I frequently make the same kind of mistake when doing reviews: when given a vocabulary word consisting of a mix of kanji and kana, I fat-finger a letter for one of the kana. I would appreciate it if the UI would wiggle and point out that I’d made an obvious error in these cases, similar to the behaviour when you enter the reading that wasn’t asked for. As an example of the suggested behaviour, if the question was 気を付けて and the answer was きおつけて it would wiggle and prevent the submission due to a mismatch with the kana.
inb4 “be more careful”, I think adding this additional safety measure would be beneficial. The brain often doesn’t see what it wrote, but instead sees what it intended to write, and therefore re-reading your answer before submitting it doesn’t work reliably.
I respect and adore the design decisions in the UI that ensure minimal information is given to the user during reviews, since they prevent my brain, ever looking for shortcuts, from gaming the system. However, I think that since my requested change would be a cross-check against the question itself, leaking no additional information, it wouldn’t conflict with previous decisions. Also, I would be against a change such as removing the need to type the kana, turning it into a fill-in-the-blank UI, since that would hinder memorization through incomplete input repetition.
I am not above bribery. Just sayin’.