We have come a great way in the localization industry, from mostly ignoring and shaming machine-translation (you all remember the funny mistranslations, courtesy of early Google Translate) to discussing whether we can actually fully automate translation. But how automation really influences localization? And is it possible to really automate it all?
The quest for perfection
There are already many perplexing concepts that will inevitably influence the future of translation. In the past, people were convinced that human language is too complex and random for machines to take over… until the rise of neural machine translation engines. This shows that our randomness does have a pattern and that machines can actually improve the understanding of the human language.
Similarly, people were concerned that machine translation engines will have difficulties understanding the sentiment behind sentences. In the meantime, we have started using machines to analyze sentiments on websites and Tweets. Sentiment analysis? Check!
Will there be errors in further improvements of translation automation technology? Of course, just like with humans. That doesn’t mean we will discard the technology and ideas. We will simply improve them, too.
Is it possible to fully automate translation?
The answer is simple: the technology is not there yet. The concepts and the ideas are there, but we don’t know how long we will need to actually make it happen. In the meantime, we will need assistance from humans, as they can also be trained to identify errors and edit machine translations.
So, to which extent can we actually automate translation? Currently, the best combination is that of computer-aided translation, machine-translation, and humans. This is true regardless of the format of the content you are looking to translate. It can be applied to files such as documentation, websites, web apps, and even software.
Translation automation to the max
In the translation process that is fully automated according to today’s standards, the content is first analyzed by the Translation Management System. Then, Translation Memories re-use the text from previous translations. The system divides the work based on the team selected for the project, the Terminology is added to the project, and the content which was not re-used from previous translations is machine-translated.
Here enters the human as the team starts post-editing the translation. For files, you can use screenshots and other references, while for websites and web apps, reviewers can work on the website itself (at Text United, it’s in what we call the Overlay Editor).
All the people working on the project can use the system to communicate, without the need to send e-mails. Once the team is done, the translations can be either downloaded (in case of files) or published immediately (in case of websites). The whole process is as automated as possible, from content preparation to pretranslation and team setup.
How far are we from the automate it all approach?
We are coming closer to that stage every day with every single project! It’s because as the technology evolves, our Translation Management System evolves with it.
What full translation automation means to you?
What’s not any less important is the fact that we look for inspiration both in the latest technological inventions and feedback we get from our clients.
That’s why if you would like to discuss your needs or ideas, feel free to reach out! Together, we may come up with the best solution to fit your needs. As always, there are no strings attached.