Clean and simple. I like it. For a more advanced puzzle, you could also put kanji instead of romaji in the "word list" (basically it would be a test of the user's ability to remember how the kanji were read.)
One bug I noticed --- it is showing ぢ for "ji". Except in a very few cases, that should be じ. "zu" should also usually be ず (but づ is not uncommon, e.g. つづける).
If you're gonna keep working on it, EDICT (a Japanese-English dictionary) may be helpful --- the data is freely reusable for both commercial and non-commercial purposes:
http://www.edrdg.org/
Yea, the romanization rules are only approximate at best. It won't properly transcribe 'doozo' either. I'd like to support some more advanced translations, but I'm trying to take baby steps -- both with the language and the program!
I've visited EDICT once or twice so far, but didn't realize they allowed reuse of their data. I'll pursue that further!
I wanted to demo my weekend project, an online Japanese word search builder. I'm currently studying the language and have been looking for some more interactive educational material than the usual barrage of textbooks. I especially enjoyed the word search puzzle included in my hiragana book and thought it would be fun to recreate it. I've built out support for roman-to-hiragana translations so far, and plan to add roman-to-katakana soon.
As for the technical details, the site is built on Sinatra and hosted on Heroku (using their new Ruby 1.9 stack!). The front-end is jQuery.
Hope someone else finds it as useful and fun as I do!
Otsukarasamadeshita! Nice app! I find it pretty useful
What kind of input validation are you doing? It doesn't seem to prevent invalid characters. For example, I put in words with 'q' and 'l'--2 letters not represented in hiragana--and was able to generate puzzles
Hehe, that's actually a typo for komban! I just quick copied some vocabulary from one of my books to seed the word list and clearly didn't double-check my work :)
That aside, this tool is intended for character recognition than actual vocab study (at least in its current state). I'm hoping to add some new translation support eventually though (straight from english to kana+kanji would be nice).
