I appreciate your insights. Indeed, I can't see how a single template classification can work for every possible datasets. Best approaches I have in mind is ei

I appreciate your insights. Indeed, I can't see how a single template classification can work for every possible datasets. Best approaches I have in mind is either determining a set of mandatory and optional metadata shared by every dataset and, as you mentioned, defining an ontology of use cases, sources or characteristics to label datasets.