
Handling multilingual content in a bilingual translation environment
Discover how to handle multilingual content in a bilingual translation environment with this blog from Patrick Andrew Hartnett.
“What do you know about multilingual source file formats?”
This question was posed to me early last year when my team was tasked with improving support for multilingual source files in Trados Studio. At the time I knew very little about the topic other than that they contain translatable content in multiple languages and the layout of these files doesn’t follow a standard format, which introduces quite a bit of complexity when we want to provide generic support.
My reaction to this question was, why we should bother with supporting multilingual formats in the first place! Trados Studio already knows how to work with a bilingual document, doesn’t it? So, wouldn't it be easier to create copies of the source file, then create a bilingual document for each target language, magically copy in any existing translations, translate the rest and finally wave a magic wand to copy all the translations back into the multilingual native file when you’re done!
Now, unless your name is Gandalf and you take windy walks on the Shire, that kind of wizardry is not so easily mastered. Instead, a project manager typically needs to arm themselves with two commands (Ctrl + C and Ctrl + V) for the tedious task of knitting the translations, first into the bilingual document during preparation, and then back into the multilingual source file after translation is completed; usually with an impending carpal tunnel syndrome shortly after…
So, what is it about multilingual formats that make them so difficult to manage in a translation environment and why don't we manage multilingual documents, as opposed to bilingual as a standard? Well, computer-assisted translation (CAT) tools are designed with features and functionality that enable linguists to deliver projects faster, while ensuring quality and consistency. They are designed to extract the translatable content from a variety of source file formats and structure it into a bilingual format (SDLXLIFF) for each target language.
To better understand the challenges involved we need to look at why we don't introduce the concept of a multilingual document in a translation environment. Let’s consider a typical scenario where we have content to be translated in multiple target languages, and we would like to distribute the translation work amongst different linguists that specialize in those languages. This process lends itself to managing bilingual documents because typically one translator handles one target language. Using a CAT tool you can distribute the individual bilingual document/s separately.
Now, let's also consider that segmentation of a paragraph of text in one language could be different to that of another. This can be represented seamlessly in a bilingual document as you have a 1:1 relationship between the source and target content, enabling you to further segment and/or merge multiple translation units (TU) while maintaining the integrity of the original paragraph. In a multilingual document this would require the skills of Gandalf, as the source content could be segmented differently for each target language, therefore preventing you to align multiple translations across a single translation unit.
You begin to realize that the real question is not 'why don't we manage a multilingual document in Trados Studio', but instead, how can we convince an environment that is predominantly bilingual to manage multilingual formats.
Initially you'd think that this would be easy for Trados Studio, because it already has core features that enable you to create tasks to extract the translatable content from a source file, create a bilingual document from it, then analyze, pre-translate etc... So, how hard could it be to automate those procedures in order to create a bilingual document for each target language from a multilingual source file? Well, as it turns out, it's not difficult at all! Trados Studio provides APIs that enable developers to create their own file types and batch tasks exactly for that purpose.
So, to solve the problem of supporting multilingual source files in Trados Studio, we discovered that the most practical approach was to develop a file type that can piggyback on the core features already available in Trados Studio and then introduce two new automation tasks. The first task, "Import Multilingual Translations", is required to import the target languages and also any additional contextual information provided to support the translator. This task is carried out after the source has been segmented using normal Trados Studio segmentation rules. We need to take into consideration that we can't always map the source to the target as there could be more, or less, sentences in a paragraph to adequately translate the source, so where the paragraph contains multiple sentences the target translations are placed to correspond to the first source sentence only. The second task, "Generate Multilingual Translations" will regenerate the multilingual source file once the project is complete. So, no more copy and pasting or creating scripts to get the translations back into the original multilingual source file.
To start off with, we have developed support for multilingual Excel and multilingual XML file formats, as these are two of the most common multilingual formats used in translation. Multilingual Excel files are the gaming industry’s de facto standard for providing strings of localization and are often used for storing terminology.
The APIs distributed with Trados Studio are an important part of our ecosystem as they enable developers to add unique features and functionality to their translation environment that align with their business needs. The multilingual file types that were developed by the Trados AppStore team are a few examples of what can be done and, as we do with most of our work, it is all open-source on GitHub\RWS Community. We encourage all developers to get involved in contributing to these projects, so that we can learn from each other and continue to provide new and innovative solutions to solve common problems together. There are also many other success stories with apps developed by third-party developers that are already available from the RWS AppStore … why don’t you check them out!
Find out more about our new Multilingual apps here: