anything in there. As long as you get the non-OCR word correct it assumes your OCR correction is right and puts it into the book database. The distinction between the two words is that the actual verifier must be optically distorted (warped, twisted, squeezed etc.) The OCR word typically is just a bad looking photocopy.
My endgame for this is to either have ReCaptcha discontinued and go back to the the original Captcha (which only includes optically distorted text for verification.) OR Force the makers of ReCaptcha to pay individual users $1 per error correction made. OR Force the makers of ReCaptcha to freely distribute a robust, multiplatform network enabled OCR application for individual users that also feeds into the ReCaptcha system.
Last edited:
