On 26.03.2012 22:28, Amgine wrote:
Are we talking about mime types for articles?
Yes, pretty much. Though technically, the mime type would only describe the serialization format (e.g. "application/json"), not the data model (e.g. "wikidata entity record") - both bits of information are needed. But essentially, yes: pages will have types, and types have handlers for displaying, editing, etc.
Otoh this means that HTTP content negotiation is not going to work, or rather, it will only work to pick the the preferred serialization format for a given content. If the client doesn't know the content model, the data is going to be gibberish to it, no matter how it is serialized.
Which is why I propose that per default, any code or client unaware of the possibility of content that is not wikitext, will receive the equivalent of an empty page when requesting content of a page that isn't wikitext. No harm done that way.
-- daniel