There is a huge amount of Arabic data that is not easily accessible or searchable (old books, government archives, etc.). We can scan these documents and make them available as images. But that is not useful for search, automatic translation or any language processing application.
The current quality of optical character recognition (OCR) for Arabic is rather low. The solution we propose is to chop the text images into little snippets and send them through our app to users who can read and type Arabic on their mobile phones. The users collect points which can be donated or cashed.