How to Evaluate the Quality of Free Online Translation (Part 2)

In the previous post we introduced the “back translation” method of evaluating the quality of any free online translation into a foreign language. It is actually double translating back and force between a pair of languages: One professional translates from the language E into the language X, then another one translates back from X into E, and a manager compares the double-translated result with the initial text in E. They should be very close to prove the highest quality of the job of translating into the foreign language completed.

You can think that the same process can be done free and easy automatically – just use an online translator from the language E into the language X and then do the same back – from X into E – with the same tool, and compare the result with the initial text. No, it is not that simple. Why?

Any free online translation tool has a huge memory and for the back translation will just search for that initial text through its memory first of all, and of course will provide your initial unaltered text as a result. And you will have your own initial text back, but not a foreign language translating job performed, see? No evaluation is possible to get this way, as you just get your exact initial text back from the database and that is it!

But the evaluation is possible, if for the back translation you take not the same, but a different free online free translation tool. It is hardly believable that its memory contains your exact initial text as well, as it is definitely placed in the memory of the original translating tool that you used for the first part of your evaluation of the quality of the text in the foreign language, thus this new tool will do the actual job of translating into the foreign language.

Again: To evaluate the quality of any machine translator by the “back translation” method, you have to translate from the language E to X with one translating tool and back from X to E with a different one. Try Google and Bing translating tools, for example, as they belong to different owners so far.

Most of the time the result will be disappointing for you. The only cases when the machine is doing a perfect job, will be for some scholarly grammatically correct and short sentences, like “I go to school every day” or similar ones. The machine is only the machine and knows only rules that had been put in its system, while complicated and especially grammatically irregular sentences require more than that to be translated properly – a human brain it is.

Plus very often idioms are only properly understandable in another language by humans, but the machine translated idioms are laughable at best.

Plus a human translator can easily fix some insignificant mistakes in a source text, but for the machine such things would change the meaning of the text completely: Compare “popular” and “poplar” or “super” and “supper” for example. Your best checking tool would not prompt you to correct these words, and you send the text for translating into the foreign language “as is”. The result?

A summary: You may play with the “back translation” method of evaluation of the quality of any free online translation, and then you better stay away from using the free online translation for your marketing needs.

Copyright © 2016 by Olga Kellen
All Rights Reserved

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s