I think it would still be fairly accurate to say that Japanese people learn kanji largely by rote and their approach isn't all that systematic.
It's true the characters aren't strictly learned in the order of frequency, but I think their assessment of "usefulness" is the highest factor in which are learned sooner, with stroke count being the second factor.
I definitely think it's backwards that they learn 飲 a year before 欠.
Japanese kids learn every single kanji they ever learn starting with the radical.
Heisig like stories are also not that rare in Japanese education either. I think every Japanese kid learns the exact same story for 親 that Heisig uses.
I won't say that the way I learned to read Japanese was perfect, because it involved a whole lot of trial and error, and patchworks of methods. But I know this much:
Learning kanji out of context doesn't work for me. Learning the abstract meanings of kanji doesn't work for me. What worked for me is doing a lot of reading, and learning a lot of words in context. If someone asked me how to learn kanji today, this is what I would tell them:
-first, focus on spoken Japanese and on Japanese written in kana (or with furigana). Learn how the language is put together and start to accumulate some basic vocabulary.
-do a lot of Japanese reading with furigana: news sites for children, manga, novels for young children. Do some Japanese reading without furigana that's on a fairly simple level with very restricted vocabulary (i.e., textbooks -- or also, take a passage that you've already read and understood, remove the furigana, and read it that way several days later.)
-When you start reading novels that don't have as much furigana (but that are still pretty simple, without much specialized vocabulary), you shouldn't have to look up a ton of stuff. You'll definitely be looking up some kanji, and some vocabulary words, but with prior reading experience and guessing-from-context, you should be able to keep the lookup down to a manageable level. Because you don't actually have to know every word on a page to be able to understand the general meaning.
-If you feel that you should learn to write the kanji, you can start this at any time in the process, but generally, stick to kanji that you already know in context. Rather than going from meaning to kanji, go from kana (in a sentence for context) to kanji.
I didn't follow this, precisely, when I was first learning Japanese. I just read a lot of manga and novels because I wanted to, and it turned out that I was learning a lot more kanji that way than by trying to deliberately study them. But now that I'm learning Chinese, that's pretty close to what I'm doing, and I have to say I'm pretty happy with my progress so far.
(Generally what I do is take a lesson from Chinesepod, listen to the dialogue a couple times, copy the dialogue [hanzi-only, no pinyin] into my SRS, and study it that way. Focusing on vocabulary, not individual hanzi.)
So, in this way that I've laid out, there's not much attention paid to writing, and you can ignore it altogether if that's what you want. Some people may find this outrageous. The thing is, I've tried learning to write the kanji when I didn't have a strong enough vocabulary base for it, and I always found that if I didn't already know a kanji from reading then I wasn't going to have any luck learning to write it. Now I'm studying for the kanji kentei, and I find that it only takes me a couple of repetitions to learn how to write a kanji, because I already know it really well from reading. And personally, I have zero regrets that I spent a while ignoring writing in favor of reading, just because I had to little reason to have to handwrite much from memory, even when I was living in Japan.
Nothing in RTK is particularly magical. Pretty much all of this stuff has been done elsewhere. All heisig did was create a systematic method for learning the kanji based on how he found it easiest for himself to learn it.
I could do the exact same thing and produce my own book, maybe my method will be even better than Heisigs, but it would probably work out to be really similar.
Ok, so what does heisig do?
First of all, he arranges the characters in a really logical order. Pretty much any book that you follow will have the characters arranged in a particular order. And surely, that order seems logical and effective to whoever made that book. Heisigs order is particularly effective for people who want to learn all the Jouyou kanji. Its probably not the best order if you just want to get to hit some baseline goal of minimal functionality in the language in a short time.
Next, Heisig says, lets forget about how to pronounce the characters for now, lets just learn to recognize them and write them. One thing at a time!
It's not exactly the most popular idea among people who teach Japanese, but again I doubt Heisig is the first person to consider it. As a long term strategy, this is fine, however it means you are still completely illiterate after learning how to recognize and write all of the kanji.
Then, the whole keyword thing.
This is merely a byproduct of not learning the readings. You need SOMETHING to prompt you to be able to produce the kanji, so if its not going to be a Japanese reading, there isn't much else to use except English. Perhaps pictures would work, but that's more complex.
This is a cornerstone of heisigs method, but again its nothing new. People have known about the effectiveness of these strategies for centuries. Its usually not really taught to students in school for some reason though, so not many people really know how to do it. When I was taking Japanese classes before doing RTK though, I saw some students doing this on their own. They didn't have any systematic method though.
So now, as for the original question, is there any alternative to RTK?
The question is framed such that you either do Heisig's method, or you learn them rote with no strategy at all. However the fact is, there are a whole lot of possibilities between the two extremes.
What about using stories to learn kanji in the same order that Japanese children learn them in? You get the most common and useful kanji more quickly, which seems like a nice benefit. It could be argued that the first few groups of kanji would be more difficult because you would be learning lots of radicals/primitive near the beginning. But since the traditional Japanese kanji order starts out slowly with like 100 kanji at the beginning, this kinda balances out.
Or what about following heisigs method but using Japanese keywords. This gives you a reading at the same time!
And about those stories? I'll bet just reviewing in Anki without stories will still let you learn them. Might be harder, and you might mix some of them up easier, but people do it.
So really, you can create your own "mix-and-match" kanji method, and that would be an alternative to heisigs method. Maybe it would even be better for you, because you could design it for your own personal goals. No one method is best for everyone.
