OpenSearch Response Conventions
This describes where the various links are located in the response, and how they are identified.
Within Atom, there is nonetheless a lot of room for variation in how responses are formatted. For integration purposes, we would like to specify conventions that address:
- How the "main" data link is identified
- How different access points to the data are described (e.g., FTP, HTTP, OPeNDAP)
- How browse is identified
- How additional metadata is enclosed: machine tags in the summary? Or an RDFa-like approach? Or something similar to the georss extension to OpenSearch?