If the paper record is not used to compute the tally, the trust problem is not resolved. Schemes to impact the election can fall under triggers to audit / reco

If the paper record is not used to compute the tally, the trust problem is not resolved. Schemes to impact the election can fall under triggers to audit / recount on paper. Secondly, why not just use paper then? The machine is adding near zero value. In the case of the photo, recording what the voter sees does not impact whether the actual record used reflects their intent. At best, we can verify they saw what they said they saw. The real problem is whether what they see actually reflects the record used for the tally. Edit: Also, if we personally identify votes, this could be used to reconstruct an election. Again, why not just use paper? Inputting a vote twice can improve mechanical to intent translation accuracy. However, the actual record still requires enabling tech to read, is still a temporary state, and is still not part of the chain of trust between voter intent and the record of the vote used for the tally.