[GiraffeNecktie]

"Flashcard systems don't work"

No one tool is going to do everything. Flashcards are actually pretty good for acquiring masses of vocabulary. But it doesn't help you put those words together coherently or even enable you to comprehend what you hear.

Language acquisition is all about growing a set of extremely complex interconnections between brain cells. You need to grow a massive number of interconnections between many areas of the brain (including to sensory and emotional areas) and you need to strengthen those connections by repeated firing.

Software, books, audio and video can play a supportive role but there's a big chunk of the brain development that is only going to come from struggling to express your own mind and to interact with real people in real time.

[hmbg]

Agreed; flashcards have worked quite well for me for learning lots of words. Also for me, vocabulary is normally the limiting factor, while Getting grammar an pronounciation up to a usable level is a lot easier. I guess that differs from individual to individual.

Trying to work with flashcards without a context (Learn the 1000 nost common words!), however, doesn't work at all. The words need to belong in a context with which I'm familiar, such as a book, film, or recurring conversation topic.

My attempt at vocabulary training (shameless plug, you call it): http://vocabulous.net. But that's pretty much only vocabulary, still on the drawing board on the rest of the components.

That said, the most important thing for the next big thing in language learning is that it's fun. Very few users will stick with it on willpower alone. Good thougts on this is at http://www.alljapaneseallthetime.com. If I ever get the idea to learn japanese I'd check his stuff out.

[dpapathanasiou]

"Flashcards are actually pretty good for acquiring masses of vocabulary. But it doesn't help you put those words together coherently or even enable you to comprehend what you hear."

In my experience with Japanese, using flashcards for kanji doesn't really help with long-term memory acquisition.

I agree with your latter point about context; it's better to find an unknown word in a sentence and look up its meaning, then you tend to remember the word longer b/c you understand how it's used.

[da3da]

Have you tried using Heisig's "Remembering the Kanji"? I have never tried to learn Japanese, but I have heard very good things about using this book to learn the Kanji.

[oftenwrong]

For anyone using "RTK" I recommend using http://kanji.koohii.com/ . I also recommend that you actually write the kanji out (on paper, or even just tracing it out with your finger in the air) when reviewing, and only consider a kanji remembered when you can render it properly, with the correct stroke order.

[dpapathanasiou]

I've heard of Heisig, but only recently.

My method for learning kanji was to count the strokes (it's relatively simple to figure the stroke count for any character: http://infohost.nmt.edu/~armiller/japanese/strokeorder.htm) then use that number to look them up in a dictionary.

But everything I looked up was in the context of a sentence, while trying to read a report at work, e.g.

[riffraff]

how did you use flashcards? I.e. did you use some "refined" form of spaced repetition or just moved a card at the end of the pile?

IMHO the difference in the quality of learning varies greatly across the paper flashcard/supermemo-ish software continuum.

[dpapathanasiou]

I don't remember which set I tried, but their characters seemed disjointed/random and, worst of all, unrelated to anything I was doing at work.

I just found it better to try to read the reports and memos being passed around at the office, and once I looked up kanji I didn't know, I would tend to remember them, since I'd read them in the context of a sentence.

[wahnfrieden]

You can use flashcards to learn kanji. The problem is trying to learn them without context and by learning their readings rather than by learning the reading via learning vocabulary which use the kanji you want to learn.

[kumarski]

"No one tool is going to do everything."

Why not? Isn't it something they should strive for?

[kd0amg]

I think in a lot of cases, effort can be better spent by accepting a combination of tools (e.g. for dealing with screws in my glasses and screws in my kitchen table, having multiple screwdriver heads is probably a lot simpler than designing one that can handle both).