ReBench has been in production use for many years. Now it is time to pull it all together, break some backwards compatibility, and apply lessons learned.

The main goal is to simplify and strip out things that are not useful, are not used, and could be done in a more consistent way. As part of it, we need to make the scope and purpose of ReBench clear. It is a tool for defining experiments and executing benchmarks, not for statistical analysis.

None goal: a rewrite. Ideally the changes stay minimal.