If that is the mapping for the dialogue font, things should be fairly straightforward (it's only kana, romanji, and some symbols, which will yield a table not much longer than the one for the Cross relative alphabet). Unfortunately, my guess is that the actual Japanese dialogue font is that butt-ugly monster with all the kanji on the next page of that thread, which means that those kanji are going to have to be examined to see if they're in any kind of standard order. If they aren't, each is going to have to be ID'd individually.
It's possible that, in order to obtain the entire script, we are going to have to map both character sets.
I'll try to compare the stuff in that thread to SHIFT-JIS and UTF-8 character orders later today (assuming that I can find that information--I know I used to have it for SHIFT-JIS) and see if there are any commonalities.
Update: I am now 98% certain that the order of the kanji in that one font is completely random, just like the rest of the characters in it (no major encoding that I can find places 人 王 山 一 near the front in that order, with or without other stuff in between). Gack. I hope that isn't the one we need.
Another update: I should have read the thread utunnels linked more carefully, as there is a conversion table already posted there that appears to cover the dog's-breakfast-kanji font. If it's accurate, it will make things a lot easier.
(Is it safe to be discussing this here? It isn't quite hacking, but we're dancing kind of close to the edge...)
