Old thread, but still a valid question.
I'm about half-way through and I should finish in another 3 weeks or so. I'm unusual since I have some Japanese (from 20 years ago, mostly forgotten) and right now I have nothing else to do (no job, school) so I'm spending whatever's left from 24/7 after subtracting sleep, food and other things that distract from study.
I wanted to make a couple of main points with this post:
1. Continue with flashcard revision
I reckon that you should continue with the flashcards for at least 1 month after completing the book before you can say that you actually remember everything you've learned. It would probably be good to continue for longer, even if it is at less regular intervals (say weekly after that for a couple of months, then monthly after that)
2. Memory "Pathways"
People talk about going kanji -> keyword instead of keyword -> kanji (what Heisig advises). I think that can work, but you have to be careful. For one thing, Heisig's keywords can be kind of arbitrary, so the temptation will be to forget the keywords after a period of using the kanji -> keyword method.
The way I see it, the brain will try to make it easier for itself to file everything away. If, while you're doing the keyword -> kanji study that Heisig recommends, you start doing other stuff like keyword -> kanji or learning kanji meanings (in compounds or individually), your brain will start to decide to dump the very thing that Heisig's method is there to teach you to remember (namely keyword -> kanji mappings). It's like if you see 勉強, your brain is going to just go "べんきょう: study" and short-circuit any activations of the part of memory that involved looking up "exertion" and "strong". After time, these memory circuits will atrophy for lack of use and at least some of the effort you've spent on RTK will end up being wasted.
So even though I haven't finished the book, I recommend making sure that when you transition to learning kanji in practice (whether it's kanji -> keyword or kanji in context in vocabulary) that you "complete the circuit", ie, do at least as much revision of the keyword -> kanji memories as you're doing for the other new stuff.
Once you have keyword -> kanji memories really solid, you can always modify them. Keywords can be arbitrary, but at least you can post mental post-it notes on them to give you pointers about how they actually are used, compounds they're used in or whatever. The great value of English keywords is that you can build up a web of associations (stories, etc.) even if you can speak very little Japanese.
3. Transitioning to Japanese (translating the system)
The main question, then, about the utility of Heisig's keyword -> kanji system is what happens after you get better at Japanese. Do you just let those memory pathways atrophy (treating them like a temporary crutch that you throw away once you're fit again) or do you try to incorporate them into your new Japanese way of thinking? I think that the latter is better.
Those keywords that Heisig chose will become increasingly irrelevant, but I think that there's no harm in remembering them. A lot of them will end up forming memories like "Oh yeah, Heisig called this X, but it's really more like A or B or ...", but at least that's a kernel that you can form other memories around and (as it's effectively a handy index you carry around with you), it gives you an easy way to trigger recall of the relevant memories.
Later on, you'll overlay more properly Japanese interpretations on your initial stories. You'll have to overcome the homophones, so you can't just translate the keywords directly. For example, say you've got rain and sweets (#2817), but they've both got the same sound (あめ), you need to index them differently, so rain -> rain, like weather -> 天気の雨 and sweets -> sweets, like food -> 食べ物の編め. Eventually, you want to be able to build up two different kind of indexes (one purely in English, with unique keywords, the other in Japanese with some sort of disambiguation built in), both interlinked. You'll probably also want to start translating your stories to Japanese. If you keep the original English keywords, this should become much easier.
If you look at the way Japanese people themselves handle the memory problem with kanji, you'll see that they take a similar approach, except that it will be more based on vocabulary. By that, I mean that if they can't remember the kanji to use in a particular compound, they'll ask someone "is that the X that's in XY or XZ, or what?" When you don't have that sort of vocabulary to use as a reference point, then the next best thing is to use your knowledge of kanji and how they inter-relate (yeah, it's made up of this bit of that, and has a similar meaning to the other---yeah, that's it).
4. Why?
The most basic reason for going through all this is that there's a fundamental literacy gap that we, as foreigners, face when learning Japanese. Without kanji, so much is cut off from us. It may seem a bit pointless having to learn them all (or at least all the Jouyou set), but I think that can be seen as a positive: it's just one pointless thing to get out of the way and then we can start having fun learning to speak and read and write just like a normal person...
