Can secure voting exist on the Internet?

Last week I was asked this question: “If relatively secure financial transactions can exist, relatively secure voting can exist on the internet, can it not?” My answer? No. Because…

(please note that here I only have in mind online voting for actual political/administrative elections. In all other cases, e.g. to vote who should win some reality show, I care much less about the whole issue)

Financial transactions are “relatively secure” only because if they go wrong someone surely notices it, often immediately, and comes asking for a refund or repetition. With Internet voting, it’s impossible to realize that something bad happened. Unless it’s not secret, which would be so bad to be half disgusting half ridiculous.

It’s absolutely impossible to guarantee that all software+hw combinations used for voting by people who by and large use their birth date as password or never update software etc… would be free of trojans, keyloggers and such. This has proved tolerable for making online payments only for the reason in the previous paragraph, it could never happen with voting. People who couldn’t be bothered to vote could even never notice that their computer was infected to vote on their behalf.

If voting happens outside a safe place, there is no protection from abuse as in any variation of “vote now what I tell you and never tell anybody, or I’ll shoot you”

etc etc. So no, relatively secure voting CANNOT exist on the internet. Period. It makes me sad to see how many people who apparently thought of this for more than 2 minutes still propose Internet voting.

And above all: WHY? Let’s assume just for the fun of discussion that all I’ve said doesn’t exist: what would be the REAL advantage of internet voting? (BEFORE ANSWERING, please also read the best (short) argument ever against online voting!)