JOSS paper

-- Renaming of the branch for the MR !46 (closed) --

Hey,

I drafted a small paragraph to get things going for the JOSS paper. I think the statement of need part should include some benchmarks, scalability plots, evidence of a large simulation, etc., basically show off Akantu's muscle. Feel free to add/modify what I wrote in the summary, I think we can embellish and add a few more things.

We should also revamp the README and write a description for the PyPI package in the setup.py file, to explain what is included and what is not (i.e. which compilation options are used and if any dependency outside of PyPI is needed---I'm thinking MUMPS, Scotch, etc.). NR: This was done in another MR

🗒 To do list (feel free to edit):

• Describe code architecture
• Parallelization strategy (overlapping domain decomposition, size of ghost zone, etc.)
• Physics that can be simulated with Akantu
• Interfaced solvers
• Dynamics schemes
• I/O
• Scaling for elastic linear statics, explicit dynamics and implicit dynamics
• Scaling for non-linear complex simulation (cohesive elements/contact/damage)

Lucas Frerot