| Abstract [eng] |
In response to massive software export to foreign countries and increasing computer and Internet usage, localization has become a crucial process of adapting products to a specific location or market. Localization is not just linguistic issues but also content, and cultural norms, and the underpinning technologies. However, existing localization models and tools have limited effectiveness in addressing the main challenges, leading to unresolved critical aspects and a poor user experience. Most software localization is done manually, which is time-consuming, expensive, error-prone, and depends on the localizer’s and tester’s competence. Moreover, support for Lithuanian language localization quality assurance is overlooked by the existing models. The study analyses scientific publications, standards, QA models, and products related to software localization problems and their detection possibilities in Lithuanian language resources. Software and cultural language-specific elements to be localized are described and systemized. A comprehensive comparison of existing localization QA tools and models is carried out. A new localization quality assurance model is created with selected cultural elements and a combination of aspects of compared models. A service implementing the model has been designed to automate the search for localization errors in localized software textual resources. An experimental evaluation of the developed service has been conducted by applying it to more than 100 projects’ localized textual resources and through an expert survey. The results demonstrated the need and advantages of the developed model compared to existing localization models. |
