Practically anyone that types Japanese on a computer uses an IME to type the pronunciation in kana (latin letters or kana directly) and then converts words to kanji. I don't like having to pause and constantly check the conversion when I write. It slows me down and is plain annoying. So, I searched around and found there have been a few attempts to develop a way to input Japanese directly.

I found three different systems: NIK-code, T-code and TUT-code. NIK-code is based on the idea of type kanji by their components like when writing by hand (each component is assigned to a key). T-code and TUT-code is based on the idea of inputting a kanji by only pressing two keys (allowing 1600 combinations, more by extending to three keys).

I have slowly learning T-code for over a month and loves it. Here is some differences I have noticed between using an IME and T-code.

Learning effort
Learning to input kanji by typing their sound with latin characters require little effort beyond memorizing the conversion chart for the kana. Possible problems comes if you have to use a keyboard layout you are not used to.

Typing kana directly and then converting kanji requires memorizing the kana keyboard layout, but nothing more.

T-code requires a lot of memorizing, since the key combination for producing a character is entirely based on its frequency.

Speed
Since inputting by sound requires conversion for producing kanji, you must stop now and then to trigger it. Other issues that slows down input is if the suggested kanji is not the wanted or that the IME have defined the word boundaries wrong.

With direct input the typing speed is only dependent on your typing speed and how well you know the key combinations.

Touch typing (aka Blind typing)
It is theoretically possible to touch typing with an IME, but it requires a lot more memorization. You must know not only the pronunciation of a word, but also were it is placed in the list by default (must disable reordering by how frequent you type a word).

Touch typing with T-code is essentially no different from learning to touch type a language like English. The main obstacle is to produce characters that are not included in T-code, but that can be solved by extending with three key combinations.

Links
Home page for TUT-code (Japanese)
T-code home page (Japanese)
Brief description of T-code (English)
A lazy man's guide to easy typing (Parts 4-5 are about Japanese, English)

Same way as you type the kanji. Kana, numbers and special characters are also typed with a two key combination.

If you use Windows, and you have a Tablet PC or purchase a pen input device for a regular PC, the handwriting recognition (both English and Japanese) in Vista is great and in Windows 7 it's absolutely fantastic. I don't bother with it much for English, because it's quite slow to handwrite compared to my ~60wpm typing speed, but for Japanese it can be competitive with typing and is certainly more natural than all the keyboard-based direct input systems you describe — as long as you know which kanji you're looking for.

Using a pen :S? please. Sounds like a complete waste of time. Japanese people write notes in kana when they really want to write quickly. I doubt you can write kanji faster than selecting the kanji after typing out its reading.

That being said, it sounds like an interesting system, but i think memorization can apply with IME too, i noticed it myself after typing out a couple of things in word.

joxn_costello Wrote:I don't bother with it much for English, because it's quite slow to handwrite compared to my ~60wpm typing speed, but for Japanese it can be competitive with typing and is certainly more natural than all the keyboard-based direct input systems you describe — as long as you know which kanji you're looking for.

I had forgotten about handwriting recognition, but as you say it is considerably slower than typing. It can only compete with typing in Japanese precisely because the conversion step (in an IME) interrupts the typing and requires you to think.

The idea behind these input methods is that, by making a lot of upfront effort, you can type Japanese faster than before. Faster because there is a fixed number of keys to press for each kanji, while a kanji have typical more than two strokes and its furigana is usually longer than one kana. You essentially don't have to think about anything like readings or stroke order.

The advantages of that system are if you know a kanji but not the pronunciation, you can type it no problem...

The disadvantage is that if you know the word and not the kanji (which mostly happens with me), you'd be at a loss of what kanji to type in.

Also, slightly off topic, but can any users who have experienced both QWERTY and JIS keyboard layouts give me pros and cons of both? I'm looking towards getting a new computer (prob MacBook Pro) and am trying to decide which keyboard layout to get.

The Japanese layout would involve completely re-memorising a lot of the keys - I'm a pretty fast typer, and I don't like the idea of my speed being effected when typing Japanese. However, I make a lot of mistakes when touch-typing Japanese, which I think could be eliminated if I switched to a Japanese keyboard layout.

Any thoughts?

julz6453 Wrote:Also, slightly off topic, but can any users who have experienced both QWERTY and JIS keyboard layouts give me pros and cons of both? I'm looking towards getting a new computer (prob MacBook Pro) and am trying to decide which keyboard layout to get.

Any thoughts?

Is the JIS layout the one with all the hiragana on it? If so then I have that on both my work computers and hardly ever notice a difference between work and home. Aside from putting in " and ' and : I can't usually remember which one I'm using. I think it took me a few months to get used to the new location of those keys. Oh and the smaller space bar. That was annoying for about a week.

Ok, thanks - I know a lot of users were complaining about the smaller space bar and the different control keys.

Another question is, when typing Japanese with the direct hiragana input (since I'm assuming that's how most users of JIS keyboards type), did it take a while to learn the location of all the hiragana? That's one of the main things I'm apprehensive about.

julz6453 Wrote:Ok, thanks - I know a lot of users were complaining about the smaller space bar and the different control keys.

Another question is, when typing Japanese with the direct hiragana input (since I'm assuming that's how most users of JIS keyboards type), did it take a while to learn the location of all the hiragana? That's one of the main things I'm apprehensive about.

Actually, I'm fairly sure the most common way of writing in Japan is the same as elsewhere. I've often heard that one of the reasons Japanese people learn romaji in school is because they use it to write.

I see. I guess the reason I'm asking is to figure out which way is faster for writing Japanese - direct hiragana input (once you learn the new positioning), or romaji to hiragana conversion (which is relatively fast for me, as I can type fast in English).

