[RFC] XI2 draft protocol specification (v 0.1)

[Simon Thum]

Peter Hutterer wrote:
> XML is merely a container format. It doesn't solve our actual problem - what
> information to send to the client.
[...]
> We're in pretty unchartered territory, with much of the UI needing to change
> anyway to accommodate for all this. Once the UI changes, the information need
> from the server is likely to change too.

A consideration from me:
In the project I'm currently on, there is an externally maintained XML
schema. We (like all other players in the field) extend it using xml
namespaces. This is really another ballpark wrt extensibility, since it
a) reliably avoids clashes b) allows arbitrary extensions and c) can be
processed so extensions are passed through.

So the flow that results is that everyone does his extensions, and some
find their way back into the standard. This wouldn't ever work that way
without namespacing.

So a dreamed up Xorg translation could be that drivers which feel their
abilities are not covered by the spec could simply extend their events.
The sever would process the specified content and pass through
extensions to the clients (optionally restricted to client-specified
extensions, e.g. via xpath or ns). If useage emerges, then extend the spec.

This flow would remedy the points quoted above. However, this is far
beyond the current XI2 scope, or at least that's my understanding.

Cheers,

Simon