Representation type about did-method-web

[gribneau]

This is important, and the JSON vs. JSON-LD issue has been raised before in the context of the file extension in #8.

My personal opinion is that preserving the simplicity of plain JSON is important for ease of adoption, but that JSON-LD will be preferable for a number of use cases. I don't think we muddy the waters too much by supporting both and requiring appropriate filename extensions (.json, .jsonld) and mime types (application/json, application/ld+json).

[gribneau]

This issue is further complicated by the IANA Considerations section of the core spec.

To get into compliance with that language, we would need the following file extensions and mime types:

.didjson
application/did+json

.didjsonld
application/did+ld+json

This would make our very, very simple file naming scheme ( did.json ) oddly cumbersome ( did.didjson, did.didjsonld ).

We might want to consider an update, as supporting multiple file extensions will necessarily involve updating the resolver at the same time anyway.