In a paper published 6 years ago a novel domain specific language (DSL) for character animation was presented. High-level action commands are described in the XML format for authoring a virtual human. The problem is, that such kind of framework is to complex for the reality. The predecessor, “Confusius” which was used for animation of a shadow play, was not very well documented and this new software XSAMPL3D has the same bottleneck. The XML schema for describing possible actions doesn't make much sense, because it's not possible to combine the motion primitives to a longer sequence. And the gantt chart in the middle of the paper looks a bit outdated. The idea is perhaps, to arrange the tasks on a timeline, but how can this be realized in XML? If we are looking how often the paper was cited, we will recognize that it doesn't attracted a large audience. A possible reason is that the concept of using a DSL can't be transferred to in-game AI problems, for example to play Atari game with the OpenAIgym.
Am i right?
- Vitzthum, Arnd, et al. "Xsampl3d: An action description language for the animation of virtual characters." JVRB-Journal of Virtual Reality and Broadcasting 9.1 (2012).