"Try holding down shift" to re-arrange the (translated) words.

It's still not working.

http://translate.google.com/#ja|en|%E6%9C%80%E7%B5%82%E7%9A%84%E3%81%AA%E7%94%BB%E5%83%8F%E3%82%84%E3%83%86%E3%82%AD%E3%82%B9%E3%83%88%E3%81%8C%E6%B1%BA%E5%AE%9A%E3%81%97%E3%81%BE%E3%81%97%E3%81%9F%E3%82%89%E3%80%81%E3%83%90%E3%83%BC%E3%82%B8%E3%83%A7%E3%83%B32.0%E3%82%92%E8%A1%8C%E3%81%84%E3%81%BE%E3%81%97%E3%81%A6%E3%81%8A%E3%82%8A%E3%81%BE%E3%81%99%E3%80%82

Does it work for you?

The above is a zero-conditional sentence:

最終的な画像やテキストが決定しましたら、バージョン2.0を行いましております。
"When the final images and text (among other things) have been
decided, then (I) will do version 2.0"

It current results in:
"Shimashitara images or text on the final decision, we 行Imashi the 2.0
version."

How about this:
Hold down a button, and you can highlight (yellow) the input and
output text blocks you are referring to. Usually, and especially with
fragmented languages with post-positional particles and inflexions
attached to the end of the sentence-final verb, it is not going to be
word-for-word or even close to that.

When highlighted you can edit the text to be natural, drag-and-drop
the word order, which ever way is most convenient for input. Bad
translations are voided if there are a greater number of better
translations (equilibrium, self-correction, etc).

Without knowing how the algorithm works, I can only guess that it's
trying to make word-for-word translations and then create a syntax
based on probability of known and registered input. That makes sense
for computers, but it needs to be less word for word if it's going to
sound like natural language.

What do you think?

-- 
You received this message because you are subscribed to the Google Groups 
"General" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/google-translate-general?hl=en.

Reply via email to