How We Integrate Technology Into Our Translations

The dream or pitch of science-fiction writers, military generals, and salesmen over the last century has been to push a button and have an instant translation. We now have accessible, instant and seemingly reliable machine translation from heavy-hitters such as Google, Microsoft, and IBM – yet translation remains over a $43 billion industry employing well over 380,000 professionals worldwide. What’s the hold up?

For one thing, machine translation is not yet as reliable as one might think. Human translators are still handling the brunt of the work. Computers can’t yet understand many things that are natural to humans – puns, nuances, brand voice, newly-coined terms, and even grammar. Just look at this MT translation from Chinese into English:

Technology in translation is not a zero-sum game – there are many widely used tools that do not do any translation, but aid human translators in producing more work than in years past. Today, this is also very true in regards to machine translation: this technology, when used properly, can definitely make the translation process go much quicker but cannot do the actual translation better than its human counterparts.

Many less splashy improvements and technological advancements have come along the way to help make the work of the translator, and their project manager, faster, more efficient, and more accurate.

Most projects depend on several people, not just one translator. The following tools are essential not only for helping a translator perform their best, but for keeping all teams on the same page to meet the client’s expectations and deadlines.

Translation Project Management

In the world of project management, translation is somewhat unique. A typical project manager may open and close multiple projects in the same day, maintaining dozens of open projects all at the same time.

Dedicated translation project management software eases the burden by combining as many project components as possible under one roof. Typically this means vendor and client databases, file management, issuing of POs and invoices, graphical workflow, reporting functionality, and practically limitless customization. In other words, you can take care of virtually all aspects of all projects from within the same platform, even doing the translation itself in some cases!

Automatic Resource Leveraging

Translators pride themselves on owning extensive libraries of specialized dictionaries, glossaries, and other reference materials. These are especially helpful when working on materials with jargon, that are very regional, or otherwise unique.

Companies and organizations have their own jargon and voice, too. This might mean using very exact terms for each component of your medical device or avoiding a certain voice when speaking to a client’s customers. It’s not efficient for a translator to keep track of multiple files, glossaries, style guides, emails, and other reference material for all of the many clients they may translate for.

Instead, a project manager, using computer assisted translation tools, called CAT tools or translation environments, can bundle client-relevant information together with the client’s translation history (called a translation memory) and glossary, right in the project itself. Those important terms, previous translations, and words to avoid (or leave in English) will pop up automatically as the translator works. No need to take time away to do research or even switch browser tabs!

Quality Assurance

There is nothing a translation project manager likes more than setting up several layers of quality assurance when seeing a project through. This begins with having a second, separate pair of eyes on the translation itself – an editor – comparing the source and target texts to ensure no omissions or meaning errors have occurred.

A series of mechanical checks, most performed automatically, comes next and are all customizable to an impressive degree. The translator’s CAT tool of choice will have built-in checks to make sure that formatting, including image placement, hasn’t been affected or corrupted during translation. The software will make sure that all portions of the document are translated and that all numbers match. The translator can further define a whole host of other factors that the technology will check before allowing the document to be delivered.

Third-party software like Xbench provides additional functionality. In addition to the typical checks for typos (including repeated words or double blanks), number fidelity, and formatting (including missing or incorrect HTML tags), Xbench can interface with other client translation resources to ensure that a banned term was not used, or that the same, recurring source sentence was not translated five times into five totally different translations.

The Bottom Line

From the typewriter, to mailing floppy disks, to collaborating in real-time on a shared server, much about the translator’s process has changed over the past decades. These changes have not served to push the translator out, but rather to facilitate their working environment and allow them to keep up with today’s increasingly agile and fast-paced world.

We will continue to see machine and neural machine translation make headlines but the need for human translators will not be going anywhere anytime soon. In fact, the number of interpreters and translators in the United States alone is set to grow by 18% in the next ten years-and new technology and translation tools will continue to be developed to assist human translators.