More thoughtful nested folder metaphors

[keenon]

One (somewhat questionable) design decision I mode in here that we can+should discuss: if you make a folder public, and later make a folder containing that folder public, currently we prioritize the parent folder, and basically hide the fact that the sub-folder was ever public (you can no longer edit the attributes of the dataset). I don't know how I feel about this. What do you think?

While reading this I could only think on how Github manages this problem. In Github you can create a Repository and you associate metadata to it (name, short description, license... etc). Then, inside of the repository you create Subfolders and Files. The repository is always on your profile in the Repositories section. Then, inside of the Repository and in each Subfolder, you can create a readme.md file with additional information about that specific folder. However, each subfolder does not show the metadata of the Repository, but facilitates you ways to navigate up to the containing Repository.

We could do an analogous structure here. In our case, we could allow users to create a Dataset, with some metadata associated specifically to it. This is what users can decide to share or not. Then, inside of a dataset, we could allow users to create Subfolders and Files. Also, we could allow users to add additional information here, like a description of that specific folder, but always keeping in mind that the folder belongs to the Dataset. I think this structure can simplify those decisions since only the parent folder can be shared and its metadata prevails over all of the subfolders (even if it is not shown explicitly, as it is done in Github). This could be less confusing for users, since the possibility of making subfolders of public subfolders public is removed. Only the higher level of the structure (The dataset) can be set as public, which is propagated to every descendant.

• If we want to keep the flexibility of allowing each subfolder to be a separated Dataset, we can do something similar to forks in Github, in a way that the subfolder is copied (or linked, if we don't want to allow modifications) to the main folder as a new Dataset, always linking or referencing its original parent.

Originally posted by @AlbertoCasasOrtiz in https://github.com/keenon/AddBiomechanics/issues/76#issuecomment-1497959792