File Retrieval From Publications

In general, data publications include metadata and one (possibly zero) or more items that have associated byte streams and nominally correspond to files. Different data models may support only one such item per publication, flat lists of items, hierarchies (analogous to directories/folders), or graphs of items (e.g. representing provenance or other relationships). Models may also differ in whether items must have unique names, whether items can exist at given levels in a hierarchy, etc. Without attempting to standardize the model in use, it should still be possible to define a mechanism to identify and allow retrieval of the individual byte streams/files within a data publication.

This task will define such a mechanism. A generic use case would be one in which an app reads the metadata for a data publication and, without complete understanding of the source's data model, is able to display a list of available data items to a user and, once the user makes a selection, is able to retrieve that item (and process it as it would a file). For larger publications, being able to display a basic tree of data items (with some defined correspondence to the source data model) would also be desirable.