RTB design principles (the desiderata): comments and suggestions welcome

[petercorke]

The design principles of RTB are written down in desiderata.md. It is both aspiration and guidance: it states the design the toolbox should converge on, not necessarily the design it has now. Where the code differs, that is drift to correct. Points are marked [done], [partial] or [aspiration] so you can see which is which.

It hasn't been shown to many people yet, and I'd like your view. Comments and suggestions are very welcome, for example:

• Is a principle wrong, unclear, or missing?
• Does something in the code you use fight with a principle, or does a principle get in your way?
• Open topics I'd particularly like input on: how grippers are modeled (Rethink how grippers are modeled, and pause gripper-related changes #884), stateless robots, which simulators and visualizers RTB should support (Backends wanted: CoppeliaSim, Viser, Gazebo, MuJoCo and others (plan first) #886), and what "symbolic-agnostic" should mean when the fast code is numeric only.

Please reply here, or on the Issue for a particular topic. I'm asking for comments rather than pull requests for changes to the design, so we can agree it first. Thank you.