Had a great time Saturday, it was fantastic meeting everyone again/for the first time and seeing all the great boards.
Obra stop making me want all your keyboards. :< First the butterfly now that symmetric stagger one, want both so bad rofl.
Oh, I found an article on the layout used on the Japanese keyboard with the thumb shifts, I totally forgot the name of the guy I talked to who had it because I am bad at names, but:
http://en.wikipedia.org/wiki/Thumb-shift_keyboard (And hah, I was right on the right long key being space. Reasonable place for it given the language.)
So it's a little like we were thinking, and there's actually a couple of versions of thumb shifting, but I think I recall yours having 半濁点 keys and not having three kana marked on any so it's probably more like the original and not the NiCOLA variant(Which actually makes a little more sense, but).
So we'll take an example key, same one used as example here:
http://museum.ipsj.or.jp/computer/word/0006.html (There's an English version of that page but the images are broken so I'll just explain it here.)
So, は on the bottom and み on the top. This one's a good example since there's actually four Japanese characters you can type with this key: は(ha), み(mi), ば(ba), and ぱ(pa). So this key is on the right side of the keyboard, which affects which shifts you use for what. So to type the bottom character, は, you just hit the key. To type the top character, み, you hit the right shift and the key. For a key on the other side of the keyboard typed by the left hand, you would use the left shift to shift to the secondary character.
There's a very deliberate pattern to which characters are on top and which are on the bottom - Besides frequency of use, none of the top characters accessed through shifting have a voiced or semi-voiced version. So here's where cross-shifting comes in. Some characters in hirigana(and also katakana) are formed by taking another character and adding dakuten(voiced) or handakuten(semi-voiced) marks to them. These look like double quote and degree symbols respectively, but the main thing is it's diacritic marks, like adding an umlaut to an o changes its pronunciation. Voiced characters are typed by using the opposite side shift and the base character - So left shift and は give us ば.
The fourth character comes from the keys where shift is on a western board in the original implementation. One of those and our は key gives us ぱ, the final Japanese character here. There are far fewer of those, just the 'h' series of kana(は, ひ, ふ, へ, ほ). The semivoiced versions barely appear in native Japanese and are mostly found in loan words. The NiCOLA variant drops the handakuten keys and moves those five characters to keys where the bottom character doesn't have a voiced version, where they're accessed just like the voiced characters, with the opposite shift.
Looking at it it's actually an absolutely brilliant system of doing it rather than JIS's mess or worse typing romaji. Sort of a Japanese equivalent of using something like Colemak or Dvorak instead of the mess of Qwerty. (Shifting with the thumbs is brilliant for the language, as space isn't used very much compared to western languages so there's no reason to have a big huge button in prime real estate. Not that there is for English either. The spacebar is a revolting waste of space.
(Hmm, center a shift and you can shift most of the keys on the board with it and one hand. I spend so much time in keyboard layout editor.)
