Translating software can be very challenging, particularly because strings (text to be translated) are out of context and in the midst of a lot of code. This content could be the name of a menu option, the name of a button, and so on. Only an experienced translator is capable of identifying the content to be translated in the most appropriate way.
Translators always undergo rigorous product training before any software localization begins.
Translators are also provided with a live software environment for reference.
Any contextual comments within the software content are provided to translators so they can clearly understand strings for translation. This step generates a reference list of variable uses with content examples.
After the translation of an application, localization testing is performed on the target language versions of an application. Linguistic testing will show translations in the application, revealing contextual errors and concatenated strings that appear incorrectly. User interface testing will uncover problems with locale-specific formatting and functionality or character encoding.
The next phase involves the use of localized operating systems, browsers, and support software. Locale-specific text, numbers, and dates are entered during testing to ensure the correct functionality of the finished translation.
The final localization step involves the logging of issues in a bug tracking system, such as Jira, where typically about 80% of the bugs are of linguistic nature and can be fixed by localizers before delivery, while 20% will require developers to modify the underlying code.