Writing
this other lengthy post helped me to sort out my thoughts on Chinese characters and increase my understanding. Shortly after, I had a mental break-through: as many characters are composed of other characters it seems that two different methods of learning are necessary: one method to memorize the shape of a non-composed (atomic) character; and another wholly different method to memorize which parts make up a composed character. Some of the recurring parts of characters are not characters themselves and they make learning harder, because you have to learn the part while also learning the composition. The key insight is to factor out those parts, give them invented names and memorize them seperatly!
This idea revolutionized my learning of the Chinese script. While the thousands of characters still remain arbitrary, there is at least an optimal, non mind-numbing method to memorize them. Don't look for a deeper meaning; just decompose and memorize along the actual shape of the character. Realizing this and using it in practice I also realized how uneffective the "etymological method" of teaching the characters is: first of all, you learn sometimes have to learn components of a character that used to be components in earlier forms, but are not any more represented in the current form. This way, you first learn a wrong character and then add the correct, modern version without any further story to memorize the modern component. You only know that it is somewhat similar and simpler than the original one, but that's not precise enough to memorize it. It is especially ridiculous when comparing it to the length and elaboration of the story explaining the old, now wrong version of the character. Furthermore, the etymylogical approach does not help to memorize small differences between characters which are very similar by appearance, but have a completely different ethymology. I
already complained how this approach makes it hard to locate a character in a dictionary. It's a trial and error process that's completely unnecessary: character lookup should be based on shape alone, since that's all that the user has when he needs this function at most!
Right when I had this insight to split my learning into memorizing of basic shapes and memorizing of of compositions, I thought that I should turn my notes into an interactive website just like zhongwen.com. Using Unicode it is technically trivial now to display characters on the web and typing them up on a website with links is not much harder then typing them into a notes file as I currently do it.
I had already started to plan the design of such a website when I thought that my idea being so obviously great, somebody might have had the same idea already. And indeed! No long websearch was necessary to reveal the wonderful book
Tuttle Learning Chinese Characters Volume 1: A Revolutionary New Way to Learn and Remember the 800 Most Basic Chinese Characters I bought it, and I really like it. This book takes the only reasonable approach: no matter how insane and complex the Characters are, just look at what's being given --the character shape-- and memorize it using the most efficient approach. If I should ever seriously try to learn the Chinese script, I will do it using this book. And I will use the same method for learning characters which are not in the book -- using zhongwen.com and other sources to research the character components.