What is with the Chinese translation?!

Hello,

I have very recently bought the 13MT900 (multi-language Partner 900) for the main purpose of learning Chinese, but the support for Chinese is a big disappointment! I also will use it for the French and German as well, though the support for these languages is superb

I have encountered a number of blocking issues including:

* The input system for Chinese words is very bad. To type a word such as 'Sunday', you have to type each character in pinyin (with number) one by one and select each from a very long list. A proper Chinese input system would allow for me to type 'xingqitian' and then this would present 星期天. On my computer, I can even type 'xqt' and the word is presented. Normally these input systems have memory so that your most used words appear at the top of the list. Instead I have to type 'xing1' and select, 'qi1' and select, 'tian1' and select. The current system means that Chinese input takes an incredible amount of time and is very painful.

* The "Text Translation" component for Chinese is a shame! For example translating from Chinese to English:

- Sunday: 星期天. The translation by the 13MT900 is "Week heaven"!!! (http://en.wiktionary.org/wiki/%E6%98%9F%E6%9C%9F%E5%A4%A9). In the Language Teacher, Sunday is presented correctly as 星期天.
- Please speak Chinese: 请说中文. The translation: "Clear say Chinese"!!
- What is today's date?: 今天几号？ The translation: Not possible because ji3 (几) does not exist. It is under ji1! ji1 is correct, but ji3 should also be there. And the question mark also cannot be added to the Chinese text. Anyway, the translation is "Today several mark"!!!
- How do you like this restaurant?: 你们觉得这个饭馆儿怎么样？ The translation: "You feel this luncheonette child how about."!?!
- I agree: 我赞成. The translation: "I agree with".
- Which country are you from: 你是哪国人？ The translation: "You are compatriot."

Seriously, what on earth?

* Again with the Text Translation program. Often it says "TTS not supported". But with Chinese, this text to speak function is quite important as this program does not show the Pinyin equivalent so knowing the pronunciation is not possible.

* The dictionary needs a good audit! Lets try English to Chinese:

Apple: Should be 苹果 but the first entry is 苹 (the second entry is the correct word for apple). (see http://en.wiktionary.org/wiki/%E8%8B%B9 vs. http://en.wiktionary.org/wiki/%E8%8B%B9%E6%9E%9C).
Elephant: Should be the second noun 象, and not the first noun 像 which means “to resemble“ (see http://en.wiktionary.org/wiki/%E8%B1%A1 vs. http://en.wiktionary.org/wiki/%E5%83%8F).

There are many mistakes in the dictionary. Often the correct word is not presented first, but is randomly placed in a long list of translations with the first translation being incorrect.

* In a number of sections, the Chinese text is too small to read. It is not displayed properly on screen. For example in the flash cards which appear as an application starts. Or the Language Teacher "1: Alphabet" section in the matching the squares game.

* Chinese native speakers cannot pronounce the words 'correctly' in the Language Teacher. Additionally the waveform analyser does not seem to take character tone into account (http://en.wikipedia.org/wiki/Tone_%28linguistics%29). I can pronounce the Chinese words with any tone and it accepts it.

* This one is not so bad, but the language teacher "2. Words" section could be better designed for Chinese. The days of the week followed by the months are taught. And then finally the numbers. But in Chinese, the days of the week are 星期一，星期二，星期三，星期四，星期五，星期六, 星期天 (literally week 1, week 2, week 3, week 4, week 5, week 6, week heaven). The months are similar 一月，二月，三月，四月... I.e. month 1, month 2, etc. Therefore there is a lot of repetition of learning of the numbers in this component. Obviously in other languages the days and months have different names, but in Chinese learning these is much simpler.

* Simplified and traditional Chinese characters are mixed together. It is mostly simplified Chinese. But maybe it would be good to separate the 2 and have them as different language choices, even though all of the content, excluding the characters, is the same.

So in the end, I was wondering if it would be possible in the future to obtain an update to the 13MT900 software image which fixes many of these very serious issues?

Cheers,

Edward

Hey Edward,

Wow this information is great and very thorough! I am definitely going to save this and make sure to send it to the product manager and the development team. First let me make sure i understand a few points.

Chinese input system(Pinyin):
-What is the "1" for when you input? Does this designate a space or hyphen?
-So in our you have to do "xing1' and select, 'qi1' and select, 'tian1' and select. To get 星期天. Properly it should just make you type in 'xingqitian'

Chinese input (Wubi):
- I was told this input method is using key strokes and a Chinese speaking native is most families with this. Can you provide any incite or feedback to this?

Text translation:
-Do you notice one way is more accurate then another? For example is Chinese to English more accurate then English to Chinese?
-I have received the feedback about the TTS, i think it may be something they are working on (but i cant be sure till a release).
-Having some sort of memory that finishes the word for you would be nice.

Dictionary:
-Often accurate but sometimes the best word is not first. IS this correct?
-Do you find the other words on the suggestion list to be synonyms?

Memorize card:
-Chinese text is too small. (This is meant to be a loading screen to teach you while you wait, however it cant do that if the text is too small!)

Language teacher:
- You said the Native speakers are not pronouncing the word correctly. Could this be due to a difference in Mandarin and Cantonese dialect?

Additional:
-Simplified Chinese and Traditional are mixed together: Is this present in one feature more then others? For example does this happen in the dictionary, phrasebook, text translation and Language teacher?

Thanks again for all the feedback. I will do my best to bring the issues to the attention of the staff and see what we can do, if we do come out with an update of the current version i will see to it that you are the first customer to test it!

This message was edited 4 times. Last update was at 03/17/2011 13:35:48

Hi,

I'll answer the best I can, point by point.

