To catch all the text, I like to dump from the first string in the pointer table to the end of string control code of the last string in the pointer table. Find the the last string (usually the one right before the text begins), go there in rom and then search down for the end of string control code. There ya go. That should grab all of the text in that particular block.

I also try to skim through the dump and check for any stray <$XX> hex codes that I haven't figured out and see if I can't do something with them. In general though, I like to try to emulate the way the strings are displayed in-game as much as possible.

One way to see if your dump is easy to work with is to "translate" a few of the strings (change to English text) and insert them. If things go smoothly now, you can be pretty certain they will later when it's time for the real insertion. Of course, after you get some practice and experience with having other people work with the dumps under your belt, you'll start develop an idea of what is smoothest for both you and your translator to deal with.

That's another thing, let your translator tell you want he thinks of the dump before he starts on it. Usually, they'll be spending way more time with the script dump than you are, so it pays for it to be easy for them to read and work with.

Makes sense, thanks a lot!

My big trick now will be how to deal with Itadaki Street's font moving all over the place hex wise. The intro has the characters in one spot, the levels have it in another, and each level also moves 5 or 6 characters around throughout... it's really odd not sure why they did it. I'm guessing I'll just have to make multiple tables which I'm working on and then somehow figuring out when each is used to get a good dump. Won't be hard to keep the english values consistent when I re-insert but the japanese... eesh.