Support removing content by replacing them with “skipped content”?

Objects in Software Heritage have an intrinsic identifier in the form of a SHWID. This means that it is not possible to remove a single directory entry without changing the identifier for the directory. Which in turn would change the identifier for the source revision… which would change the identifier of one or more releases, and the relevant snapshots then.

That is why we currently limit removal to snapshots and origins.

For the case of Content objects, we could easily allow their individual removal by replacing them by SkippedContent objects. First question, though: is this a good idea?