If the source code and compiler are public, the code is placed on an easily-inspectable memory cartridge which can be hardware write-protected, and the code is run in an inspectable computer which is physically incapable of running code from any other source, where's the "black box" aspect?
So you think hundreds of thousands of such voting machines will be inspectable? You think a chain of custody could not easily be subverted?
We are not talking about procedures for safe transit and storage of nuclear weapons by highly screened military personnel.
We are talking about machines passed around the public domain among persons who often have barely a high school education and who do this infrequently, and where procedures are locally controlled and subject to last minute local policy changes by persons unfamiliar with anything computerized.
That’s the black box aspect of it.
The problem discussed years ago in the Florida elections on this subject is that people who fashion themselves as computer technical problem solvers do not understand when low tech solutions are more appropiate because they think everything can be solved by code and procedure.
In the arena of voting, the appropiate tech is to have interested persons watch the polls and witness the handcounts. There should be no computer counting at all. Results should be reported by telephone from authorized persons to election officss.
The problems of voting today are not solvable by computer. Computers in fact throw a layer of distance between counters and voters, and by doing so computers create uncertainty because vote monitors cannot ever know for sure if they are secure. One has to take another’s word for it.
The problem today with voting has to do with identity and absentee ballots. The non tech solution is clearly known but the political will has not yet reached critical mass.