In my last post, I talked about the composition of Japanese kanji characters, and how it’s been giving me difficulties as I try to create a good database model for my kanji quiz program.
Probably the biggest problem is that not every sub-character is a complete kanji character. The four I mentioned above—sun, temple, earth, and inch—are all kanji characters that can be found alone quite commonly. However, the grass radical (which can be found in kanji such as “tea”, “flower”, and of course “grass”) is not a kanji character itself. You’ll never see it on its own—there’s no Unicode character for it—and there’s not even a standard way of referring to it. One source might call it “ku-sa-n-mu-ri”, while another calls it “Bushu 140, Variant 2″ while yet another refers to it simply as “Element #1783″. How should my database refer to it? Should I give it yet another arbitrary number, or should I use one of the names somebody else uses?
Another problem is that, of these many sources I’ve looked at, none are complete. In fact, most of them are not only woefully incomplete, but in some cases simply wrong. So I not only have to deal with incomplete data, but I also have to deal with incorrect data, and ensure that whatever format I use in my own database, it’s easy to change or update when I come across incorrect data that I imported from elsewhere.
So, after puzzling through this problem all weekend, and attempting to drastically simplify all my assumptions and use cases so I could cut this down to something manageable—even if it had to be reworked significantly later—I find myself no closer to my final goal of having a working database for my kanji quiz program. I’ve caught myself going down numerous dead ends, realizing the flaws in my implementation or data model, then heading down another path that failed for different reasons. I keep trying to carve off pieces of the headache-inducing problem, trying to get down to a smaller and smaller piece until I’ve finally got something small enough to chew, but I still end up with something too large for me to reasonably tackle on my own.
I feel like I’m getting closer—every so often I can catch a glimpse of the light at the end of the tunnel—but then I crash into a wall I had forgotten about or a new wall I hadn’t yet encountered, and I wonder exactly how far away I really am.
Entries (RSS)