XLIFF: Multilingual Translation Memory Management among Divergent Language Families

Abstract

For the purpose of localization, only textual output is not sufficing the need of Machine Translation unless until it is in a usable format. In Indian scenario, localization as an industry has not been recognized yet which had led to lack of Language Standards leading to varied translation quality. Localization is the process of adapting a product or service to a particular language, culture, and desired local "look-and-feel" Machine Translation is one of the most important activities under localization but it is not complete unless until it is adapted by end user in a desired manner. In this paper, we are introducing format retention utility using "XLIFF" as an important tool to English to Indian Language Machine Translation. Machine translation (MT) is the technique of translating source text of input language into the target language text. This process uses bilingual data set along with other language assets to frame language and phrase models which are used while translating text in machine translation. Here, Machine Translation System (MTS) that uses the Tree Adjoining Grammar (TAG) is considered. Here, uniqueness and complexity of task has been discussed. In this paper, we are proposing a design and architecture to support the system along with experiments, results and future aspects. It is closely related to the long-term vision of enabling code to support local, regional, language, or culturally related preferences.