Linking reduced data(derived dataset) and metadata of its parent(raw dataset). #4
Description
I just talked to Max in SIMS team, and the conclusion was that we should avoid duplicating metadata in a derived dataset.
It's because there is a risk that a metadata is updated in the parent dataset but not applied in the child dataset.
Instead, Max said they can provide a simple query(through their backend API) that searches all the derived datasets based on their input datasets' metadata.
I thought it was important to you since it'll be the first step to find the input of the analysis software.
I also wrote that down in our repository.
scipp/essreduce#27
And Max also wanted us to write down a use-case scenario so that they can find the best solution.
I'll scribble things here first so that you can review it before I send it to him.