And that is definitely something that can be done I think.
Being tamper proof is a thing that you address in a systemic way, with information, analysis, materials and procedures.
So a paper ballot is not tamper proof but a box that is viewed by many that is opened in the presence of many and counted, and re-counted by many can yield a result that all will likely agree on to a degree that the democratic process can be executed.
You could try to exchange the box, but there would be many witnesses, you could try to put extra ballots in it, but it would be detected and so on. The witnesses make all the difference, they turn a tamper prone piece of paper in to a tamper proof election device.
A computer in stead of that box with paper ballots is open to any number of ways of tampering that would go undetected, both locally and remotely as well as from the outside before the election has even been started.
I think the statement ‘The paper ballot process is practically tamper proof’ is essentially correct and as such don’t think that the statement of the Chief Election Commissioner with regard to voting machines in itself is incriminating. All I want to say is that saying something is ‘practically tamper proof’ is a meaningful statement.
The security (I use the term lightly) of electronic voting relies on not having knowledge about the system.
I don’t think that’s correct. I have a hard time imagining any system that is ever really tamper proof. Much is possible with enough resources, even tampering with the paper ballot. Tamperability doesn’t seem like a binary property to me.
When changing one vote costs you a few thousand Euro and the likelihood of the tampering being detected is still substantial (that would be my ballpark estimate for the paper ballot) the system is practically tamper proof and that’s good enough.
Sometimes we may even increase tamperability just to get the result we want. Absentee ballots are much easier to tamper with than paper ballots but we have them to allow as many people to vote as possible. That goal trumps the easier tamperability.
The problem is not that tampering with voting machines is at all and somehow possible – that’s just the same with paper ballots – the problem is that tampering with voting machines has the potential to scale much better than tampering with the paper ballot. The cost per vote is substantially lower as is the probability for detection.