DigiLing courses have been localized into Croatian, Slovene and German. With the multitude of content types used in each interactive course, localization proved a challenge both from the technical and the linguistic aspect. Our primary concern was to localize the video presentations in our courses, which reside on Youtube and can be used both within a Moodle course and independently through our Youtube channel.

The localization workflow for each video was as follows:

1. Provide the English transcription. Most videos are transcribed automatically by Youtube’s speech recognition, however these transcriptions had to be corrected for spelling, punctuation, terminology, timing etc.
2. English caption files were downloaded and pre-translated with the Translexy machine translation system,  a Neural MT engine developed specifically for the educational domain.
3. The machine translated files were aligned with their originals and exported as TMX, a translation memory file which can be used by the translator / post-editor in any translation environnment.
4. Pre-translated files were post-edited by translation students and revised by their peers or course instructors.
5. The revised translation were exported and uploaded to Youtube.

The entire localization was performed within an international student project co-ordinated by members of the Centre for Translation Studies of Leeds University in collaboration with the Department of Translation in Ljubljana and FTSK Germersheim. Over 90 students contributed so far, yet not all DigiLing contents have been translated yet. Localization is planned to continue through similar projects in future.