Something better than Babelfish?
May 5, 2009 3:14 PM Subscribe
Do you know of a reliable translation programme that will take English text and convert it to written Chinese?
Many of my company's customers speak only Chinese languages, but we persist in sending them letters and other documents written entirely in English. I would like to make things a bit easier on them and save our staff from having to do over-the-phone translations on a regular basis.
I want to be able to preserve formatting at least roughly, and basically reproduce a MS Word document in Chinese.
I don't want to have to use a Chinese version of the operating system or get new keyboards.
Upon Googling, I see many options but would like to hear of other's experiences and recommendations.
Thanks in advance!
Many of my company's customers speak only Chinese languages, but we persist in sending them letters and other documents written entirely in English. I would like to make things a bit easier on them and save our staff from having to do over-the-phone translations on a regular basis.
I want to be able to preserve formatting at least roughly, and basically reproduce a MS Word document in Chinese.
I don't want to have to use a Chinese version of the operating system or get new keyboards.
Upon Googling, I see many options but would like to hear of other's experiences and recommendations.
Thanks in advance!
I use NiceTranslator -- I only go from Asian characters (Chinese, Japanese, Korean) to English, not the other way around, but the translations are good enough for me to understand the main points.
posted by rabbitrabbit at 3:31 PM on May 5, 2009 [1 favorite]
posted by rabbitrabbit at 3:31 PM on May 5, 2009 [1 favorite]
If you're thinking about machine translation, I think the state of the art products are Google's translator and Microsoft's.
Both are pretty good, given the enormous difficulty of translation without knowing what any of the words mean. But the output will be far worse than what a bilingual human could do. Both would produce a decent "first-pass" translation that a Chinese-speaking editor could correct.
posted by molybdenum at 3:57 PM on May 5, 2009
Both are pretty good, given the enormous difficulty of translation without knowing what any of the words mean. But the output will be far worse than what a bilingual human could do. Both would produce a decent "first-pass" translation that a Chinese-speaking editor could correct.
posted by molybdenum at 3:57 PM on May 5, 2009
Best answer: I'm a professional translator (Japanese to English) and I feel obliged to step in and tell the OP that he (or she) is asking the wrong question.
Find a random Chinese document on the web (say, the Wikipedia article "bicycle") and compare it to Google Translate's attempt at English. Imagine, for the sake of argument, that the English-to-Chinese machine translation reads about as well to a Chinese speaker as the Chinese-to-English machine translation reads to you. Is that how you want your company to be represented? This is what professional translators are for.
The output from machine translators, especially for difficult language pairs like Chinese-English, is bad enough that it cannot be edited into a usable copy by a monolingual editor and still have any chance at factual accuracy, never mind decent readability. A bilingual editor might be able to do a better job by being able to refer to the source text, but at that point, you're really retranslating.
posted by adamrice at 4:21 PM on May 5, 2009 [2 favorites]
Find a random Chinese document on the web (say, the Wikipedia article "bicycle") and compare it to Google Translate's attempt at English. Imagine, for the sake of argument, that the English-to-Chinese machine translation reads about as well to a Chinese speaker as the Chinese-to-English machine translation reads to you. Is that how you want your company to be represented? This is what professional translators are for.
The output from machine translators, especially for difficult language pairs like Chinese-English, is bad enough that it cannot be edited into a usable copy by a monolingual editor and still have any chance at factual accuracy, never mind decent readability. A bilingual editor might be able to do a better job by being able to refer to the source text, but at that point, you're really retranslating.
posted by adamrice at 4:21 PM on May 5, 2009 [2 favorites]
Response by poster: Hmm...
On reviewing the answers so far, especially adamrice's, I'm wondering if perhaps this project will cause more trouble than it's worth!
I do have many bilingual staff members who do all of our oral interpretation. Depending on the quality of a machine translator's output, I suppose they could edit the written results as well, though anything extensive would probably render the exercise a net loss in productivity.
The professional translator option is just not practical as we have many hundreds of letters going out each week, with different contents.
Anyway, it seems that I might be barking up the wrong tree, but I do thank those who have responded!
posted by Pomo at 5:20 PM on May 5, 2009
On reviewing the answers so far, especially adamrice's, I'm wondering if perhaps this project will cause more trouble than it's worth!
I do have many bilingual staff members who do all of our oral interpretation. Depending on the quality of a machine translator's output, I suppose they could edit the written results as well, though anything extensive would probably render the exercise a net loss in productivity.
The professional translator option is just not practical as we have many hundreds of letters going out each week, with different contents.
Anyway, it seems that I might be barking up the wrong tree, but I do thank those who have responded!
posted by Pomo at 5:20 PM on May 5, 2009
Given what I know about the state of the art in machine translation (and the kinds of research questions that do not have real answers, and are not within years of having real answers), I'd be very wary of doing this for documents that have any kind of importance. (Especially if they may have legal importance, as seems plausible from a previous question you asked.) A similar suggestion to adamrice's is that you try translating one of your documents to chinese and then back to English, and see if the result is comprehensible -- this will magnify error beyond one translation step, of course, but it may give you a sense of what the output will be like.
posted by advil at 5:21 PM on May 5, 2009
posted by advil at 5:21 PM on May 5, 2009
This thread is closed to new comments.
Not sure how good it is but the couple of times I used it for class it didn't need much revision from the teacher.
posted by mrt at 3:21 PM on May 5, 2009