Bootstrap dump creation activity

[azaroth42]

From IIIF/discovery#81, and enabled by IIIF/discovery#83, it would be possible (without any official change to the spec) to create a bootstrap 'dump' file, in the spirit of the Resource Dump from ResourceSync.

When seeing the creation of one of these, a crawler would know that it can retrieve the dump file, process it, and have all of the resources in the state they were at the time of the dump. This would cut down the number of retrieval requests, and optimize for compression over multiple files.

In the situation where the crawler does not understand the profile, format or type of the dump, it would continue and see the individual references in the stream. Thus the dump should come immediately after any entries for resources that are in the dump, and before any further changes. In a 'refresh' scenario, the creation of the dump would be the end of the refresh process.

The challenge is the alignment of filepaths or other identifiers for the packaged bitstreams with the equivalent URIs. This is solved in ResourceSync with a Resource Dump Manifest, with a mandated filename. The members of the TSG on the 2021-01-06 call did not want to expand the scope of the charter to include solving this issue, but consensus was that enabling such experimentation was important to include in 1.0

[azaroth42]

Agreed on 2021-02-03 call to defer until 1.1

[azaroth42]

Propose transfer to iiif/api, keeping defer