Welcome to my blog. The space where I document my passion about Localization, Project Management and Leadership
When Is Good, Good Enough in the Translation Industry?

When Is Good, Good Enough in the Translation Industry?

Another Sunday that it's raining!

It seems to me that the slogan of sunny Spain needs to be changed! A lot of rain lately!


Thank goodness I like reading, writing on this blog or just turning on my PlayStation and play a Liverpool-Real Madrid to start warming the final of the Champions League 🙂
Lately, my life is very indoors :) but hey, speaking of clouds ... every cloud has a silver lining!
I just finished the book Good Enough Now: How Doing the Best We Can With What We Have is Better Than Nothing
And the message of the book is quite interesting, it explores around the idea of doing something that is good, with the means we have; Sometimes we are obsessed with excellence and sometimes excellence is not necessary, and sometimes Good .... it is good enough ... even in our Translation industry.

I usually read posts, news or articles about the battle between translators and machine translation, a kind of the script from Terminator. Translators against machines, machines against translators, there can be only ONE! and my opinion is that in this fight there is a middle ground, a balance in which machine translation and translators can collaborate and get a result like the title of the book that I just finished ...Good Enough ... and how do we get those results?

Well, I think it is achieved with the combination of the power of machine translation and the interpretation capacity of a translator. When we combine machine translation with post-editing the result can be very powerful. And let me develop this concept a bit more in the following paragraphs.
First, a definition of machine translation and post-editing to make sure we are all aligned with this concept.
In the excellent page that Microsoft has in its "translator's corner", Machine Translation is defined as

Machine translation systems are applications or online services that use machine-learning technologies to translate large amounts of text from and to any of their supported languages. The service translates to “source” text from one language to a different “target” language.
— Microsoft


Post editing, on the other hand, is defined as

Post-editing (or post-editing) is the process whereby humans amend machine-generated translation to achieve an acceptable final product. A person who post-edits is called a post-editor
— Wikipedia

What I like about this definition of Wikipedia is the word "acceptable" which is the same as the theme of the book that I just read "Good Enough".
How can we confront this "Good Enough" mentality in the world of translation via machine translation + post editing?

Well, I think we can achieve it by having a mental state that we can refer to as:

Soft Good Enough MT Post Editing

This approach requires the execution of 3 steps

  1. Align expectations with our client
  2. Adapt cultural references / idioms
  3. Fix mistakes

And it also requires 1 step that we should NOT execute

  • Improve the text "flow"

Let's elaborate on this bullet points ... One of the first steps to take is to make sure that the expectations of the client and the translation team/ vendors/freelancers are clear. Both sides must agree on the quality goal of the completed target translated document.

Is this an internal document for a wiki page or maybe customer internal support?

In those cases where visibility is not very high and what is important is to get a rough idea of the message; the soft good enough MT approach is a valid method. What I mean by a rough idea is basically a post editing review where the target document is grammatically correct, basically, it has a quality good enough that native speakers of the target language might understand it.
In this "soft" approach the goal of the post editing phase is to keep as much MT content output as possible.

Yes, we would correct spelling mistakes as we would edit.

Yes, we would include appropriate idioms.

But any temptation of changes related to the flow of the text is out of scope in "soft" approach. And moreover, it's quite likely that a translator is not paid extra for improving the flow, because this is the soft approach ... because this has been agreed before with the client

Basically, the focus on Soft Good Enough MT Post Editing is

With this post I do not intend to underestimate the work of a translation professional, a great translator can take advantage of MT or NMT, but at the same time, it is necessary to recognize that not every content translated should be premium quality. The level of demand that we can require to the instruction manual of a microwave must be different from the level of accuracy that we need in the web page of our bank.
For the microwave, Soft Good Enough MT Post Editing is good enough ... at least for me.
Which are your thoughts about this? Do you always aim to top-notch quality? Or sometimes readability is good enough? Which are your thoughts generally speaking for the good enough mindset? Thumbs UP or Thumbs DOWN?

I would love to read your comments below …

Have a wonderful week!

Continuos Localization is not Agile .... follow up of my master class in Barcelona Videogames incubator

Continuos Localization is not Agile .... follow up of my master class in Barcelona Videogames incubator

Localization as part of Global Business Strategy –Decoding the air C-level Executives breath

Localization as part of Global Business Strategy –Decoding the air C-level Executives breath