Still plugging away at things, trying to improve edge case handling for when folks have errors.
A few changes:
When login or sign up fails, errors are actionable and styled
The forms will attempt to re-fill themselves out so you don’t have to start over (esp. that API key)
Error messages can now be transmitted via the API, including fatal messages which can stop the app in its tracks in the event it enters a looping condition
Fixed a couple such looping conditions (example: if the request to load an item failed, the app would continually request it ad infinitum, wasting everyone’s CPU/bandwidth)
KameSame’s search feature will now also search Kanji & Vocab readings when you supply romaji in the search field. This is super handy when you’re on mobile and need to quickly hunt for a word during conversation without the lag of switching keyboards or pecking at the kana keyboard.
This is something Wanikani doesn’t do, but I feel like search is more important to KameSame because its audience is mostly people who’ve already cleared items in WK and therefore “know what they know”. Additionally, KameSame is meant to be used on the go (like, for ad hoc lesson addition) when you encounter a word you struggle to produce, so getting search right is really important.
In terms of short-term roadmap, there are a few major features I want to tackle soon:
In lessons, allow users to choose which types of lessons to pull:
Random order (current state)
Frequency (descending; I don’t think I"m going to find a good enough corpus for this, though; I experimented on this one but wikipedia terms way undervalue a lot of 日常会話 terms; “refrigerator” ranked dead last, for example)
WK level ascending
WK level descending
WK incorrect rate ascending (words you know well, requires me to download everyone’s review statistics, which will be a bit of work and migration hassle)
WK incorrect rate descending (words you struggle with, “”)
User account maintenance
Change your WK API key
Change your password
Forgot password reset workflow
Verify your e-mail (including a backfill workflow for this chore for current users)
Download user synonyms for all their WK items, and then cross-reference these for alternate matches when quizzing them. This is really important IMO, but in order for it to be performant, will take some reworking (right now, all the alternate matches for each item are crunched in advance and takes a nontrivial amount of time, which won’t be as feasible for N users
Implement a KameSame synonym system for Japanese synonyms to accept for given challenge words (in cases where multiple japanese words mean the same thing and you want KS to accept other answers)
Implement a back-button UI in nested views (particularly show item) so that KameSame will be PWA-friendly on most mobile platforms as soon as Apple gets their act together and un-breaks PWAs (hopefully soon)
Another fun update today: you can now customize the order of your lessons!
When you start KameSame lessons, the default has been (and remains) to pull available lessons in randomized order. Later we added minimum WK progress levels (e.g. only show “Guru” or “Burned” items). Then we added a checkbox to exclude Kanji cards.
Hi @searls noticed one bug that I didn’t see mentioned above ( sorry if I missed it): My level seems to be stuck. Now it seems to still update the vocab fine (e.g. items that increase levels seem to apply), but both profile and front page seem wrong.
Also just a suggestion for the akai/aka problem - how about just putting the ‘part of speach’ information on the card. E.g. if adjective put a small icon somewhere to the side. This could get complicated but for practical purposes adjective or not seems to be the main information to glimpse from there.
I see you mentioned you don’t intend on integrating with Jisho. Is there a particular reason for this?
I don’t see what exactly this integration would buy users, to be honest. Every integration increases the logical complexity of an app (not just in keeping it working, but maintaining data integrity, and dealing with cases where data is 1…n places but not 0…m other cases), so the value proposition needs to be pretty extraordinary to really be worth it from a maintainability perspective. (This isn’t about “I don’t want to spend the time”, at all, it’s just a fact of life that eventually apps fall over on themselves and become harder to maintain the more complexity they have.
If someone can make a clearer case about why they’d want Jisho integration I’d be interested to read it.
About the 赤 vs 赤い vocab items, will you be handling edge caes like this so it’s more polished overall?
I’ll probably start printing parts of speech and also adding part of speech hints into the placeholder. (e.g. change 単語を入力して… to ｀形容詞を入力して…` for adjectives.)
No, this is a completely reasonable idea. I’ve been using Japanese IMEs on Windows/Mac for over 15 years now, so “hitting enter twice” is something I had become inured to (and was surprised to see feedback about).
This might be a reasonable option, but I have a concern that enabling it might throw off people’s flow, since each card would no longer take a predictable number of enter keystrokes to continue.
That said, it might just as well give you greater pause when you get one wrong. On WK I usually get 85%+ scores on my reviews and that means I just mash enter to get the next card but will usually whiz past my wrong answers and have to get them wrong twice.
Understood. I hope I did not come across as entitled when I asked my question, as that was not the intention. It was curiosity above all. Thank you for the response
I think the case for Jisho is about pulling synonyms from there so users don’t have to add them in themselves.
My plan is to pull in user meaning synonyms from WK itself. That way, if you get two words confused that roughly mean the same thing you can ensure they’ll show up as an alternate match (instead of an incorrect answer) in KameSame.
Separately, I’m going to add Japanese synonyms to KameSame at some point to allow you to specify alternate correct answers (necessary for silly cases like 手洗い versus お手洗い)
@searls This is extremely embarrassing, but when do you plan to implement password resets? I created my account on my work computer using a LastPass generated password and forgot to hit the “Save” button after creation, so now I’m tied to only my work computer
Hey, sorry about that! I hope to get to it soon-ish, but it might be a couple weeks. For now (and I know this is not super secure), email howdy@kamesame.com with your registered email and I can set it to something else manually
Parts of speech are now declared underneath the item.
That means alternate meanings are currently not being shown. This is because I’m working on a mechanism for pulling in user synonyms, and how to display either or both of them will probably result in my having to stash them someplace else.
Fixed a bunch of minor rendering issues, you ought to see somewhat smoother and less janky rendering performance now across all devices and form factors
Fixed positioning of a bunch of elements that didn’t look right on every display
Add an icon for the account menu and a logo to the login screen
Thank you this app has really been such a good supplement to my WK reviews.
I have one tiny problem though, whenever I try to insert 「のま」into my IME and try to get the kanji for “repeater” it does not give me the kanji that the site is looking for…
I don’t think it’s a problem with the IME I think it’s just something that I’m doing wrong and have been trying to figure it out by myself for a while but still can not figure it out. The kanji has the same kunyomi and onyomi readings according to Wanikani so there are no alternate readings I can put in the IME to get the kanji out.
If you hit space bar two more times before confirming the IME would have submitted のま. The easiest way to type one quickly is probably to type a word like 次々 and delete the leading kanji
