Authoring your text with translation in mind can help improve the quality of the translated content, help develop your tone of voice in each language and even reduce costs.
The more value your source text has to your organisation, the more it is valuable to think about optimising it for translation – not only because the positive effects will be multiplied into each target language document, but also because any error will be multiplied if not caught.
There are good reasons to aim for consistency and clarity in your source texts. Not only will the quality and translatability depend on this, but rewriting your content every time also impacts on the scalability of your processes. Translation memories also ensure cost savings for repeated content rewarding consistency.
So what makes a difference? What would allow you to have more authors create more source texts that are consistent in style and content?
Depending on the style and kind of your content, there are several solutions:
- use DITA to organise your content;
- apply grammar checking tools;
- use a standard controlled language approach;
- reuse previous content with an author memory;
- create style guides for your organisation.
DITA and organised content
DITA stands for “Darwin (or Document) Information Typing Architecture”. It is an open XML data model for authoring and publishing. It uses the principles of specialisation and inheritance, which means each topic has a defined primary objective and structure.
It also enables good Content Management Systems to store your content as modules. Each module can be marked-up with information about its use and content.
This way, you only describe once how to fill a pot and then reuse that description and make the necessary changes – like fill a pot with water, or fill it with milk. Either way – you will use similar phrasing.
Grammar and spell-checking tools
There is a range of tools, starting with professional writing aids, that allow to set a particular target tone (“formal”, “technical” or “casual”) and run grammar checks that suggest changes to your text. This may address issues such as passive voice or the sentence structure, and make it a repeatable process. Large corporations, for example in automotive or the aerospace industries, use more sophisticated checking tools.
Some of the most common tools used for English are Grammarly and ProWriting Aid.
Controlled language takes those grammar checking tools a step further. There are several types of controlled language with different approaches, but generically one can see rules that request the author to create content that:
- uses active voice rather than passive;
- has short sentences that make no more than one point;
- uses a clearly defined vocabulary on a “one word – one meaning” basis (for example, to avoid noun-verb mixups like it may happen with “hammer”).
Don’t aim to be Shakespeare, who used more than 28,000 different words in his plays.
Writing the same thing over and over again, with different words, will have a negative impact on consistency. The use of authoring tools (e.g. ArborText, AuthorIT) takes care of this by allowing the author to search text portions and sentences base, for example on a keyword search. The same function can be used as a QA method. So check your authoring tool for boil egg, or fill with water, and you might find a match.
It is not a coincidence that companies like Microsoft have developed detailed style guides for every aspect of their content writing. So, for example, the question of when to use “assure”, “ensure” or “insure” in a documentation is clarified.
Apple, for example, addresses this issue like this: