Open martinetd opened 3 months ago
I've been thinking about this too, but I feel like adding an artificial mistakes for repeat might not be very intuitive and flexible. As an option there might be an additional slider on the practice saving step with how much mistakes to add for each repeat, this way it will keep flexibility, but it still feels weird since characters can have different difficulties and stroke count...
Yeah, artificial mistakes is definitely not intuitive -- from a user point of view I'm sure I wouldn't understand it myself. It works for me as a workaround, but it's not a solution.
However just changing the mistake count doesn't work for me, I'm actually pressing bad with 0 mistakes from time to time, so there's no way that'd catch up on these. I think we can keep the mistake count as is as a hint/reminder for the user, but base the good/bad selection in recap screen purely on whether "bad" has been pressed, e.g. track it as a different variable. It's a bit intrusive in the code but I think it's the best way to go: it reflect what the user thought about the kanji at the time one saw it during the review itself.
When reviewing (write practice) I sometimes do small mistakes that the recognition is too nice about and accepts, or just make a single mistake but still click "bad" to reschedule a review.
However, at the end screen only kanjis with more than 3 mistakes are selected for re-scheduling, regardless of whether bad was pressed or not. That means that I often need to go through the list and try to remember where I pressed bad or not.
I've made a local workaround that just adds 10 mistakes whenever bad is pressed (diff below), but that's not really acceptable for submission -- I guess we could add a separate "repeatMap" and flag kanjis as bad in PracticeSavingState regardless of mistake count, but that's a bit of code shuffling to get it all the way there so I've been lazy at this point.