Open jeffberhow opened 10 years ago
Still working on this! Albeit glacially, hope you don't finish RTK before I'm done!
Haha! You'd better hurry! I'm about 24 days away! Just kidding, this will be a great tool afterwards as well, and maybe I can think of suggestions to extend it's lifetime outside of RTK.
On Wed, Aug 6, 2014 at 4:56 AM, Ahmed Fasih notifications@github.com wrote:
Still working on this! Albeit glacially, hope you don't finish RTK before I'm done!
— Reply to this email directly or view it on GitHub https://github.com/fasiha/kanjiwild/issues/10#issuecomment-51324908.
Just saw the "Reading" pane at Koohii: http://kanji.koohii.com/sightreading this is similar to what you want (but not exactly the same).
@jeffberhow I made a bunch of changes. Take a look at it now: http://fasiha.github.io/kanjiwild/ (The app at http://aldebrn.me/kanjiwild/ link is still running the old version, and I'd like to retire that server. Github provides nice, reliable hosting, so why not use it.)
It's not quite what you describe, but I implemented a bunch of low-level things that will make it very easy to add the features you asked for. The continuously-updating tally and overall percentages will be easiest to add. I had a prototype of a click-induced popup for the keyword, but left that out---I'll add that once I get some more feedback on the keyword interaction so far.
I also added a couple more features that are important now that I'm doing RTKLite and am skipping some RTK kanji..
The realtime counter is now in ("1 kanji found, 6 left").
How do you feel about the box appearing below the block of text? I know when I paste a big chunk of text, it can get annoying to have to scroll down to enter the keyword, but at that point, I'm just having fun clicking a bunch of kanji, then scrolling down and entering keywords, and then back up to kanji.
I think that would be fine. How difficult would it be to implement all ideas and have a boolean debug-type switch to try each one?
I can see the benefit in both options and they may work different parts of the brain or, like you said, one is more fun at times.
By the way, Ahmed, I notice I am using this more after finishing RTK.
So for my own clarity let me list the ideas so far and compare to current functionality:
By itself Idea 2 might not seem like a big improvement over how things stand currently, but combining Ideas 2+3+4 would certainly add new functionality: the current way the kanji turns green when you click on it makes this app mainly a kanji recognition app, with only a (distant) secondary focus on keyword production---it doesn't even track the number of correct keywords. I can certainly see the benefit of 2+3+4 for someone who really wanted to practice the keywords. I personally have been aggressively annotating my keywords with extra definitions and stock phrases to help differentiate between keywords, so producing keywords is a hit-or-miss affair for me (usually I remember the general sense of the keyword and its various senses, but don't ask me to remember whether 身 is "someone" or "somebody").
Let me think on how to make it more useful to people wanting to practice keywords.
I enjoy pasting in large books (like Botchan by Souseki Natsume) and seeing how many non-RTK1 (and even non-RTK1+3) kanji they contain. Glad you find it useful too, and hope to share in your post-RTK bliss someday soon!
By the way, I'm on Mozilla Firefox 32.0.3 via Arch Linux 64bit kernel version 3.16.3-1. When I click on the kanji, it does bring up the string "Enter English keywords here:" but no box appears. I will comment on the rest later; time to eat. Oh, attach "someone" to an actual person and somebody will have a vague sense. I attach mine to my fiancee Lucy, so all stories are related to her. I hope that helps. bbl
Regarding Firefox: thanks for the bug report!!! Fixed with befaf2e6...
Wow, I really like the clickable idea where the boxes show up at the bottom; much more-so than the hover boxes, which may get in the way of a click-fest. Unfortunately, I can't be certain without trying, but that's not important. 2+3+4 up there sound great and I can't wait until it's implemented.
I will try Botchan out.
@jeffberhow check it out now: https://fasiha.github.io/kanjiwild. After clicking on a specific kanji, if you hover over any instance of it, you'll have a box to type in the keyword, plus the "×" marker to "unrecognize" the kanji. I also added some basic statistics display for keyword accuracy.
It took a bit of work but now these hovering keyword inputs should play very nicely with the keyword input section below the text, e.g., when you add new input text or remove existing ones or change the number of kanji you know (by reducing/increasing the Heisig number, etc.).
I haven't spent a lot of time polishing the styling of the hovering boxes, so they still look at little 1990s (I'm really regretting the monochrome theme I started with because the pastel green on gray is really ugly), and I'm not totally happy about the positioning (the hoverbox covers up the lower-right quarter of the kanji, which is too much, but if I decrease the overlap, it becomes harder to keep hovering to get the mouse to the input box), but the basic functionality should be there.
This gets us to 1+2+4. The only thing remaining is (3), which is to change to color of the kanji to indicate correct keywordage; I'm trying to sort out the color scheme with this in mind (I am not good with colors :worried:).
In IRL mode, you click on a kanji in the diplayed text and a text box opens allowing you to enter a keyword. If the keyword is correct the kanji turns green; otherwise it stays black. A tally of the percentage of known (green) kanji is always updated.
This allows seeing the kanji in the wild with other kanji instead of the columns of single kanji. It also allows to get a good idea how many of the kanji are known in the text --percentage-wise.
In the future, this could be a plugin for firefox and will work in all settings.