Software products today are made for global consumption, so globalisation is an important step in ensuring engineering efforts are sufficient for worldwide markets. As part of globalisation, internationalisation and localisation are attributes that organisations have long taken care of. However, there is another important piece that is becoming recognised: the linguistic part of the engineering.
What is linguistic testing and QA?
So, what is linguistic testing?
Linguistic testing and quality assurance (QA) are essential processes in localisation projects, especially for websites, software and applications. They ensure that you have high-quality translations that best communicate the brand and the vision of the company to the target audiences.
The use of the word "linguistic" should bring up thoughts of typos, terminology inconsistencies, grammar errors and glaring chunks of untranslated text. Linguistic testing is a language quality assurance step. This language-specific testing must occur directly in the final product because software and website localisation often takes place using tools and technologies "outside" of the actual running application.
Why Linguistic Testing is Important
The linguistic testing and QA processes provide the linguist with a picture of the final product in order to be able to review the translation as a whole and correct any truncation or contextual issues so the final content flows smoothly. It helps finish the final product in a way that guarantees the highest quality.
When a linguistic tester takes a look at the built product, running exactly as local users in a given country will use it, they will see the problems that the out-of-context translation process may have caused.
Translation errors in a UI or web page can impact usability as much as functional errors. If the localised product is full of errors, you will have to update the product once you become aware of them, outlaying more time and money. They are almost always going to be a few linguistic issues in your product after localising it, but fixing them could be easy and quick, versus painful and costly.
An eye-opening conversation
I recently met a linguistic expert talking about this area in great detail and it was such an eye-opening piece of information for me.
He talked about the two main categories called ‘Overt’ and ‘Covert’ that need to be accounted for in the test coverage.
Overt issues are some of the more obvious ones such as grammar errors, typos, translation issues, confusing instructions, missing voices, complex animations etc.
Whereas, when you look at the covert issues, some of the subtle but very important cultural issues surface – these include high and low power distance taking into account geographical powered usability, colour profiles (which we often look only from an accessibility angle) – for example, red in China is a colour of power, but is a colour of danger in the US, symbols and icons, fonts, regional names, border disputes etc. For example, the Japanese market typically calls for more animation. Depending on whether the calendar is Gregorian or Lunar dates will change, holidays will change – he talked about a few examples of how companies like Coca Cola and Nike had to face serious repercussions for their releases that were not tested for linguistic accuracy.
Few more examples I read somewhere:
A westerner travelling to Asia who is offered food with chopsticks or is asked to eat with their hands may have a huge culture shock if not aware of these practices. Considering this small experience that may last just a few days, you can imagine what it would be like to deal with a culture shock quotient in applications to be used for a long period of time by millions of users.
Pepsi had a campaign that said, “Come alive with the Pepsi generation.” This slogan fell apart in the Taiwan market, where “come alive” has the connotation of reanimating deceased ancestors!
These examples highlight the need for sensitised attention to the use of content, colours, and images in global products.
The engineering team, including the stakeholders, needs to understand that there are cultural attributes that need to be engineered in and tested for as part of the localisation process.
What happens if you opt not to test?
If you skip the linguistic testing phase, you run the risk of releasing a localised product that contains not only text errors but also functional or display errors. Linguistic, formatting and technical errors can interfere substantially with an end user’s experience, making quality assurance testing critical after translation and prior to the release of the final product.
For example, take a table with numbers - in English the numbers after the number ‘one’ are plural. But in Arabic and Sanskrit, for instance, numbers after one are not plural and are mostly written in words instead of numerals. They have three cases to be applied to objects followed by numbers: singular case, dual case, and plural case.
If you don’t engage a linguistic tester in the early phases of your product, chances are higher that you will have to subsequently make changes at a design level.
Linguists performing this type of testing look for many things, such as:
- Untranslated text
- Incorrectly translated text
- Grammar and punctuation issues
- Typos
- Inconsistencies in terminology throughout the UI (such as between menu items and dialog box titles) and between UI and help files
- Language-specific characters that are broken
- Inappropriate cultural references, images, symbols, colours and sounds
- Improperly adjusted country-specifics (names, addresses, zip codes, phone numbers, etc.)
- Mistakes in date, time and calendar formats
- Errors in currency conversions and monetary symbols
- Incorrect metric conversions, numeric formats, separators, negatives
Nutshell
In general, linguistic testing is critical after translation and prior to the delivery of the final product. It ensures the accuracy of the final translated content, whether this content is part of a document, application, multimedia, website or software project.
If you have discovered linguistic errors after localisation, what impact did that have on your product release?