Aiming an easier interoperability between the engine and other components (e.g GUI app or Blender addon) the following improvements should be addressed:

• Minimize IO operation: file system / network reading and writing, for instance simulation.toml and materials config files.
• Provide an alternative interface to pass simulation configuration, geometry data and sources/receivers information from memory as opposed to reading from config file
• Agree on a standard set of data structures to pass the above mentioned data. At the moment Odeon format is being followed, ideally this should be the exception, not the rule.

These improvements will simplify the mechanism of passing data to the engine and running the simulation in batch mode.

Let's create a few milestones and which features to be implemented.

What do we want/need for:

1. a publication?
2. an alpha release of a commercial product

etc...

This is WIP, a separate repo will be released as soon as a minimally featured version is available.

Please could you detail your "readme" file? I tried to run reflection_test,py on pycharm, but failed. What is the installation steps? or do I need to run ra_cpp file in VS paltform?

I also see some blender info and matlab info, so what is the relation and how to run this project correctly?

Add support for Pybind11 class pickling:
https://pybind11.readthedocs.io/en/stable/advanced/classes.html#pickling-support

This is useful for load/save state of the simulation.

It's been mentioned the Blender addon should implement some sort of layer mechanism. It's not completely clear for me, so here's a thread to discuss this asynchronously.
Why do we need layers? Please share some use cases.
ATM the addon makes use of Blender Materials which allow multiple objects to share acoustic material properties like absorption and scattering. Do we need an additional grouping mechanism?

It's been agreed so far that a list of materials should be kept separately with materials absorption coefficients, something along these lines:

# id	name	alpha	description
0	metal	0.56 0.11 0.35 0.15 0.48 1.0 1.0 0.83	Some description about this material
1	wood	0.26 0.44 0.75 0.08 0.38 0.34 0.0 0.18	And so on
2	polymer	0.0 0.12 0.07 0.0 0.45 0.82 0.3 0.22	And so forth
3	ceramic	0.6 0.4 0.0 0.0 0.0 0.0 0.19 0.14	Ceramic material from Uranus
4	cloth	0.36 0.77 0.44 0.21 0.37 0.6 0.71 1.0	Cloth

Question: should this data base maintain the list of scattering coefficients as well? What are the implications/advantages/disadvantages of keeping it outside/inside the same file?