[reproducible research] Collect meta data
To reproduce results it is important to collect meta data. This includes simple stuff like date&time of the run, but also runtime parameters e.g. output of the logging parameter tree, runtime data like number of Newton iterations, and compile time specifications like discretisation. For extracting meta data from a simulation run, we have to find a suitable mechanism.
Ideally all such data can be collected by one object and be formatted in a structured way, e.g. XML or JSON format (e.g. https://github.com/nlohmann/json). This output could then also be used to automatically generate meta data for other places, like DataVerse (e.g. using the ExtractIng project). Or just stored alongside simulations results elsewhere.
This is to collect ideas, MRs, issues related to how to design such a system within Dumux (in a way that the data can be easily collected by other tools but also be useful on it's own).
Some meta data could be also interesting for testing to compare against reference meta data.