How fast is wubi

Radiolab is supported in part by the Alfred P. Sloan Foundation. Enhancing public understanding of science and technology in the modern world. For more information about sloan, at www. Jad: Science reporting on Radiolab is supported in part by science sandbox, a Simons Foundation initiative dedicated to engaging everyone with the process of science.

Jad, Radiolab. Back to producer Simon Adler. Simon: Before the break, Professor Wang had seemingly solved this massive technological linguistic challenge and saved the Chinese character. Jad: Thinking back to the beginning when you took me to that cafe Simon, and we heard about all the different ways people were using the keyboard in that Hong Kong Starbucks.

How did we get from Wang making his method to suddenly infinite ways of typing? Simon: First of all, while Professor Wang really cracked this thing open, he wasn't alone.

There were others who had been hammering and chipping away at this problem as well. From the beginning, you had a few variations, a few different ways to type. However, after Wubi, things do really explode because underlying Wubi was this subtle but spectacular departure. Tom: The keyboard changed from something where, what you typed was what you got, to a system where you were telling the machine certain features or characteristics of the Chinese character that you wanted on the page or I guess on the screen.

Tom: It seems like a minor distinction when you say it, but once you do that, once you have entered into a reality in which A is not equal to A-- I pushed a button that has a little symbol A on it, and I no longer expect that symbol to appear on the paper or the screen.

Effectively, I can set the letter A equal to any property of the Chinese character that I want. Simon: A could equal that water component or that work component or something far more abstract. Tom: Anything goes, and so in the early s, different ideas about how to do this started to flood in. Simon: Just a couple of quick examples here. Some of these broke the characters into components that looked like English letters. Jad: Does that mean you could look at the characters and be like, "I think there's a D in that picture.

Then placed those components on their English look-alike key. A represented a mountain peak looking component. Others looked to English spelling. The component for tree was represented by the letter T. Others had you input just what was present in the four corners of the character. Then going even further afield, some of these don't even use letters at all, they just use the numeral bank of the keyboard.

You know that square number pad on the right side of most keyboards? In essence, every character was given its own numeric code that you would tap in there, , dog, , fire, almost like a clerk ringing up vegetables at a grocery store checkout.

We're just scratching the surface here. Jad: It started to dawn on me what you mean when you say if we go to that Starbucks. Everybody would have their own preferred way of going from those 26 Roman letters to the thousands of different Chinese characters. Simon: Ming says at one conference he attended, someone actually had to be thrown out because of a fight. Simon: What they were fighting and arguing over was just like with the typewriter way back when, speed. Every single new input system, the inventor claimed we haven't achieved maximum speed yet and that my system, it's easier to use and faster.

One way they went about this, pushing the limits of the speed was by trying to predict what it was the typist was trying to say. Tom: Both predictive text and auto completion were anticipated in Chinese information technology decades before they were in English language computing and new media. To get to the character you want faster and faster. Simon: The way this began was you'll be typing in the components of the character, but before you'd finished typing them all in, it would guess what it thought you were going for and offer you a couple of options.

Tom: They would give you those options ranked by the probability that this is the one you want. Tom: Almost immediately, people started to think about next character suggestions. Simon: Predicting and suggesting not just the character you were trying to type, but also the next character, the next word you were going to type. Tom: If someone types in the character bei meaning north, it is a very high likelihood that the very next character is going to be jing for Beijing or maybe beifang for northern.

I'll give you that as a suggestion. Simon: Keep in mind, this is the s, a full decade before we had anything comparable here in the United States. Anyhow, right as all these technological changes were taking place, the Chinese language itself changed.

Reporter: Tomorrow, ABC News will begin conforming to the Chinese standardization of its languages, spelling and pronunciation. Pinyin it's called. Tom: Pinyin is a way of using the Latin alphabet to spell out the sound or the pronunciation of Chinese characters and words.

Reporter: The big advantage of Pinyin is that it more accurately reflects the actual Chinese pronunciation of a name or a place. Pinyin had been around for awhile, but in the s, right around the time professor Wang saved the Chinese character from the threat of computers, the Chinese government started to prioritize Pinyin in the classroom. Tom: So that when a Chinese kindergartner begins developing literacy and reading and writing, they learn Pinyin at the same time or even earlier than they start to learn Chinese characters.

Simon: These computer scientists who had spent years trying to figure out how to visually relate Chinese characters to the letters on a keyboard-.

