payload formats and methods, was: Proposal for work on an efficient, browser-friendly, HTTP-based communication protocol for fine-grained information exchange

On 14.08.2010 03:58, Geoffrey M Clemm wrote:
> One of the things I'd like to see with this work is to settle the
> various syntax debates, by just defining mappings to a variety of syntaxes.
>
> One key part of this is already identified below, i.e. define a GET URI
> scheme for any method that involves reading data.

Right. I think if we do this right it might address the biggest 
shortcoming of WebDAV (the non-adressability of PROPFIND results, and 
its implications on every other part of HTTP, such as cacheing or range 
requests).

> In addition, define a POST body syntax for any method that is not in
> HTTP/1.1.

Do you mean for non-read methods? From a consistency point of few that 
sounds logical.

However, I'm not convinced it's needed; the main issue with extension 
methods nowadays seem to be with broken intermediaries, and that can be 
addressed by using https (which is often ok for authoring operations).

If we did this, we probably would need to spec X-Method-Override (shudder).

> Finally, define an AtomPub and JSON variant, in addition to the XML that
> WebDAV would normally use.
>
> This would then allow us to focus on the semantics, and avoid all the
> syntax debates.
> And if a particular syntax has trouble expressing some of the semantics,
> that would highlight why one syntax might be preferable to another.

+1

Best regards, Julian

Received on Sunday, 15 August 2010 10:34:21 UTC