Automated Text Generation for the Localisation of an Online Game
|Title||Automated Text Generation for the Localisation of an Online Game|
|Publication Type||Journal Article|
|Year of Publication||2010|
|Authors||Arthur, KA, Brandt, B, Fedane, A, Hannan, D|
|Journal||Localisation Focus - The International Journal of Localisation|
|Keywords||cost reduction, grammar rules, localisation translation, online games, String concatenation, string parsing, translation review|
The problem being examined arises in the gaming industry, where the addition of new and varied content is of great importance to gamers. New content addition is a cost driver and so, finding ways to allow scalable content, while managing costs will be of interest to producers. A solution is presented in the paper to the problem of localising a large volume of sentences of text with similar structures. An automated text generation tool was written in Ruby on Rails, which implemented a set of grammatical rules for each of the project languages, and generated grammatically correct text to appear in the user interface of the online game. The costs of full product internationalisation and rules based translation are compared in a simple model. We conclude that the rules based approach to automatic text generation is less costly, and scalable for this particular case.