We have all experienced them, those irritating unreadable characters that we can barely see, let alone read.
A problem I met last night, though that was only part of the aggrevation of trying to buy tickets on-line from the British Museum, the main problem was an incredibly badly designed user-unfriendly website. A friend had tried the previous night to buy tickets and eventually gave up.
A computer selects characters, presents them to us in an image which we then have to type in, if there is a match, we are ok. In theory, these images are difficult for a computer to read, but as pattern recognition gets better, the task for us mere humans gets harder and harder as the characters embedded in the images become more and more difficult to read.
The task takes on average ten seconds. Only a few seconds if you can read the first image shown, but as you keep cycling through the images to find one that can be read, the time taken gets progressively longer.
Can this time be put to productive use and if nothing else, partly offset our frustration?
The answer is yes.
You will have noticed now we get two words.
Many books are being scanned, digitised. The older the book, the more difficult the task with an error rate running at 30%.
Humans do better, but very costly, even if outsourced to sweat shops.
Why not use those ten seconds whilst we are trying to determine the illegible characters?
We now are displayed two words. One is as before, a computer generated word, the second one of the digitised words from a scanned book or document. If we get the computer generated word correct, and we cannot get further if we get it wrong, means there is a high probability we got the scanned word correct. But the final word is not taken from one person, several people will have ‘voted’ on the word.
Can this crowd sourcing be used in other areas?
Yes, a project called duolingo has been set up on the net for language translation.
If people were paid, even in offshore sweatshops, it would be very expensive. Why not get people to do it whilst they are learning a language?
Why would anyone do it for free?
They already do it in other areas, freely give up their time to benefit others, Open Source Source Software, this blog …
But they are not doing it for free. They are doing it whilst learning a language for free.
As with digitised words, a ‘vote’ over many translators. http://keithpp.wordpress.com/