Linguistic testing

When software is properly globalized, all language-dependent text can be isolated from its context. This speeds the work of the Multilanguage vendor, who then translates strings without their context. However, that places the burden of checking text in context on the Multilingual QA department. Linguistic testing is no luxury: it is the first time that the translated resources are seen in their context. Multilingual QA’s linguistic testing routines cover the following issues:

• Typographical, grammatical, and contextual errors (for instance, is the tabbing order from right to left in right to left languages?)



• Consistency between printed documentation, on-line help, error messages, interface resources, and command-key sequences



• Language-specific settings and resources, such as regional defaults, spell check, speech engine, etc.



• Cultural appropriateness (for instance, order of first name and last name) and politically sensitive issues (for instance, use of male and female adjectives, culturally sensitive images, icons, and terminology)



• Market-specific data, such as contact information, local product support phone numbers, and the format of local addresses