Games have existed throughout the years in dierent forms, ranging from card games and board games to digital games, each having various subgenres and specic markets. Digital gaming is nowadays one of the biggest markets in the world. It has mostly been used in the context of entertainment. However, games induce eects that can also be useful for other purposes, such as training and learning. A person's concentration level is much higher when he is performing something he enjoys, like a game. Furthermore, video games can be so immersive that the player becomes unaware of its surroundings in real life. Therefore, there is an increasing interest in so-called serious games.
The goal of serious games is to learn while playing. Failure in a virtual world does not have the same consequences as it would in real life. On top of that, even failure is a way to learn. However, the creation of digital games is time consuming and expensive. Therefore, in order to allow for the development of more serious games, tools that support and shorten this development process are needed. This thesis is a contribution to this goal.
The subject of the thesis is a simulator for the semi-automatic replay and verication of virtual scenario models, created with the ATTAC-L language specication. ATTAC-L is a Domain Specic Modeling Language (DSML) developed in the context of the Friendly ATTAC project, aiming to allow non-technical stakeholders to be involved in the modeling of scenarios for serious games. Virtual scenarios are modeled using the natural language based syntax of ATTAC-L and exported in a JSON format. The simulator we created takes this JSON as input and simulates the scenario in a 3D environment. A limited number of 3D environments are possible. Inconsistencies in the scenarios are detected and reported. The user can stepwise go through the scenario and explore different branches.