- TECH SPEX -




Step 1 - Finding the Font

The first steps were taken when Disnesquick removed the font graphics from the ROM, which somehow found their way into my hands. (get font graphics) There are approximately 6,000 characters in the font, however, upon closer examination I discovered the font is actually just a standard Shift-JIS ordering of kanji. Instead of having to make a 6000-entry table by hand, I coded a program that would build one for me. (get font texfiles)




Step 2 - Preliminary Decoding

Darkforce, a good friend of mine, helped me find the location of the script in the ROM. I coded a text dumper to see what I could see, but there was very little legible japanese in the script. After much hacking and hex editing there, I narrowed the unknowns down to two main obstacles - a form of LZ compression, and various control codes. (get preliminary notes) I gave up when the compression algorithm got over my head. (get Demi's bullshit attempt) Over the next two months, I sat and waited until two men helped me on this. (get decompressor sourcecode) The first legible script was decoded on November 28th, at just over 600k. (get Shift-JIS decoder sourcecode)




Step 3 - Control Codes

Now that we could read the script, it was time to decipher the remaining garbage. This is slow work, because you have to run the game again and again, trying to notice what's different after changing a certain byte to a new text code. There are 40 codes to be isolated. (get control code sheet) When this sheet updates, my Shift-JIS decoder will update accordingly. Once all 40 were isolated, it was time for the Japanese translators to do their thing.




Step 4 - Translating

In a clever lure to attract more and more people to my cause and hopefully hire a translator or two, I released a "teaser" patch on August 15, 2002 with the first ten or so minutes of the game translated. Actually it was pretty much just begging for help, but hey, what are you gonna do. Anyhow, it was only then that I started to receive some serious offers. Don't get me wrong, I would get emails from potential translators all the time, but very few were actually willing to sit down for more than a week and tackle any significant amount of text (the script is about 700 pages, which I can admit is somewhat intimidating). The best guy who approached me is now a good friend of mine, who goes by the nickname MO. He managed to translate and edit about 150 pages for me while his girlfriend was over in Europe and needed something to keep his mind off things, which I can understand completely.

But, the majority of the translation fell on my shoulders. I have a BA in Japanese, and I've spent three or so months in Japan a couple of years ago, but I wouldn't exactly call myself fluent. I relied heavily on dictionaries such as EDICT (supplied with JWPce, the Japanese text editor I used for Radical Dreamers) and Japan's Goo online dictionary. However, these left me hanging sometimes, and occasionally I would hit a tough spot in the translation. When this happened, we would discuss them on MO's wwwboard (a great place for translators to hang out, see link below) and try to figure out the best possible way to convey what we thought was being said.

In the end, it all worked out fine, I'd say. I'm much better with Japanese now than I've ever been before thanks to all this. I translated the script's final Japanese sentence on the chilly afternoon of March 12, 2003. From there, only a few more bits remained to be edited. Once that was out of the way, private beta testing began, and soon after that we had the first formal public release! Wooo#@!$$





Step 5 - ROM Adjustments & Script Reinserting

By June 2002 I had refined and readied the script decoder (a program that takes the script in the game's own format and turns it into a readable, editable textfile) and reencoder (another program that does the reverse). I had to learn 65816 assembly in July when noone came to my call for help with some low-level coding issues, but I didn't run into anything too difficult. I'd really like to mention Gideon Zhi here; his DMA textfile really helped me with learning SNES assembly. If more of the experienced translating community was to start making tutorials as assible as this, I know it would help out the learning curve in the scene immensely. Anyhow, within about two weeks, I had a new font system in place, and managed to make it so that the script didn't even need to be recompressed when being reinserted back into the ROM. I really lucked out, being able to avoid all the compression issues completely.




Step 6 - Beta Testing

I chose a few guys, known for their continued support of the translation scene, to be my beta dudes. I set up a board, having them post the mistakes they saw. After a few runthrus, we were ready to release!