Filled with curiosity

[Oriphia]

I'm helping a friend developing his new Chinese IME, it is not basing on romanization, but the structure of characters, so someone can type a 汉字 without knowing its pinyin.

So my question is: as a non-native, how do you type a character you saw on papers and have never seen before?

1. Seek pronunciation in dictionary, and then type with keyboard.

2. There are already IMEs other than pinyin can solve the problem.

3. Someway else.

I'm keen to get your opinion on the matter.

Thanks so much.

[skylee]

There is a very popular input method called Cangjie which is based on the shapes of the characters instead of the pronunciations. Take a look -> http://en.wikipedia.org/wiki/Cangjie_method

There are easier versions of Cangjie. The users are only required to input part of the codes of a character and then choose the right one from a list of characters which match the codes.

There is also a stroke input method called jiufang which breaks all the characters into nine basic components and require the users to input using the number pad only.

The T9 input method, which is used on mobile phones, is even easier. You input the characters by simply keying in the 5 basic strokes following the stroke order.

There are other non-pinyin based input methods like "Wubi". If you search these forums and the internet you would find them.

[Oriphia]

Thinks for the helpful information, I know a little about such IMEs, in fact, I don't think Cangjie and Wubi are suitable for newbies, my friend have an idea of enhanced T9, it's no need to input strokes, but character radical instead, different from Wubi, it is no need to remenber radicals on the keyboard at the beginning, ... I don't know how to draw a picture of it, but I'm sure it is easier to kick off, and will be faster than T9 when you are experienced.

At the moment, we have no confident that what kind of user would like to try shape/stroke order basing IMEs when they encountering new characters. We don't sure if speed is more important, or maybe usability do. If we know what you thought, we would able to revise it on a right track, for example make it more friendly to elders or speed up for youth.

:wink: :wink:

[AlexBrit]

Sounds like a worthwhile project Oriphia.

I only use the pinyin IME on the computer keyboard. I use a pinyin input on my mobile phone also, and have tried the 'stroke order' method on the mobile too, but it is slow.

If you guys could develop a simpler method I'd be impressed.

My question is, do you mean you are developing a method that requires the user to be able to recognise the radical of the character?

[Oriphia]

His idea is to show all the first radicals on the screen, while input the first radical, radicals replace, and all the optional second radicals will show on the screen combining with the first radical you have input, then the third...

For example, someone would like to input 想, seek and input the first radical 木, and then radicals changed, a new radical 相 show on the screen, after one more beat, 想 show on the screen. So it is no need to remenber any radical at the begining, since each radical is fixed on the certain key, blind typing is possible. He expect one can reach the speed of 20-30 characters per min. after 3 month training.

Another feature of the IME is, both Simplified and Traditional radicals are supported.

Does it work?

[imron]

What you propose sounds quite similar to 五笔字型, but 五笔字型 maps several different character "roots" to the one key, and also has a few completion rules so that you need only 4 keystrokes to type one character, regardless of how complex it is.

What I'm a little unclear of with what you propose, is that if there are only 26 keys and close to 200 radicals, how do they map so that blind typing is possible? Also, without needing to remember which radicals map to which keys, does that mean you will show these 200 radicals on the screen at the same time? That doesn't sound very effective from a usability point of view.

Also, 20-30 cpm after 3 months doesn't seem that great. If you spend that much time training in the other character based methods, I'd imagine you'd have a much higher typing speed.

Anyway, I don't mean to seem too negative, it's just that I don't see what advantages this method has over what is already out there. If I've misunderstood something, please feel free to point it out.

[Oriphia]

To imron:

Exactly, that is what he's facing now. The key issue is how to make a balance between efficiency and simplicity.

五笔 is exactly the most efficient IME I've met, some senior typist can rush 100 characters per minute, that is marvellous!!!! But it's not a easy thing everyone can manage in 2-3 months, 拼音 can also reach 20-30 cpm, but it will be helpless when one frustrated to pronounce the character, it always happen.

I don't know my friend's IME whether success in encoding the radicals, I think your question is great, I will forward your suggestion to him, much thanks for screen out the question.

PS: Anyone have tried Chinese Handwriting Pad with OCR software? Compare to IMEs, pro's and con's?

[skylee]

I use handwriting recognition to input chinese to my pocket pc and I think it is fairly painful. Painful because I don't really write any more except on the handheld and it is quite tiring and slow. That said, the recognition accuracy is pretty good. But I would have liked to use pinyin input if my handheld supported it.

I would only use an input method that builds on what I already know (e.g. stroke order / pinyin) and is really easy to pick up (e.g. T9). Personally I would reject any input method that requires a lot of techniques and practice (e.g. Cangjie).

[AlexBrit]

Ok, thanks for explanations Oriphia. I understand. It looks interesting.