Thanks for all the help so far.

julz6453 Wrote:I see. I guess the reason I'm asking is to figure out which way is faster for writing Japanese - direct hiragana input (once you learn the new positioning), or romaji to hiragana conversion (which is relatively fast for me, as I can type fast in English).

That's like asking which is faster: QWERTY or Dvorak. Maybe with enough practice direct input can be faster. Then again, maybe not. But it's a moot point because almost no one uses direct input. Sit yourself down at a computer that's not your own and you'll be reduced to finger pecking an a ridiculously slow rate. Makes you feel illiterate.

There's an IT guy I knew at my old work who swore by Dvorak. Said it made all the difference in the world. But he had to carry around a specially-made Dvorak keyboard that he could plug into computers he needed to work on, because otherwise his QWERTY finger pecking was just too damn slow. Somehow I just never saw where the efficiency gain was.

Well, it's quite easy to realize that direct input is faster since you only have to type 50% as much. However, like mafriend said, it's all about getting used to it and it might not be worth it. I do not plan to work with writing Japanese on a computer so I don't need it even though I'm an IT guy and I plan to work in Japan... If you want to be a Japanese secretary, it might be worth the effort though.

mafried Wrote:There's an IT guy I knew at my old work who swore by Dvorak. Said it made all the difference in the world. But he had to carry around a specially-made Dvorak keyboard that he could plug into computers he needed to work on, because otherwise his QWERTY finger pecking was just too damn slow. Somehow I just never saw where the efficiency gain was.

You don't need a dvorak keyboard to use the dvorak layout.
Just change the OS settings and touch type.

Tobberoth Wrote:

julz6453 Wrote:Ok, thanks - I know a lot of users were complaining about the smaller space bar and the different control keys.

Another question is, when typing Japanese with the direct hiragana input (since I'm assuming that's how most users of JIS keyboards type), did it take a while to learn the location of all the hiragana? That's one of the main things I'm apprehensive about.

Actually, I'm fairly sure the most common way of writing in Japan is the same as elsewhere. I've often heard that one of the reasons Japanese people learn romaji in school is because they use it to write.

I've never met anyone who can or does type in direct kana input. It's just a holdover from typewriters afaik.

Kana method is faster for cell phones, Romaji method is faster for keyboards (IMNOHO)

Jarvik7 Wrote:I've never met anyone who can or does type in direct kana input. It's just a holdover from typewriters afaik.

Pretty much everybody at the place I work at (in Japan) types using kana input.

I would like to see some figures about the input preferences as I haven't met any Japanese people who are good enough (at romaji<>kana) to use the romaji input.

What is there to be good at with romaji-kana conversion? In both you still have to use an IME to select the final output. Everyone knows that eg. か is written with ka in romaji...

I find it hard to imagine that you've never met anyone who uses/can use romaji input. Perhaps you misunderstand what kana input is?

I haven't found anybody that uses kana input. Granted I'm in not in the clerks office, maybe they use it there and maybe it's faster but I doubt it since none of my coworkers do.

mistamark Wrote:

Jarvik7 Wrote:I've never met anyone who can or does type in direct kana input. It's just a holdover from typewriters afaik.

Pretty much everybody at the place I work at (in Japan) types using kana input.

I would like to see some figures about the input preferences as I haven't met any Japanese people who are good enough (at romaji<>kana) to use the romaji input.

What are you talking about, every Japanese person above the age of 10 should be perfectly fine at using Nihon-shiki...

Jarvik7 Wrote:What is there to be good at with romaji-kana conversion? In both you still have to use an IME to select the final output. Everyone knows that eg. か is written with ka in romaji...

Yes, you still have to use an IME for both... but direct kana input requires about 50% less typing, like Tobberoth said.

I would also like to see some statistics about romaji->kana versus direct kana input. There seem to be some pretty differing opinions about it.

What I'm really looking for is evidence whether typing Japanese with direct kana input is faster than with romaji->kana input, once the key positions are learned obviously. Any opinions from people who have switched from one to the other would be greatly appreciated. Thanks for all the comments so far.

Just out of curiosity, why do you guys care so much about typing speed?

In my experience, typing speed isn't an important factor compared to the time one spends thinking about what to type next.

It only makes a difference for some specialized jobs where you have to do a lot of dumb retyping.

For me, typing speed is everything - I can normally think of what I want to say long before I ever get it down. That's why I hate writing on paper - my writing speed is so much slower than my typing.

Even though I don't use it for professional situations (unless you count school...), fast typing is a necessity for me. That's why I want to try to find the fastest possible way to type in every language I frequently type in, which happens to be English and 日本語.

I'd be curious how much more efficient the kana input is. More kana can be typed directly, but at the cost of some of them being farther away from the home keys, and still extra keys need to be pressed to get dakuten & handakuten, plus shift for little kana.

julz6453 Wrote:... but direct kana input requires about 50% less typing, like Tobberoth said.

Actually that's not at all clear (EDIT: vosmiura, you beat me to the punch). There are too many kana to fit on the top, home, and bottom row of the keyboard, necessitating some of the less frequently used kana to be placed on the numbers-row of the keyboard, or otherwise in positions too far to be reached from the home-row touch-typing position. Reaching these keys significantly slows down typing performance. Unless someone can find some stats on this, it's an open question as to which is faster.

But, if you *really* want to go your own way, I'd recommend a combination of romanji IME and the Colemak keyboard layout. Colemak puts the a-i-u-e-o on the home row, which should significantly speed up a syllabic language like Japanese...

Tobberoth Wrote:What are you talking about, every Japanese person above the age of 10 should be perfectly fine at using Nihon-shiki...

No need to get angry I'm just telling you what I see every day at work. Also if I ever need to use the shared computer, it's always set to direct input.