Interestingly enough, while during my BS CS undergrad stint (1999-2002), I learned both Ada95 and MUMPS (along with Java, Scheme, Lisp of course)...
Kevin O'Kane was teaching MUMPS because (at the time) he had one of the only MUMPS compilers out there (translation to C++ and then compiled down)... Did a lot of work in textual indexing and searching (both medical data and genomics).
So a few things... the video through the audio jack, probably not at the rate that the video is displaying...
Notice also in the begininng demonstrations that he has to get the device super close to the screen in order for it "to work"... and when he moves it less than a foot away, it stops.
Now.. when he ties it to the balloon, it would require that the balloon would happen to fly within less than a foot of the screen. Errr..
So, the next clue is the muxing couldn't be accomplished without having input to how many screens there are. For example, he holds it up to the video screen with 4 screens... How would the device know to map the video across four screens? It would have to be doing 4x the work as well.
Met this guy sitting outside of Facebook a few months ago. Nice kid, smart, savvy. We chatted about Greplin and it sounded like he was making good progress... Let's hope he keeps it going...
Kevin O'Kane was teaching MUMPS because (at the time) he had one of the only MUMPS compilers out there (translation to C++ and then compiled down)... Did a lot of work in textual indexing and searching (both medical data and genomics).
http://math-cs.cns.uni.edu/~okane/mumps.html
I really miss the B-Tree arrays...