Tom: -they think to themselves, basically we have the Chinese educational system teaching a way of relating the Latin alphabet to Chinese characters. It would be kind of foolish not to exploit that. Simon: We should start inputting characters by typing their sounds in Pinyin.

Now of course, professor Wang was staunchly opposed to this. Yang: When we use Pinyin to type, we lose sight of the Chinese characters form and the form is the soul of a character. It's like you're grabbing hold of a person and doing a way with their flesh. You can't express the meaning of a Chinese character by its sound. The more people use Pinyin, the more screwed Chinese characters are.

Simon: To the point that as Yang Yang told me, if you go into a Starbucks in China today, yes, people will be typing using different methods, but-. Yang: That's one of the things that actually saddens me after this interview and because by all means, professor Wang, he is right about it that you do forget how to write Chinese if you are so used to typing in pinyin and that happens to me.

Throughout our interviews that lasted so long, I didn't have the heart to tell him that I couldn't type in Wubi, which just to confirm that young generation has no hope in preserving the Chinese culture anyhow. Simon: Even as young Chinese people, I don't know, as they sit down at their computers or stare down at their phones, are being drawn away from this long rich history of Chinese characters towards this Pinyin phonetic future.

The allure of speed and the search for the fastest way to type continues. Tom: Absolutely. The question still remains, what is the best fastest way to do this? What you have today in China are these typing competitions.

Simon: Where these different methods and different typists face off. These things are a big deal. Simon: This audio is from the finals of a competition back in , took place at China's e-sports hall in Beijing. The broadcast opens with the audience looking down towards a young lady MC who's standing in front of 10 or so desks each with a computer on them. Before the race can begin, she invites the contestants out to stand with her on the front of the stage, this crew of lanky glasses and t-shirt wearing Han Chinese folks.

They introduce themselves one by one and then also-. Simon: They declare which input method they'll be using, because oftentimes the folks who designed the input methods have actually hired and trained these super speedy typists to use their input method. With the introductions done, the MC sends the typists back to their keyboards, some of which are interestingly blank. They have no script on them at all. Tom: In essence, what happens is a text appears on the screen that no one in the competition has seen, the same text for everyone in the competition.

Then the stopwatch starts and the race is on. Just like [typing noise]. It's unbelievable the speed at which they're going. Simon: The room is totally silent other than the clacking of Keys. The cameras cutting between contestants capturing these over the shoulder shots of their screens just filling with text.

When they do linger on one typist's screen long enough and really you'd need to almost go frame by frame to catch this. What you see is a typist inputting a string of nonsense letters, which prompts a little tiny box to pop up with five or so options which they then select from with one final keystroke. Tom: What? That's insane.

I didn't understand. I did not understand, Simon. That's so fast. Oh, my God. My dad who's the fastest typist I know, he could only do like That's wild. Simon: While they're still in this competition, the Winning typist was using Wubi. Simon: Yes, the guy who typed characters a minute was using Professor Wang's Wubi. Tom: Wow, whoa.

They're clobbering us for speed but also able to do that in a way that preserves character writing. Unnamed Speaker: This is not uncommon. Oftentimes in these competitions, it's these older Wubi-like input methods that win.

Simon: Ironically, by all accounts, their top speeds are faster than the top possible speeds of phonetic input. Tom: Wow. Wait, but then he's made this thing that is so blazingly fast and also is able to preserve Chinese way of writing, goes back thousands of years, why is it that these other input methods, these phonetic based methods are winning in terms of usage? Simon: Right, well, the reason there is pretty much the Chinese government.

The Chinese state promotes the idea of phonetic based input systems, really for one major reason. One of the same reasons they prioritize teaching Pinyin in school. Simon: Because although when we think of the Chinese language, we think, oh, there's Mandarin and Cantonese.

In reality, when it comes to speaking, there are dozens of different Chinese languages, Cantonese, Shanghainese, Fujianese. Languages that sound totally different, but on the page look the exact same because they're all using the same characters.

Now, with a structure shape-based input like Wubi where you're describing what the character looks like. You can type and still maintain your spoken language. Jad: It doesn't care if you speak Cantonese, or Fujianese or something or so forth, because you're typing it based on what the character looks like, not how you pronounce it. If you get people having to learn phonetic based input systems, they have no choice really, but to learn to type and speak the standard pronunciation of every character.

Look no further than Wubi. Simon: You can argue that which typing method you use, how you type has a real impact that goes beyond the death of the Chinese character or beyond the government's desire for unification of the language, beyond China itself. Let me give you an admittedly small example of this.

