On paper voting

Many years ago I volunteered on a campaign for a friend running for a city hall alderman position (like a councillor, or whatever, you can figure it out).

Voting was done with paper ballots, and the way the votes were counted is very interesting, and has bearing with the recent news that foreign countries were able to hack in to American voting machines and actually change results.

The city had a ballot counter present at each precinct (voting hall). Each candidate was allowed to send in one representative to verify the count; that was my job. Scrutineers during the day had verified voter ID, so we could assume all the ballots were legitimate.

Each of us maintained our own count of the ballots. In our case, my total and the opponent's rep's totals were in agreement with the city worker's, so she simply phoned in the total into a central office. I don't remember what happened to the ballots, but I assume they were packaged and also sent to a city government office.

This system has so many good points.

1. It scales up, even nation-wide.  If there are no challenges, ballots would be counted precinct-by-precinct, so counting up something like 300 ballots with, say, 30 questions (perplexing to Canadian readers, but not so unusual for the American readers, which is one of the reasons Americans have such low voting turnouts) would take maybe 6 hours (assume 9,000 separate votes each take two seconds to process, so that's about 5 hours, with another hour to do verification). The networks won't be happy that they can't release results at 8:10 PM Eastern time and call a winner at 11:15 PM once the west coast reports, but as we've seen, the price to pay for getting instant results is too high.

In a state-wide system, each precinct would report its totals to something like a county office, where the same n-way verification would take place. In this place each person would type in the totals and add them up in a spreadsheet.  Yes, there's some copying and typing going on, with room for human error, but because there's an audit trail all along the way, discrepancies should be readily discovered and fixed. The counties then report their totals to a central state office. And anyone who watched the 2016 office knows that each of the networks is capable of adding up the sums for the electoral votes in each state.

So while I'm staying we shouldn't use computers to record the votes, there's nothing wrong with using them as a tool to help do simple math. In the absence of another Intel bug like the one 25 years ago that caused some spreadsheet-based errors, we're talking small integers here, the kinds almost any computer has no problem with, and spreadsheet templates can be prepared in advance.

2. It's not easily hackable. Maybe a ballot-checker could be a plant from an opposing party, but there's still a theoretically neutral government official counting the ballots.

3. It takes more time to fill in a ballot with a pen than using a computer. On the other hand, you can have many more people filling in cardboard ballots in parallel than having to wait for machines; you're only limited by the number of desks that you can put a privacy shield on and the number of pens to hand out. I expect this would be a win.

4. It involves more citizens in the democratic process, as we'll need a lot of people to help report the counts. This is a good thing.

5. There could be a public audit trail for the whole counting process, so anyone can see totals by precinct up to the district or state level, something we didn't have back in the paper-ballot days. You don't need a block chain, just a web site with PDF and CSV files. Ballot count verifiers should be given a card with the precinct count; if they see a different number on an official site it should be easy to launch an investigation. We couldn't have this back in the paper-ballot days because we didn't have web sites, let alone PDFs and CSVs, or even files.

This combines the best of electronic, paper, and human processes.

Or we can stick with using closed-source computers with undocumented but discoverable ports, and keep seeing people with questionable legitimacy step into positions of unearned power.