In our last post about usability, we briefly mentioned the role of machine translation and the tools that can be useful to achieve the desired final product. This time, we will focus on the types of content or settings where machine translation (MT) is more appropriate and effective to represent the human element, which is a factor that still cannot be ignored.
Although some of us are still in shock, the Neural Machine Translation (NMT) is a technology that has been around for a while, and its development has produced very interesting results in the name of usability.
Continue creating new platforms based on this technology. For example, Amazon enters the panorama as the titan emerging who want to dispute the current site of Google, and is currently developing a highly functional engine that is easily integrated in CAT toolsused around the world, as Memsource or Memo Q.
However, before inserting any text in any of these tools, another key factor must be analyzed: the type of material that will work.
It is not the same using MT, not monitored for poetry than for a standard employee manual or brochure of guidelines of the company. The texts are always different in nature and some carry more subtleties than others. For some of them, MT is not yet mature enough to be able to discern the correct terms itself.
Which brings us to another useful tip: never neglect the human element.
Machine translation has come a long way, but has a still longer to go way. Trust the specialists if they can afford it, because they will always know how to make the translated material sound more organic. Why? Simply because they are! (organic, I mean: P).
Do some real end-users to check the usability of the material. It is not always just the consistency of the template and the structural design. It is mainly that the meaning is intact, because inconsistencies without sense are never usable.
Machine translation and then professionally edited material can go a long way. Just be sure to choose the best available options and you will be directed in the right direction with good chances of success.
Therefore, if they try to prove this "usability", remember to think about the users first and how will proceed to actually achieve that. Since a poorly done job is a job that will probably have to make (and pay for) two times