More on Testing

We've been having a lot of discussions recently about testing. I wonder if we can distill some of that into the manifesto. Specifically if/how the model checker and its traces can be better leveraged to reduce manual work, and how to think about how much to test at each level of testing (unit tests, integration tests, compliance JSON tests, etc.).

Review the existing manifesto and the prior testing discussions first. Frame guidance on how model-checker traces could reduce manual work and how testing effort should be divided among unit, integration, and compliance JSON tests; done means these points are clearly distilled into the manifesto.

Written by the indexing model from the issue text.