Can of worms: the non-native translations
Why might a non-native translator be better (sometimes) than a native one?
First, a good, specialised, non-native translator might have a better command of the specialised language than a generalist native. Second, some language combinations, such as most combinations with Romanian, simply do not have enough specialised translators. Third, the best non-native translators are those who are very good at translating into their native languages, too.
That leaves us with the quality assurance issues. How do we make sure our non-native translator is as good as they claim?
Tests are one way to do it. While an experienced translator should be great in their native language and should not require any testing, I would strongly recommend specialised, native-reviewed paid tests. You will most certainly get a great return on your minimal investment.
Assuming the test went well and that you’ve already selected your translator, it is time to make sure the end product is up to par. My preferred method (because I do translate in two of my source languages), is to have the translated texts reviewed by a specialised native, even if monolingual. They might spot errors, misinterpretations and request clarifications, thus ensuring a spotless end product.
Also, if you are the translator in this equation, please be aware of your limitations. I, for an instance, am comfortable with fields I often translate into Romanian in (technical, medical or legal translations), or texts read a lot about, such as food and beverages or tourism content. I would not dare translating in my source languages in others, because my foreign language vocabulary is simply not there.
Let me say, again, as the quality-obsessed translator I am, all parties should act responsibly. If you’re the client, don’t err on the side of cheap. If you’re the agency, make sure you can provide native review. If you’re the translator, know your limitations.