Shading languages are probably the most significant development in recent years for real-time graphics. The ability to write special-purpose routines to handle effects directly on the GPU has a number of benefits that we are still discovering today. Just think about a few of these rewards: flexibility, expressive power, and integration with the movie industry. But it appears that the current crop of shading languages is just one step in the evolution. We are advancing in the right direction, but are not quite there yet. Most likely, advances in the shader arena will dominate our craft for years to come.
So, where can we turn when looking for a glimpse into the future? Who can offer guidance or ideas for new techniques? I must say the answer is probably hidden somewhere in the movie industry and in the Renderman software. Although it is true that offline shading languages such as Renderman and real-time languages such as HLSL or Cg are a completely different business, their goals are the same. Renderman has been around for some time now and has an expressive power, which is, as of today, superior to current real-time shading languages. By taking the time to learn new techniques, you will hopefully make real-time languages a more sophisticated solution in the years to come.