My Tsumego quiz program got a lot of action last week! Many people tried it out, and a dozen or so people gave comments. One person even emailed me directly, asking for a particular feature. So, time to stop working on the Japanese quiz program and get back to work on the Tsumego quiz program, right?
Well, not quite. First of all, this week somebody also emailed me about the Japanese quiz program, suggesting some improvements to it. But more to the point, these programs are my hobbies that I work on in my spare time, so I work on what I feel like working on, not necessarily what people are clamoring for! And I wanted to work on the Japanese program.
Last week, I mentioned that I’d done the final bit of refactoring to abstract out the nasty flat-file system so it could be more easily changed to be backed by a database. So, this weekend, I really wanted to get the database schema sorted out, and a solid plan for how things should look in the new, faster, easier-to-maintain world. Unfortunately, this was not to be. Once again, I ran into the problem of how to model the kanji radicals.
As I’ve mentioned before, each kanji character in Japanese can be broken up into smaller pieces. For example, the kanji for “time” is made up of two smaller characters, one for “sun” and one for “temple”. The kanji for “temple” is also made up of two smaller characters, one for “earth” and one for “inch”. Native Japanese speakers don’t typically think of these sub-characters any more than we think of the etymologies of the words we use, but they can be very useful for non-native speakers to use as mnemonics for both meaning and for stroke order.
Unfortunately, there are significant complexities when it comes to modelling this data in a programmatic way. I’ll talk more about these nasties in my next post.
Entries (RSS)