Extending ATTAC-L to support code generation for serious games (AVAILABLE)

Type of Thesis: 
Master Thesis

Serious games are games developed for a primary purpose other than pure entertainment.

Successful serious games include a compelling narrative context and empirically validated pedagogical intervention methods. In order to create such games, design teams must consist of a multidisciplinary group of technical and pedagogical experts. To facilitate the communication and thus the collaboration between designers with different expertise, a domain specific modeling language ATTAC-L has been developed for the specification of serious games narratives. 

A domain-specific modeling language is a kind of modeling language that uses a dedicated vocabulary and provides abstractions that make the specifications of solutions easier and more accessible for domain experts. In general, domain-specific modeling languages are graphical (visual) languages because graphical or visual specifications are easier for the communication with non-technical people than textual languages; they are also helpful for conveying complex models and designs as they can help people to grasp large amounts of information more quickly than large listings of text.

ATTAC-L is such a visual modeling language. Below is a screenshot of the tool supporting ATTAC-L. ATTAC-L combines flow chart principles and a natural language based syntax to specify the storyline of a serious game. The output generated by the ATTAC-L results in a formal specification that can be processed automatically to generate part of the game code.

 

149

 

ATTAC-L also provides a mechanism, i.e., annotations to specify additional information (e.g., pedagogical relevant information) to parts of the storyline model. Annotations allow the modeler to specify and add relevant, but non-storyline related, aspects on top of the storyline model, such as pedagogical aspects. This prevents that the specification of different aspects (for instance learning and gaming) are entangled. It allows for a clear separation between the narrative content and the other aspects while the modeler can still relate the latter to the story flow.

Currently, the annotations are merely used to specify pedagogical aspects. The purpose of the thesis is to investigate whether this annotation system can also be used to specify information needed for a more complete code generation, such as graphical issues, game rules, or emotions for avatars. After an investigation of the type of information required for a more complete code generation, a proposal for new annotation types should be made. Next the existing tool should be extended with these new annotations, as well as the current JSON output generated. Next, a proof-of-concept code generation should be developed. This can be done by extending the current similar for ATTAC-L. The complete approach should be evaluated with a case-study.

For more information contact Prof. dr. Olga De Troyer.

Background Knowledge: 

·      Good programming skills

·      Knowledge of HTML5/Javascript is an advantage

·      Knowledge of game development (e.g., using Unity) is an advantage

Technical challenges: 
  • Providing abstractions for game implementation issues by means of annotations
  • Translating conceptual annotations into code specification
  • Integrating (weaving) code for different aspect into an integrated whole
Academic Year: 
2016-2017