I understand and agree why WK doesn’t allow retrying a question after you’ve answered wrong, but I’ve noticed it’ll forgive you if you enter English/Romanji into a “Reading” question. When I’m typing fast and misread the question, sometimes I’ll mistake “Reading” for “Meaning” or reverse (not just me??? Please???).
I know it would be difficult (I’m a CS major, I develop software and tech, it’s hard) to try and guess that someone accidentally input the letters for Hiragana instead of a meaning (how do you parse the characters into Hiragana? How many characters can be between? What if the combination of letters could be two or more Hiragana if we remove certain characters, but not both? Yuck.), but what if there’s an exact match?
For instance, I was trying to answer a “Meaning” card for 又 and input “mata”. The exact Kun reading is また, which I would type as “mata”. Particularly because it’s exact, WK should be able to tell I saw the Kanji / Vocab and just flashed the first thing I knew about it. That being said, I know nothing about the backend of WK, so maybe it is really difficult and that’s why it hasn’t happened / isn’t an item of interest at the moment (or there are better things to do?). Thoughts?
Ah, I see. You’re saying the suggestion isn’t just a mirror feature that affects the reading questions instead of meaning questions, because what happens with the reading questions is input sanitation.
I think that’s a good point, but I also feel the result is the same. One might might consider an attempt to enter Hiragana when English is expected as invalid input, particularly if one suspects (within reason) that the input was the wrong type.
Not to mention, I feel like a few levels in you know the Hiragana pretty well, and invalid input on a reading question is likely to be a mistake or misinterpretation.
I don’t know if I understand correctly what you mean, but I’ve recently implemented a “try again on reading/meaning confusions” feature into rfindley’s Double-Check userscript. However, I don’t know if it will be accepted. Also because it doesn’t really add much value to its functionality - the script already allows to undo answers that were counted as wrong.