A paper trail is a must. The last two elections I voted using this system. There was no tangible record of my vote. Only that I had "participated" via the log of who in my precinct voted that day. Your proposed system would be a vast improvement over what is currently the case. Hopefully we won't conveniently "vote away" many rights before a verifiable system is in place.
This is NOT nuclear physics -- hell's bells, any bright college freshman could design a lot of assorted systems very well, postulating only that he or she had some amount of experience in/with such systems. From a data standpoint (and from a systems standpoint, too), the struc and proc of a competently designed balloting system is absolutely child's play compared to, say, six-sigma production quality control or multivariant real-time sims or any number of other apps.
Why in the world the ''problem'' (sic) of effectively tamper-proof e-balloting has not been quite thoroughly solved is a complete mystery to me. Now, I understand perfectly well why, once solved, such a solution might not be implemented; that's easy. Some politicians have a strong vested interest in NOT having a tamper-proof system...but why (apparently) IS there no implementable solution yet? Baffling.
FReegards, and Happy Hols!