Postby sphere » September 28th, 2007 3:40 pm
hmm.. I don't know if it can be generalised like that (having no momentum to verify with a large test case). The best way is : find some place to learn it ;p
For hiragana/katagana, find a book or a chart with the strokes numbered (I saw one somewhere at the language school where I took lessons).
For kanji... well it is almost a life-long never-ending accumulation effort. But a lot of characters has overlapping parts as the same radicals (or "sub characters") are used repeatedly in a lot of characters. the way to write each sub character is *almost* always the same (there are prob exceptions. have yet to see a natural language which has no ambiguity or exceptions). Then one just have to learn the general order of writing sub characters (which prob follow the general order jemstone suggests, though it can be broken down further in the general ways characters are composed by it's sub parts... e.g. single (1 way). double (top-down, left-right), triple (top-down, left-right, left-(top-down), (top-down)-right, top-(left-right), (left-right)-bottom ....) it goes on and on....
(e.g., the most complicated I've learnt so far is (鬱陶しい) (can't even see this character clearly unless I crank up the font size a few times over) The first character is hierarchically composed of 7 subparts...
Although clear rational methods are fast ways to jump start learning a language, this kind of methods are often too simple when you learn more (and as you learn more exceptions to the "rule"). once you picked up enough "patterns", you will probably gain the "feel" to write most other characters even if you are seeing them for the first time.
just sharing my own exp. hope that it can be of reference.
regards