The models use the El Kebir framework and the Vogt framework. Both frameworks use OpenGL for graphics. Neither is a true framework in the sense that it is not a single library that can be linked with the code for a variety of specific models. The source code directory for each model must contain its own copy of the framework. This is clearly far from ideal. In the future we will either enhance the El Kebir framework (since it is superior to the Vogt framework) to be a library or move to using an existing ABM framework, such as Repast HPC.
The choice between continuing with the El Kebir framework and moving to an ABM framework will depend on the whether or not an ABM framework can be found that is widely used, reliable, actively developed and meets our current and future needs. Those needs include the modeling needs - being able to specify the models, including hybrid ABM-ODE models and multi-scale (ex. cellular scale and molecular scale in the same model) and multi-compartment models (ex. lung model interacting with the lymph model), performance needs - obtaining model results quickly enough for large sets of runs (ex. LHS runs with hundreds or thousands of model runs), and reporting needs - being able to obtain model results in the forms required - both numeric results and 2D and 3D graphics, as single images and movies.
Using an existing ABM framework has the obvious advantages of using packaged software rather than custome software. The basic framework is managed and supported by a dedicated team rather than our lab staff. We can the concentrate on designing, coding and running the actual models. An existing ABM framework will also be a true framework (by definition, otherwise we wouldn't switch to using it). We would have the support of a large community of users. Simply by using the framework they act as a freely available test bed for the framework itself. The user community can also serve as a source of enhancements that we can take advantage of. Any general contributions we make to the framework can be tested and refined by the community as whole, not just ourselves.
The current lung, brain, and breast cancer models use a modeling framework built on the Qt application framework. It was originally designed and implemented by Mohammed El Kebir, a computer science student from the Netherlands who built it as part of doing his master's degree thesis at the Kirchner lab. There is the original 2D framework for running 2D models and a 3D version for running 3D models. The 3D version displays both 2D and 3D graphics simultaneously in separate windows. The 2D graphics display one slice of the 3D grid at a particular vertical coordinate. For example, if a model has grid that is 35x35x35 compartments, the 2D graphics can display any of the slices at the 35 vertical coordinates. There is a GUI control that specifies the slice to display in the 2D graphics window.
The lymph node model Robert Vogt SDL based framework.