Better support for DCAT

It is now quite clear that federation across different catalogues will be based around DCAT, CKAN should have a good support for at least a particular implementation. We should probably favour a JSON based implementation, although XML-RDF could be also considered on a second stage if there are resources available.

This should include:
- Expose CKAN's metadata in a DCAT-like form. There is already support for this to a certain extent (eg http://demo.ckan.org/dataset/adur_district_spending.rdf), but it should be exposed on a standard endpoint.
- Consume other catalogue's endpoint (via harvesting extension) to create CKAN datasets.

Unfortunately there still isn't a clear standard way of implementing these endpoints, even the actual schema. Different efforts include:

http://spec.datacatalogs.org/#entity-schema
http://project-open-data.github.io/schema/
http://open-data-standards.github.io/

Some interesting discussions:

https://github.com/project-open-data/project-open-data.github.io/issues/23
https://github.com/project-open-data/project-open-data.github.io/issues/27
https://github.com/dataprotocols/data-catalog-spec/issues/7
## Implementation work

https://github.com/okfn/ckanext-dcat
## Colophon

Originally at https://trello.com/c/WdUKTjG4/3-better-support-for-dcat