> Chinese input system(Pinyin):
> -What is the "1" for when you input? Does this designate a space or hyphen?

The numbers are the tones. Chinese has four tones (and a neutral 5th), see http://en.wikipedia.org/wiki/Four_tones or http://upload.wikimedia.org/wikipedia/commons/6/64/Zh-pinyin_tones_with_ma.ogg .

> -So in our you have to do "xing1' and select, 'qi1' and select, 'tian1' and select. To get 星期天. Properly it should just make you type in 'xingqitian'

The standard pinyin input systems do not require the entirety of word to be typed. There is a lot of information here http://en.wikipedia.org/wiki/Pinyin_input_method. Note this is not complete! This may require fundamental infrastructure work to implement something like the IBUS system that I use under Linux (http://en.wikipedia.org/wiki/Intelligent_Input_Bus).

> Chinese input (Wubi):
> - I was told this input method is using key strokes and a Chinese speaking native is most families with this. Can you provide any incite or feedback to this?

I have quite a few Chinese friends, and they don't know wubi. The key problem with this system, from http://en.wikipedia.org/wiki/Wubi_method, is "A major drawback to learning Wubi is its learning curve. Memorization and practice are key factors for proficient usage." I can say that the Chinese computer systems I've seen in operation all use pinyin. If an advanced input bus infrastructure is developed, then this is beneficial for the input of all languages. Pity the systems run on CE, otherwise the IBUS method could be used directly. If this was implemented, the a number of different input methods could be utilised including standard pinyin (http://en.wikipedia.org/wiki/Pinyin_input_method), wubi (http://en.wikipedia.org/wiki/Wubi_method), any of the other Chinese input systems (http://en.wikipedia.org/wiki/Chinese_input_methods_for_computers), the advanced Japanese input methods (http://en.wikipedia.org/wiki/Japanese_input_methods), and the Korean and Indic input methods (http://en.wikipedia.org/wiki/Input_method).

> Text translation:
> -Do you notice one way is more accurate then another? For example is Chinese to English more accurate then English to Chinese?

My Chinese is not good enough for me to validate the English to Chinese route. A nice measure (for development or debugging) is to use Google translate (http://translate.google.com/#en|zh-CN|). Note on Google translate the choice between simplified or traditional Chinese. However a proper language teaching book with both English and Chinese side-by-side would be best.

> -I have received the feedback about the TTS, i think it may be something they are working on (but i cant be sure till a release).

It would be good to have full support. I seems to be something quite doable with the current infrastructure built into the 13MT900 software image.

> -Having some sort of memory that finishes the word for you would be nice.

That together with word prediction would bring the Chinese level up to a good standard (http://en.wikipedia.org/wiki/Pinyin_input_method#Word_prediction).

> Dictionary:
> -Often accurate but sometimes the best word is not first. IS this correct?

I would say half-half. The correct word is not first. An incorrect word is presented for 'apple' and 'elephant' first.

> -Do you find the other words on the suggestion list to be synonyms?

I've been told by native Chinese speakers that these additional words are most definitely incorrect. The dictionary looks like it needs a comprehensive audit.

> Memorize card:
> -Chinese text is too small. (This is meant to be a loading screen to teach you while you wait, however it cant do that if the text is too small!)

It is useful if you are using this to learn English if you are a native Chinese speaker. The reason is because you can recognise what the characters are supposed to be. But I cannot recognise them because I don't know them to start with. Additionally, I don't know how you would pronounce them, so the pinyin spelling with tones could be useful to have there too.

> Language teacher:
> - You said the Native speakers are not pronouncing the word correctly. Could this be due to a difference in Mandarin and Cantonese dialect?

No no, the device is clearly pronouncing with a Northern accent (that of native Mandarin speakers). Those who had problems were from Beijing, so where speaking with the same accent as the device (the Beijing dialect is Mandarin). The problem, well I am guessing here, is that the waveform processor is having problems because Chinese words are short - hence the waveform is shorter than in Indo-European languages - and that the intonation is not part of the current infrastructure. They just flat out cannot pass the "2: Words" section of the Language Teacher - it requires a few tries before a "not bad".

> Additional:
> -Simplified Chinese and Traditional are mixed together: Is this present in one feature more then others? For example does this happen in the dictionary, phrasebook, text translation and Language teacher?

This is ubiquitous. You only have a choice of Chinese. This seems to be predominantly simplified, but traditional characters are also in there. For example in the dictionary (Chi-Eng) if you type 'men2' (gate), you get 门, then 門. The first is the simplified, the second is traditional (which looks like a gate). But if you look at the following, you'll see that the choices are split between simplified and traditional:

http://translate.google.com/#
http://babelfish.yahoo.com/
http://www.yellowbridge.com/chinese/chinese-dictionary.php

> Thanks again for all the feedback. I will do my best to bring the issues to the attention of the staff and see what we can do, if we do come out with an update of the current version i will see to it that you are the first customer to test it!

Thank you, I hope this helps, and that the Chinese component of the Ectaco devices improves enough so that I can use it to successfully learn Chinese.

Regards,

Edward d'Auvergne

Hi Edward,

This is great info. We have over 50 primary languages and close to 200 others so i am always happy and willing to read more information about the intricacies for each of them.

I'm going to show this post to management, development and also use it for technical service so they get a better understanding on different features and functions. Hopefully they are compelled to revise some of the software for future versions and edit some of the things you have pointed out.

If i hear anything interesting back i will certainty let you know and if you run across any questions or comments in the future please feel more then welcome to share them with me! Thanks again and have a great weekend.