The tampering with the vote takes place in the "programming" of the machines. This is "inside" the machine, accessible only to the manufacturer of the machines and the people who maintain or do the specific programming for each election.
As one expert has pointed out (and a Republican too), it takes a lot of time and trouble to alter 100,000 paper ballots. It takes three seconds for one person to change 100,000 votes. Plus, since the mistake would be inside the machine, a result of malicious programming either at the factory or at some point later on by some "insider," probably not some lone hacker (though that's possible), it would be impossible to discover the crime unless the guy who did it confessed, which isn't very likely.
The best method is the opti-scan machine which merely "counts" the vote and leaves the actual paper ballot so that it can be audited and recounted as needed. If you have the paper, you can always "audit" the results by counting the paper in a few randomly chosen precincts and if needed, recount the whole election by counting the actual paper vote.
BTW, do you have a link to an article about this Louisiana guy who tried vainly to prove that the machine could be compromised and failed to do so? It would be interesting to see who he is. There are perhaps a hundred articles about studies at many universities showing how easy it is to "compromise" the machines. I'm sure if you google "voting machine hacks university studies" or something like that you can get a huge number of articles. Bev Harris in the documentary HACKING AMERICA which showed on HBO in about 2006 takes part in an experiment in which a Finnish computer whiz uses a memory card to alter the results on an opti-scan experiment. It comes at the end of the documentary. But there are many many other ways to alter the results of an election with almost zero possibility of being detected.
= new reply since forum marked as read