+- kanji koohii FORUM (http://forum.koohii.com)
+-- Forum: Learning Japanese (http://forum.koohii.com/forum-4.html)
+--- Forum: Learning resources (http://forum.koohii.com/forum-9.html)
+--- Thread: KO2k1 Optimized RTK Deck (/thread-11930.html)
KO2k1 Optimized RTK Deck - MaxHayden - 2014-06-26
When I started doing RTK, there wasn't a deck that did exactly what I wanted. I tried for a while to make do by suspending and unsuspending cards, but it got really aggravating, so I broke down and merged several of the available RTK decks into an Anki deck that did things the way I wanted. The deck is still a work-in-progress, but it's usable as-is so I'm posting it here in case anyone else finds this useful. I'd love to hear any suggestions for improvements and would be happy to work with anyone who wants to make this even better.
Here is the link to an Anki package. I can also provide the information in another format if people need it. To save on space, I didn't include the stroke diagram media files, but you can get them from this deck.
What's in this deck:
*The RTK1/3/4 Kanji
*Indexes for both 4th and 6th edition frame numbers
*Stroke diagrams, Koohii stories, and a bunch of other information for the 6th edition RTK 1 kanji.
*A special sort index that groups the kanji by KO2k1 level (plus needed primitives) and then sorts based on RTK frame numbers within each level.
What it's for:
Research shows that you can learn more and more quickly if you learn different kinds of things. So, according to the research, you are better off doing some vocab and some RTK at the same time instead of going straight through RTK and only doing vocab afterwards. If you wanted to implement this research for Japanese, there are two ways to go about this, either you can do RTK in-order and use a vocab deck that sorts using the RTK order (i.e. Kore), or you can vocab in morphological/frequency order (i.e. Optimized Core) and do RTK in "batches" based on frequency. I wanted to do the second option, but there wasn't an existing deck that worked with the 6th edition. Hence my making this one.
What needs to be done:
1) The RTK 3/4 Kanji are missing pretty much everything except kanji, keyword, and number of strokes (and one of them is even missing a kanji). Most of the fields could be screen-scraped with a script though.
2) Because the deck is skipping around the book, it needs to have "dummy" cards entered for the unnumbered primitives so that you know where to go in the book when it pops up in Anki. (Ideally, the dummy cards would tell you what kind of primitive this is. Some of them are just rare kanji. Some of them are radicals. Some of them are graphemes. And a few of them are probably made up for purely mnemonic purposes.)
3) The existing entries need to be checked against the errata.
4) The ko tags need to be checked against the KO2k1 list. I trusted the deck I merged that information from, but it had errors in other fields, so there are probably errors here as well. In particular the stuff tagged as a necessary primitive needs to be checked to make sure it really is necessary.
5) The sort order is currently sub-optimal. Right now, I split the kanji into 4 groups depending on KO level and then sorted each group by RTK frame number. But RTK isn't strictly linear b/c the primitives don't have linear dependencies. If someone had frame dependency information compiled, or if I compiled it while filling the fields in, I could do a much better sort by introducing the primitives in the "right" order to maximize your frequency coverage.
I hope you guys find this deck helpful.
Edit: As I'm working with this deck, I notice that the list I used includes some but not all of the prerequisites for the KO2k1 kanji. In particular if you have two simple kanji that make up a third and that third is used as part of a fourth. If the fourth is on the KO2k1 list, the two simple ones will be, but the third will not always be included.
There are a couple of ways to fix this. The easy way is to sort by the number of primitives each kanji is built out of and then to sort within each level by frequency. The problem with this is that it would end up jumping around RtK quite a lot and it would also force you to do more simple kanji than strictly necessary. On the other hand, it is easy to do. The *better* way to do it is to use the frame dependency information, but that might be a little complicated. So what I'm inclined to do is to tag the kanji that get used to make others and only put them at the front of the list. I'll play with it and see what I can come up with in a spreadsheet without having to write and special sorting code.
KO2k1 Optimized RTK Deck - Inny Jan - 2014-06-26
MaxHayden Wrote:If someone had frame dependency information compiled[...]Would this be of any help?
http://www.ravenbrook.com/tool/kanji-learning-tools/?action=kanji&keyword=picture
KO2k1 Optimized RTK Deck - MaxHayden - 2014-06-26
That's great. Thanks. I'll see what I can do with this.
KO2k1 Optimized RTK Deck - murtada - 2014-06-26
Thanks for making this, please tell us when you're finished
KO2k1 Optimized RTK Deck - MaxHayden - 2014-06-26
Can anyone who has been around longer tell me if there are existing tools that I could use to screen scrape the missing information from koohiii and an online kanji dictionary or if I'm going to have to make my own?