> On Mar 16, 2015, at 7:37 AM, Daniel Gruno <[email protected]> wrote: > > To me, it's not a question of "should we try", but more "do we want to?". > I can push the changes to the repo and it should just work (maybe with a few > more nudges), and would only show up if you add the "?interactive=true" to > the URL, so nothing will change WRT the old system. > > If we don't, then it's probably another year before we can get a lot of > people to try it out at once and see what feedback we get. > What we could do is try it for the test ballots, get feedback and then decide > based on that. After all, it's an option for people, no one will be forced to > use the interactive page, they can pick between the old one or the new one as > they see fit. > > I realize we're quite close to the election (which is what made me start > working on it in the first place ;)), but the math inside the new page is > sound and it's easy for people to double-check that the resulting STV string > matches their choice and order, so even with all my bias, I'd say we should > at least give people the option of trying it out for the test ballots. If > they don't like it, we can wait till next year for the real election. >
I tend to agree. The test-ballot will help verify that all is well, plus if there is some weirdness w/ the impl messing up the order of the votes, when the person gets their verification, they should see it.
