Um estudo sobre o desenvolvimento de jogos orientado a modelos com codificação manual

Abstract

The electronic gaming industry has recently witnessed an exponential growth in terms of titles and overall complexity of the games. This growth is responsible for many advances in computational technologies. But it also revealed problems related to the development process. It is very common that game projects exceed budget and schedule, not to mention the high number of bugs delivered to the final consumer. Current tools that support the development of games, such as game engines, allow great flexibility and artistic freedom, but require professionals with a high degree of technological knowledge, which means that a high financial investment is often needed. To help solving these problems, the literature suggests the use of Model-Driven Development (MDD). The increased abstraction level made possible through MDD can lead to higher reuse levels, and the automation achieved by code generation helps to reduce development time. But for automation, the developer often has to pay the price of reduced flexibility since the MDD exists a lack of flexibility in designs, but with manual code that can be overcome. Given these challenges, it is proposed an approach for game development that targets models to offer efficiency but without losing the flexibility in projects. This goal can be achieved through the integration of multiple DSLs with handwritten code, with the help of design patterns. To evaluate this approach, it was developed a tool capable of assisting the development of games following the concepts of MDD, which was subsequently subjected to experiments to validate its effectiveness. According to the experimental analysis, it was possible to observe that the proposed approach can provide the desired efficiency and flexibility in game development. It was also noted that due to the automation offered by code generators, inexperienced developers could create their projects with greater ease. However the study indicates that abstraction offered by MDD hinders the learning of the technology used.