I spent the thanksgiving break in Lyon, doing great things. I profited to advance a bit in Japanese and especially in Kanji. Someone pointed me to the WordLookup program which is cool to get some information about a specific Kanji (you can even input the Kana, or the English word). The problem is that this program is slow and commercial.
I discovered that the dictionary itself is freely available, so I had this idea: why not writing my own dictionary UI? Well, that's what I did, while my girlfriend was watching the Va Vinci Code movie (which I don't really like).
The program is of course written in Ruby. It's indexing the JMdict version of the dictionary (a 40MB XML file!). I experimented Ferret for the first time, and it's a great library (modulo the fact that you have to separately analyze the non-ASCII strings, but it's not really Ferret's fault). Of course the parsing/indexing part takes some time (20 minutes on my Mac Book Pro), but it's only necessary once, and after that searches are instantaneous. I'm already using my tool to learn more Kanji :)
I will clean the sources and dump a tarball somewhere in the very near future. I still need to add an history feature though, without it it's hard to navigate in the dark world of Kanji.
Previous Topic
