We've got this!

  • Vivaldi Team

    Vivaldi developer Patricia Aas shares her story of a personal Twitter campaign challenging Norway’s ballot counting system and being heard – just 10 days ahead of a parliamentary election.

    Click here to see the full blog post

  • Thats a pretty cool story. Glad things worked out.

  • @lonm: Thank you!

  • Moderator

    For german 🇩🇪 federal election (Bundestagswahl) end of September a official but uncertified software is used for the digital (unsigned!) transfer of votes (ball paper) from polling station to central servers. But the transfer can be faked, login passwords for the software can be found in the WWW or even data like vote counts for a political party can be changed.

    Now i know why our politicians fear Putin's hackers who may influence the upcoming elections.
    But we have White-Hat-Hackers like the Chaos Computer Club or other IT specialists who investigate such bad software.
    See https://www.ccc.de/en/updates/2017/pc-wahl

    You see that eGovernment and eVoting is nice but not always trustable in Gemany.

  • Vivaldi Team

    We are all very proud of Patricia. It takes guts to do what she did. Guts and knowledge and she has plenty of both!

  • Moderator

    @patricia-vivaldi: Good story, Patricia! Thanks.

  • Without going into detail, what were the main concerns with the Norwegian system? Is a manual recount simply to validate the electronic ballot? What happens when the manual count is wrong (which it will be)?

  • Moderator

    @lonm It certainly is a cool story. Well done, Patricia!

  • Vivaldi Translator

    💐 ❤

  • Vivaldi Translator

    @mossman: If I followed correctly (not knowing norvegian), the main problem was that the ballot were optically scanned, and never counted by hand. And the integrity of the softwares of the scanner was never demonstrated. So trusting only in them was stupid.

  • :Knight: Even here in the US it is hard for some researchers to get through to companies of the vote counting machines. I know more than one time a manually count of votes has been demanded and has happened many times in many different elections.

  • @mossman: The main concern has been to have something to compare the machine count to, as @cqoicebordel says below. Now that we do, we have to have a procedure for handling the instances when they don't match, like you said. Currently that is not completely in place. But there are parts that are very useful (this is my current understanding):

    1. During the preliminary manual count the ballots are sorted and packed according to party.
    2. When the machine count is being done, the operator(s) know that all ballots in the package is supposed to be from a specific party and how many there are supposed to be.
    3. If the counts don't match you can now do a partial recount of just that one package.
    4. How that recount is done is, unfortunately, not specified from the government, which I have been tweeting about this week. But it should, imo, be both a manual and a machine recount of that package.
    5. This system would ensure that the process was auditable by election observers and volunteers. They don't have to trust the machine or know how it works - they can directly observe the result for a single package.

  • @gwen-dragon: i'm posting that on Twitter - thank you!

  • @mossman: "Right" or "wrong" isn't a matter of extreme exactitude. Yes, humans make mistakes and it's unlikely that a manual count is perfect, but the important thing is that the correct result is reached. A systemic error in a manual count is unlikely given the presence of impartial observers (or observers for all candidates, or both), but a systemic error in a machine can and does happen, a lot.

    Unless the race is extremely close (a few hundred votes or less), a manual count should have sufficient accuracy to validate the accuracy of the machine.

  • @jking I think you misunderstand me - a hand count will nearly always produce a small error. I'm not saying there's anything wrong with hand counting, but it's highly likely that the machine and hand count will disagree - so even if the machine is always right, the recount may repeatedly say it's wrong. The question was if that would be considered margin of error on the human side or would people tend to think there a systemic issue with the machine...

    But I'm satisfied with the detailed answer from Patricia since it does address this.

  • Oh. When I saw a new post in the blog with an envelope on the picture I thought that email client is finally here...

  • @fifonik If (when?) that becomes available it'll be after a major version jump, not partway through the snapshot cycle. But yes, I too look forward to that. 🔜

  • So what was the beef? I missed that.

  • Moderator

    @moondawg That there was no human verification of electronic results (which are hackable).


Looks like your connection to Vivaldi Forum was lost, please wait while we try to reconnect.