I still believe that a screening phase where people with limited support below a certain threshold can quit the race or be removed is the best way to have a functional ballot... to me it's just simpler this way. Even at real-life elections you need to show some signatures to access the race.
If, after weeks of debate, a person get 1/5th of the support of an average candidate, it simply does not have a real chance. I point out again here, this would not be an additional phase, it's just something that can be done in parallel to the presentation of the candidates. For example, at the nth support signature, you enter the ballot.
For some reasons, some people assume that "plurality" means that everybody
can join, but a crowded ballot is just chaotic. For n
places to be selected, you should not give more than 2n-3n candidates on
a final ballot, IMHO. Especially if you want to use certain electoral methods.
I tried to revise all 70 profiles and it was really boring. So after a while, I just put 10 names I kinda liked and that's it, I probably missed some of them. I also had negative feedback... which went wasted but could have also helped. Maybe in this scenario, the old method of "positive-neutral-negative" tipping box per each candidate could have also worked better than a STV ranking.
In any case with the other election I could more or less predict the probable final output (gender balanced, with actual limited chance for so-called GS), here it's almost impossible, the vote will be diluted so much and I really cannot focus on all the candidates. This ould probably mean that bugs of UI (fixed display of candidates, problem of selecting from menu if initial letter has an unusual accent...) might influence the outcome more than usual.
Alessandro