Chris Wilper will start to move forward the ideas of projections (based on Tom's PR), and brainstorm how to move forward these ideas
Start at brainstorm level / use cases. Think about naming projections based on use cases (instead of making the REST Client decide every single field to use). One goal mentioned is to make things easier on the client using the REST API. E.g. if they want a set of fields for "browse" or "search" maybe they should be able to select a 'profile' of that name, instead of having to individually request 5-10 metadata fields.
We also should add an implementation summary to DSpace 7 REST: Coding DSpace Objects , as this wiki page provides a guide to other REST API developers on how the underlying code is structured.