Simon: This is an English study. It was initially done here in the States in the early s. They did a bunch of tests on people trying to find what feelings they associated with words.

What they found was that people like words that have more letters in them typed from the right hand of QWERTY keyboard than not. Jad: No way. Simon: Yes. People like O more than they like E. This has been found in English and Spanish and German and in Dutch, both for right-handed people and left-handed people. Jad: It could not just be that the keyboard was designed so that the letters that we like happen to just be on the right side.

Do you know what I mean? Is it a chicken or the egg type of situation? Simon: It likely is not that. It's likely not that those letters were intentionally placed there. One of the indisputed facts is, if you look at the top row of your keyboard, QWERTY row, it has all of the letters of the word typewriter in it. Simon: They're all there. The story goes that the reason it was laid out this way is because you had these salesmen who would show up and want to demo the product, demo this typewriter, but these guys didn't know how to type.

They put all the letters for typewriter on the top row so that they could very quickly punch out the word typewriter in their demo. Jad: It's totally arbitrary. It was put in the order it was put for reasons that have nothing to do with anything we're talking about.

There is some evidence that the layout of the keyboard created those left-right preferences rather than the other way around. Just a couple of years ago, research has asked, okay, has our feeling towards letters changed over time?

What they did was they got social security records from the s through and they looked at names of babies being born. Let's look at the prevalence of names with more right-handed letters than left before and after, and it spikes after It's crazy. Jad: So suddenly a lot of Paul's and a lot of Leah starts to appear. That is bizarre. Simon: Simon is four right hand one left-hand. Jad is one right hand to left hand. You and I bear out the idea. Jad: It's funny. Was it Wittgenstein? I don't think it was Wittgenstein.

Heidegger, was it a Heidegger thing? Somebody, one of those nihilistic German philosophers had this idea that the hammer isn't just a tool, the hammer actually feeds back.

The hammer changes the hand. It's interesting to me that this arbitrary leftover arguably outdated QWERTY keyboard that we're all stuck with is actually influencing our preferences when it comes to naming our offspring. Who knows what else it's doing? It's probably doing all kinds of weird things to us. Just to get back on track. They had some Chinese grad students actually who wanted to see if it applied back in China, but in part, I think, because there are so many different ways to type, they weren't methodologically able to figure out how to do it, but I will say the idea you bring out or you bring up of the hammer changing the hand, like where Chinese typing is going, I think is the hammer changing the hand on steroids.

Tom: In the United States, I would say the way that people are most familiar with this, is the Google search bar, that when you start to type, it will give you suggestions not based on the absolute mathematical probability of the frequency of a word that you might be doing, but really what's hot in the news and what other people are searching for.

Tom: In Microsoft Word, this is not a search field, this is Microsoft Word and you say, "Okay, in the news today, some star has done something terrible and fallen from grace. The system is smart enough to say, "Okay, this user has never entered this person's name before, but up in the cloud, millions of people are entering this particular person's name. Let's give this local user that suggestion based upon what users elsewhere in the cloud are doing.

Simon: With this cloud-based input, everything you write, every keystroke, every word is being, in some way, influenced by what everyone else is typing. Jad: It is totally unparalleled in the Western world, there is nothing even close to this. In fact, now, arguably, over the last two decades, there has been an inversion in which Chinese in the computational world is arguably the fastest language in the realm of typing.

Simon: We're the ones now looking East seeing these technologies and wondering like, shit, how do we catch up?

In the course of 40 years, China, they've leapfrogged us. Simon: But with this cloud input, there's also a question of do we want to catch up to that? Tom: It's both invigorating, exciting, strange and also eerie and post futuristic because right now it's guessing what the writer already wants to say, but what happens when the speed of suggestion outstrips the speed of thought and the speed of intention?

And what it says is, "Simon, what if you did this? Thank you. Yes, I will do that. Jad: Producer Simon Adler. This story was reported and produced by Simon with reporting assistance by Yang Yang, original music throughout the piece by Simon.

Special thanks again to Yang Yang, without her, this story would not have happened. Next week, we're going to stay international but in a very different part of the world. I'm Jad Abumrad. Thanks for listening. Dylan Keefe is our Director of sound design. Suzie Lechtenberg is our executive producer. Do not learn Cangjie to type simplified characters you can do it, but it is an unnatural workaround and do not learn wubi to type traditional characters. So I would strongly advise learning pinyin which you will naturally do when you begin to study Chinese if you haven't already ; in the meanwhile, you can make learning either wubi or Cangjie a side project to the extent that you feel it works for you.

