After a software or web application has been localized, it is usually tested before being released to the target market. This step is critical to prevent usability, functional, and linguistic problems down the road.
Why are localized products tested? Because the localization process can introduce linguistic and functional problems in the software and often reveals shortcomings in the software’s design, especially if it hasn’t been properly internationalized.
Those problems can be caused either by over-translation of some system variables that are invisible to the target user and are not to be translated, or by modified functionality, which sometimes must be implemented to meet local market expectations or requirements. Letter wizards and spell checkers are typical examples of this.
Linguistic Testing
Linguistic testing confirms that the text in the user interface appears correctly and completely and is not truncated, mistranslated, or misapplied. This testing generally requires a high proficiency in the target language and strong familiarity with the product itself.
Some problems that linguistic testing reveals are:
- Missing translated text
- Text that remains in the source language
- Mistranslations based on the context in the user interface
- Translations that are cut off or wrap to the next line
- Inconsistent appearance or usage
Functional Testing
In functional testing, testers follow prescribed test scripts to run through all aspects of the product to make sure it functions as designed in the localized environment. Often functional testing includes verification that the localized product is compatible with various operating systems and third-party products. Functional testing generally requires good knowledge of the target language and total familiarity with the product.
Some problems that functional testing reveals are:
- Incorrect error messages
- Localization bugs that may cause software crashes
- Incorrect or missing UI elements, graphics, and windows
- Errors generated by installing the localized software on a localized OS
Requirements for Proper Testing
Localization testing requires both source and target language versions of the product installed in the environment that a typical user would use. This means testers must have the correct version of the operating system, language, regional settings, required third-party software, and more. It is critical for the source and target software builds to match, so the expected functionality can be guaranteed.
Testers typically follow strict test scripts that clearly define steps and expected results. Any bugs found are well documented in bug-tracking software. When the bugs are corrected, the software is re-tested to make sure the defect is removed.
In the past, localization testing was largely a manual process, but many tasks have now been replaced by automated software that can discover truncations and other defects. However, linguists are still required to discover errors in meaning, context, and function.