Re: Use cases

On Fri, Dec 31, 2010 at 12:46 PM, Julian Reschke <julian.reschke@gmx.de> wrote:
> What's essential is that payloads are self-describing, and media types that
> support compound information (for instance, through XML namespaces), are
> completely OK.

We're not talking about random media types, but text/html
specifically. Hypermedia formats have a special place within REST
architecture:

http://roy.gbiv.com/untangled/2008/rest-apis-must-be-hypertext-driven

Sprinkingly some namespaces around does not magically produce a
document that software can turn into a human-friendly hypermedia
interface.

Software must be programmed to present the represented semantics to
the end-user. With text/html, we have a lingua franca consisting of a
set of generic document semantics (HTML/MathML/SVG) that we expect to
be mostly understood by today's software agents and universally
understood by future agents.

Arbitrary elements from other namespaces are not part of that lingua
franca. Their introduction would babelize text/html, making it not a
generic hypermedia format but just a catch-all data serialization like
application/xml or application/json. It would no longer be the case
that you could request text/html and expect to get back a
representation you can understand and turn into a hypermedia
interface, which would make text/html unsuitable as the default
representation of a RESTful service.

If we want to talk about expanding the lingua franca, let's do that by
all means, but the right approach to doing that is to add to HTML
through mutual agreement on common understanding of the new semantics,
not to support Joe Blogs developers adding random XML gobbledygook to
the lingua franca in the deluded hope that one day all user agents
will be specially programmed to do something useful with it.

--
Benjamin Hawkes-Lewis

Received on Friday, 31 December 2010 13:18:20 UTC