By the time you start learning Chinese, you should have alread started learning Pinyin. So there is no additional "rules" to learn before you can type. Speaking of speed, Wubi was fast in s because Pinyin IMs had to deal with many characters with the same pronounciation, while Wubi rarely had two characters having the same 4-letter encoding. But thanks to the advancement of processing power and the machine learning technology, nowadays state-of-the-art Pinyin IMs lets the user type the Pinyin of a phrase or even a sentence rather than a character, and predicts the correct characters using advanced Markovian statistic language models In layman's term, the program reads your mind!

Basically you just keep typing. The longer phrase you type, the more accurate it is. And as the computer system supports more and more characters, and the IM starts to have bigger and bigger word lists, Wubi starts to have many words or characters having the same encoding, and it basically undermined its own advantage. Most Pinyin IMs such as Sogou Pinyin let you toggle between simplified and traditional Chinese mode at any time with a hotkey. In conclusion, it's already the 21st century.

Use Pinyin, and let the technology handle the rest. I will give a blanket answer to your question. This might seem like a discussion and sound harsh but I hope it helps. Which is "easier" to learn? Both Wubi or Cangjie are hard to learn if you are new to Chinese. I am basing this on your goal of learning to read and type Mandarin and Cantonese, fyi, character wise, they are the same, only with different pronunciation.

While a pinyin-based input method requires only that you are familiar with the pronunciation, Wubi and Cangjie require that you are already familiar with the characters you want to write. Once you have mastered the characters, in terms of speed and which is faster, both are equally fast once you get used to it. How long do they take to learn? Both takes about 1 month to "learn", but many more months to practice.

This is only if you have mastered reading Chinese first. Now you brought up typing Japanese. This will be a completely different input method so Cangjie and Wubi will not help. At the end of the day, typing is only an additional "soft skill" but you will have to master the "hard skill" of learning the language first. There are very very few people using wubi in Chinese, and zhengma is much less used than wubi.

Because they are hard, and you have to learn them to use them. We use pinyin, because we learnt pinyin even before we started school. And it is fast if you use it for a long time. Also it can type simplified and traditional. Only old people use wubi in our mind. Pinyin would be easier to learn considering you or so I'm guessing are a native English speaker. I speak Mandarin and Cantonese so I can't tell you how much of my advice relates to learning Japanese. While Cantonese and Mandarin are both Chinese dialects, they are very different in terms of tones and sentence structure.

My best advice would be to get your conversational skills down first. Translating your thoughts from English into Chinese words and sentences will take some practice. You can choose either Wubi or Zhengma, but I don't recommend Cangjie, as it is is not compatible to how a character is normally written.

I agree with others that wubi is a mess which is why I stopped studying it It's also proprietary and found almost nowhere. Cangjie, by contrast, is extremely intuitive, ubiquitous, insanely popular There's even a cangjie fan club in Malaysia as well as many spin off input methods such as sucheng and qingsong , and very versatile. I often find myself picturing the characters in my mind as I type them and this visualization carries over into writing.

Usually, I think of a sentence, and my fingers just start flying, spitting out the characters. If it's a character I haven't used for a while, I will visualize it to type it. There is no code memorization necessary. Another advantage of Cangjie is its accuracy. Especially with names, it is very efficient vs. For most common characters there is only one code.

Some key combinations might yield 7 characters, but 6 are ancient obsolete characters. I can pretty much type with my eyes closed and know what's going to be on screen. All this leads to very fast typing. I've seen speeds higher than that on some Chinese blogs, but I can't remember now where I saw them.

I'll add them if I come across them again. Also, to type simplified characters in boshiamy, you need to type the traditional character and use a toggle switch to convert. If you only remember or see the simplified character, you're SOL. It's my understanding that wubi traditional character support is the same with the added problems that many times 1 simplified character can map to many different traditional characters.

With Cangjie, you type whatever you see on paper or in your head. And you can't beat that in my humble opinion. You can also search partial codes using? By the way, if you are interested in Chinese, you could try taking an online Chinese course. The Wubi input method solves these problems. Rather than typing in the pronunciation of characters you want to input, you hit keys assigned to particular strokes or shapes.

There is little to no ambiguity, so any Chinese character can be typed in four keystrokes or less, without scanning and selecting from lists of potential characters.

The system even incorporates quick codes for common combinations of characters, so you can type a two-character word with four key strokes. These two factors combined can up your typing speed significantly. The flipside is that the system takes a lot of time and effort to learn, especially compared to phonetic input methods.