Federated search best practices

We should have a standard way of knowing that a harvested dataset comes from an external source (and so probably shouldn't be edited on a local ckan instance) How do we determine this right now?

As more governments work towards offering federated search of datasets we should have a set of best practices for managing harvested datasets and ckan instances containing both harvested and original datasets.

What other use cases should we consider? Should harvested datasets be allowed to have new extra values that are maintained on a local instance? How do we make sure that these values will be kept when the source is harvested again?

No files, tests, or implementation entry points are mentioned. Start by reviewing this issue and the main CKAN Discussions, as the repository is deprecated; the work is only complete once the open questions about harvested datasets, local edits, and federated-search practices have a decided scope.

Written by the indexing model from the issue text.