Solving the embedded OpenGL puzzle - making standards, tools, and APIs work together in highly embedded and safety critical environments

Embedded graphical Human Machine Interfaces (HMIs) are increasingly making use of the OpenGL rendering API as a standard for defining and rendering screen graphics. This trend is supported by the emergence of hardware accelerated graphics subsystems and commercially available driver software. Meanwhile, embedded graphics tool and software vendors have adopted OpenGL in various forms as the rendering API they support. For highly embedded and safety critical environments, however, full OpenGL is not a narrow enough standard. In order to achieve low-cost/low power hardware implementations and reduce driver complexity to achieve safety-critical certification, OpenGL subsets must be embraced. In recent years, the mobile graphics industry has benefited from the efforts of industry consortiums to define capable OpenGL subsets. These subsets, or profiles, exist in various versions intended to facilitate the development of applications for widely differing embedded markets, from cell phone graphics to safety critical high-powered embedded graphics subsystems. It is clear that such well-defined standards can and will have a beneficial impact on the embedded and safety-critical graphics industries, offering unprecedented portability and simplicity for HMI applications. What is not as clear is the level to which graphics tool and software vendors are supporting the new standards. The stakes are high for the end developer, as reliance on API capabilities that are either unsupported or difficult to certify can present serious system integration and certification pitfalls. This paper presents recommendations in such areas as tool selection, standards to levy on vendors and developers, approaches for achieving user interfaces and font rendering using the OpenGL standards, and recommendations to ensure the successful engineering and wide deployment of HMI software.