kevenbify wrote:

I have recently discovered Anki and have been using Tae Kim's Guide to Japanese Grammar, Heisigs RTK 1-3, and Japanese Core 2000. However, I'm not sure if I should follow the JLPT decks for kanji and vocab, or just keep the core 2000 deck? Any suggestions? I'm sorry if this question has been asked!

I think it's important to focus on that which is the most relevant to your interests as motivation is the difference between grudging forcing yourself through 50 reps vs spending all day doing 1000+ and fully enjoying yourself. So I'd recommend custom tailored subs2srs deck > premade frequency based vocab decks (eg, Core) > mostly arbitrary test requirements (JLPT list).

When/If you do later on decide to care about JLPT, I'd use MorphMan to match the JLPT words you still don't know to subs2srs sentences.

Tori-kun wrote:

When/If you do later on decide to care about JLPT, I'd use MorphMan to match the JLPT words you still don't know to subs2srs sentences.

How to do that exactly (in steps)?

0) Create a known.db containing all the morphemes you know.  MorphMan 2 will do this automatically if you let it manage your decks, otherwise you can create one manually: in browser, select the facts you know well and do Actions>It's Morphin Time>Export Morphemes, provide the field you want to get morphemes from (eg, 'Expression') and the name of the db you want to save it to (eg, 'known.db')

1) Create a DB of morphemes you want to learn. You can use Export Morphemes as above on a selection of cards, import morphemes from a text file of a book/article you want to read, combine various databases together using set operations like union/intersection/difference, filter by part of speech, etc. In your case, generate a DB from a JLPT deck and subtract your known.db to get the remaining morphemes you don't know.

2) Open a deck with cards you want to learn from, make sure the sentences are in the 'Expression' field and add a 'matchedMorpheme' field to your model (these weren't made configurable like the other tools but you can modify the source if you care) and do Actions>It's Morphin Time>Morph Match. Provide the database you want to learn from and wait.  Matching a DB with 2000 morphemes against 2000 cards might take a few minutes depending on your machine and I'd recommend testing it with smaller numbers first (it's doing a maximum cardinality bipartite matching which can be slow).

I have my card layout then show the matchedMorpheme field in big blue letters and retain some of the sentence card stuff (screenshot, audio, context sentences, etc) although obvious it's not necessary.

Last edited by overture2112 (2012 March 09, 9:34 am)