The prize is measly but a fascinating problem, no?
My initial idea is to dump all the material into a large room-sized box. Fans would then blow the clippings into the air (like lottery balls or something) and stop to allow it to settle.
Next a high res camera scans the jumbled scene closely from above (using an X-Y plotter). Upload the images onto a server.
Repeat many times.
Then process the data. First, images of individual clippings are extracted from the raw images.
Finally, images of the original paper documents are assembled by statistical algorithms similar to those used by Craig Venter to map and sequence the human genome.
It seems to me that by cooperating with schools and introducing assessment, Khan risks losing his way. The educational value will not be improved by becoming more schoolish -- schools are the problem.
As hackers know and as Mitra showed, learning is fun. However it ceases to be fun when one is told what to study and pressured to compete for meaningless grades/scores/badges.
(IIRC, Mitra doesn't assess children directly in the sense of giving them grades.)
Also, now that Khan knows that children will be assessed on his video material, I expect that this will skew his future presentations in a bad way.
My initial idea is to dump all the material into a large room-sized box. Fans would then blow the clippings into the air (like lottery balls or something) and stop to allow it to settle.
Next a high res camera scans the jumbled scene closely from above (using an X-Y plotter). Upload the images onto a server.
Repeat many times.
Then process the data. First, images of individual clippings are extracted from the raw images.
Finally, images of the original paper documents are assembled by statistical algorithms similar to those used by Craig Venter to map and sequence the human genome.