RTK didn't work for me. I got to about 800 characters, got the stories all jumbled and confused (or couldn't even think of good stories) and was not enjoying the process at all. I know the OP claimed that with Heisig you write the kanji max. 3 times to remember it, but with me, that was not the case at all. I was writing them over and over again, and still forgetting.
What worked for me was learning vocabulary, and using them in context.
The Kanji Odyssey books worked really well for me. Their ordering made more sense to me than Heisig's way. They ordered the kanji by frequency - how often do they show up in native materials, which was much more helpful than grouping them by 'primitive' and ease of writing. They also included 3-4 vocabulary words per kanji, along with example sentences showing how to use those words. So while I admittedly worked slowly through the book (about 11/2 years), I can now recognize and sound out 1100 kanji along with vocabulary that uses those kanji.
Currently I'm working my way through the Core 6k to fill in any gaps, along with reading Japanese every day. Simply reading manga with furigana helped me more with learning kanji than RTK did.
Honestly, I think exalting one method of learning as the be-all-end-all is not going to help new learners. AJATT for example, says that you absolutely have to do RTK and pretty much infers that you will fail at Japanese if you don't. I've seen many people on this site who have tried 2 or 3 times to get through RTK and cannot seem to finish it. That tells me that maybe they need to try a different way.
When RTK didn't work for me, I felt pretty disillusioned with AJATT and a lot of the advice I was seeing online. My beginner textbook, which was teaching me vocab, kanji and grammar all at once, was so much more helpful. The only valuable thing that I took from RTK is that I cannot learn the kanji by themselves. I wish that there was a more fair presentation of the different methods, with pros and cons for each, instead of pushing RTK as "The One True Way".
Edited: 2012-08-30, 10:43 am
I've been having good success with using KanjiBox in conjunction with Henshall's book. When a kanji is easy to memorize, or is obviously built from other kanji or radicals, I just memorize it; otherwise, I look it up in Henshall to get some history and information on its components. I sometimes create a Heisig-like story to remember particularly devilish kanji. If a character contains some elements that only occur as inidividual characters in Chinese ("CO" characters), Henshall points this out, and explains the component's phonological and semantic use in the character. I love Henshall's explanations of how each character has morphed into its current form.
I've mostly been using the Kanji and Readings features in KanjiBox, and intend to follow that up with using the KanjiDraw feature (which I've only used sporadically up until now). KB groups everything by JLPT level, which has enabled me to work through the kanji in tiers. I completed all of the Jouyou kanji last week after about three months of intensive study, and am now working my way through the N1 readings.
I combine this with extensive reading, which helps me to remember the most important kanji in context.
Edited: 2012-08-30, 12:43 pm
I came to RTK already with an intermediate level of Japanese, and I was extremely frustrated at how the book doesn't teach the readings and how the meanings are sometimes way off. I tried to go through it two or three times before I gave up on doing it the 'Heisig' way.
After experimenting with other methods (including the movie method and the lazy kanji method) I decided to take parts from different methods and make them into my own(-ish). What I did was:
1. Go through a list of primitives (that someone had provided) and rote-memorized them.
2. Went through the jouyou kanji list and picked a more accurate meaning for each Kanji instead of the ones Heisig provides.
3. Picked one 音読み for each Kanji and added a sound indicator to each of the stories so I know how to pronounce them.
To learn the kanji, I use the Lazy Kanji method: kanji on the front, various English meanings on the back. Look at the kanji (<1 sec), look away from the screen and write it, give a ballpark meaning and the reading. This method is extremely effective for me, especially as I am able to quickly learn Kanji that I've seen already (i.e. in class, online, manga/ラインベ, etc.)
I learn the kanji by groups of onyomi, and I noticed that it's easy for me to learn how to pronounce them because of the radical on the right hand side that appears in several kanji of the same group. I eventually connect that radical to a sound, and I'm able to produce the onyomi even if I forget the sound indicator for the story. (I believe this is what warakawa was trying to say - it seems as if people tended to brush him off because of his harsh language, but what he says is true for the most part.)
Another thing people will say is bad about methods like this is that "you shouldn't go kanji to keyword" - I think it's alright to do this if you're going from kanji to meaning, instead of Heisig's keyword. Because how else are you going to learn Japanese if you don't know the meaning of the word??? You're going to have to learn it anyway, associating it with the kanji should make it easier to learn words down the line.
I didn't learn kanji in any particular order.
The first kanji I was able to recognize (hundreds of them!) were in fact proper names (film directors, actors, actresses, writers, models, towns), movie titles, book titles, etc. And long before I even started to learn Japanese.
I've always been interested in good books and movies and had a nasty habit to check the literal meaning of the original titles.
When I finally decided to learn Japanese, I did some research, learned about kanji (bushu, components, stroke order rules), learned Japanese bushu names, but never bothered to learn kanji in any 'proper' order. I relied on massive comprehensive exposure to texts (audio + transcript + translation) I liked or was interested in for some reason.
If I were to use Heisig and SRS (Supermemo or its clones - Anki, Mnemozyne, etc), I'm sure I would rather kill myself.
I've never had trouble understanding why '色 colour' and '色っぽい sexy' are related and use the same kanji.
They're the same kanji because the word いろ means both color and sexual related things.
Japanese kids do write kanji over and over sometimes, but they have other things they do also -- their schoolbooks have etymology, they do reading and writing exercises, and they're seeing the kanji in their other subjects as well. So even if they are writing the kanji over and over it's probably not an integral part of their learing.
Why is this even being bumped?
Everyone here should know very well that no non-Heisig related method of Kanji acquisition is as close to as efficient or fast.
Edited: 2012-08-30, 9:03 pm
LOL @ RTKers who don't understand trolling.