| HN Mirror

That app is one of the best I found, but it's sadly based on data that is flawed at fundamental levels IMHO. The even more sad part is that it's actually hard to find good data. In fact, the data that I would like to exist doesn't exist, or is so hard to find that I haven't found it. But there's probably some matter of taste in there, so I won't bother you with that, but I'll list my gripes with the data a lot (if not most) apps use.

- Jim Breen's JMDICT/EDICT. While very valuable, its "word priority marking" data is based on word frequency from old newspapers. That makes it both outdated and biased (newspapers don't use the same range of vocabulary as conversational japanese or books or other forms of expression)

- Jim Breen's KANJIDIC: It contains all sorts of readings for characters, some of which don't even appear in many japanese dictionaries, without any kind of differentiation whether they're rare or not. I also recently found that the Kang Xi radicals it contains don't all match the Kang Xi radicals contained in the Unihan database from the Unicode consortium (I reported the discrepancy to M. Breen).

- KanjiVG: Used in many apps, including Kanji Study, to show stroke order, and document radicals/kanji parts. Sadly, that last part is full of fantasy. To give a couple examples:

- The parts it gives for 息 are 自, 目, 心. Yeah, 目. Because, you know, if you remove the small stroke at the top of 自, you can see 目, but that's highly irrelevant. Furthermore, the only meaning listed for 自 is "one's self" except in characters like 息, 臭, 嗅, 鼻 it's "nose".

- The parts it gives for 専 are 寸 and a vertical stroke. Sure, there's barely a character for the top part of that character (many fonts don't have it, but it exists: http://glyphwiki.org/wiki/u24c14), but a vertical stroke? really?

Consequently, when looking at characters associated with a radical in Kanji Study, you end up with lists larger than they ought to be if you look at simple radicals like 目, because everything that contains 自 or 首 is listed as well.

Maybe that helps beginners in some ways, I don't know, but feeding them with useless information doesn't sound very useful on the long run.

Now, another problem I have with all these resources (and others you can find on the web) is that the most comprehensive data tends to be in English. Which is great for native English speakers, but really, studying a language in yet another language that is not native to you is not the greatest experience.

Finally, past a certain level of japanese (from my experience), it's better to use japanese-japanese resources. Because things don't map 1:1 between languages, and past a certain point, you get a sense of how some japanese words differ from what you would normally find in dictionaries to your native language. And it's then better to have other words defined in terms of those you already know than in terms of yet other not-quite-matching native words. Sadly, I haven't found good online (or in-app) japanese-japanese resources, except for Kanjipedia.

OTOH, I realized a while ago that I'm completely useless at translating japanese into my own native language. I'd take a japanese sentence I understand fully and be completely stuck trying to convey its meaning in my native language. This may or may not be related to the fact that my japanese abilities grew mostly through japanese than through non-japanese resources. Either way, translation is definitely a skill that is acquired separately.