1

Topic: Translator Google and

At a forum in a subject about translators wrote that to Google translate connected . I led some tests and drew for myself following outputs: 1) To translator Google really connected ; 2) Translation quality from it became worse. Here some transfers: The priest had a dog, he loved it-> The priest had a dog, he loved her The priest had a dog, he loved her-> the Bottom was a dog, he loved it the priest had a dog, he loved her-> the priest there was a dog, he loved it A priest had a dog, he loved her-> the Priest there was a dog, he loved it Priest had a dog, he loved her-> the priest had a dog, he loved it Once in ice cold winter time I from wood quitted, there was the strong frost-> The cold day in winter time I came out of the forest, there was a strong frost-> Cold day during winter time I came out of the wood, was the strong frost Was enraged a witch malicious, and snow , started up, escaping, in the fine child-> He enraged the wicked witch, and grabbed the snow, let, running in a beautiful child-> It is angered a malicious witch, and seized snow, let, works in the beautiful child Obviously, is "clever"  on the basis of any last trainings defines, in what cases to translate priest as "priest", And in what as "priest". In the translator there is an option "to Offer correction". Probably,  it is permanently trained on these corrections and if millions people offer corrections, transfer will become gradually all it better and better. And all the same I very much doubt that the translator ever learns to translate normally. The reasonable translator in this case translates priest as "priest" and in any way differently. To translate priest as "priest", it is necessary to guess on a context that the word priest has the negative painting; in the given case a phrase for this purpose too small.

2

Re: Translator Google and

Hello, Khimik, you wrote: K> 1) To translator Google really connected ; K> 2) Translation quality from it became worse. If I correctly understood,  connected not so much for translation quality improvement, how many for the sake of simplification of transfer of the language pairs containing at least one  language.

3

Re: Translator Google and

Hello, Khimik, you wrote: K> At a forum in a subject about translators wrote that to Google translate connected . I led some tests and drew for myself following outputs: K> 1) To translator Google really connected ; K> 2) Translation quality from it became worse. K> here some transfers: Result hardly more vital examples. (Original) Why you delayed payment? When from you money arrive? The task which you delivered, has been made. That the customer meant something another, is your problem, instead of mine. (Transfer) Why are you late payment? When you arrive on the money? The task that you have set, was made. The fact that the customer had in mind something else, is your problem, not mine. (Reverse transfer) Why you untimely payment? When you arrive on money? The task which you delivered, it has been made. That fact that the client meant something another, it is your problem, instead of mine.

4

Re: Translator Google and

K> In the translator there is an option "to Offer correction". Probably,  it is permanently trained on these corrections and if millions people offer corrections, transfer will become gradually all it better and better. No, the translator turns to the big dictionary.

5

Re: Translator Google and

When you arrive on the money, Srini? You still owe me pipty bucks.