KanjiMood
Member
Registered: 2009-04-06
Posts: 132
So recently I've been SRSing Japanesepod101 dialog (which is working great for me). I use audacity to quickly cut out the audio, and add English > Japanese cards on what I think is appropriate. But as I'm progressing (currently got about 300 cards) I'm wondering when I should stop doing English > Japanese and begin making them as Kanji > Kana (or Kana > Kanji? not sure which is best) cards instead.
I only ask because I recently started seeing more words like 自動販売機 and 公衆電話..
I'm also doing core6k, but it's a lot less interesting than the above deck IMO..
Thanks for any suggestions! 
SomeCallMeChris
Member
From: Massachusetts USA
Registered: 2011-08-01
Posts: 787
I don't see anything wrong with doing English->Japanese cards as long as they are working for you, but I would also to Japanese->English cards.
The reason people don't generally do English->Japanese is just that it takes effort to create cards where there's only one correct answer, and as your vocabulary grows and you learn more synonyms the problem becomes worse. So basically... you should stop doing E->J when you start having trouble answering because you know there's more than one right answer.
However, if you don't -also- do Kanji->reading+meaning, then you won't be testing what you need to know to actually read.
Also, let's -not- discuss the source of ryuudou's advice, we don't need another flamewar.
SomeCallMeChris wrote:
I don't see anything wrong with doing English->Japanese cards as long as they are working for you, but I would also to Japanese->English cards.
The reason people don't generally do English->Japanese is just that it takes effort to create cards where there's only one correct answer, and as your vocabulary grows and you learn more synonyms the problem becomes worse. So basically... you should stop doing E->J when you start having trouble answering because you know there's more than one right answer.
I agree, there's nothing fundamentally wrong with notion of doing E -> J production cards alongside J -> E. They are definitely not mainstream around here though...
The single answer card can be difficult to create, but it's pretty simple to use a core deck and add an extra field and add any info you need (or just add the English sentence) as you go along. I know the pain of knowing more than one answer from iknow!
Production is also 'slower' than recognition -- especially written production, so it takes longer to complete the core deck. Some may see this as a waste of time in comparison to pure recognition.
I've changed from just one recognition card to 3 card types per fact, one of which is English production to reading. Personally, I found this helps my comprehension which I found was also the case when using iknow (which features production). You do end up with more reviews, but you can always up the review rate to compensate a little -- mine went from 100 to ~200.
Still you have to experiment with your cards and see what works best for you.