On Sep 25, DJ Adams wrote:
> Jabber is not just a network protocol. There are two main cornerstones
> that make Jabber what it is - the protocol, which could arguably seen
> as a 'meta-'protocol (designed with extensibility from the outset),
> and the server architecture, within which hubs (routers), components
> (satellite services) and clients (connector stubs) interoperate and
> can be used to build all sorts of solutions.
>
> Jabber is a development platform much more than it is an IM protocol.
> Clearly, because the first major application that Jabber was used for,
> and arguably the most public, was IM interoperability. But if you look
> at the development efforts and directions in the community (JECL,
> Jabber-RPC, Jabelin, to name but a few), it becomes clear that any
> firm coupling of Jabber and IM, or even the view that Jabber is 'just'
> a protocol, is missing the point, and doing an injustice to Jabber.
I'm _really_ familiar with Jabber, honest. Jabber is much more than
an IM gateway. I'm saying that there's no call for a top-level namespace
unless the Jabber community is going to populate it with a variety of
internals hooks, in much the same way that the Apache community has.
AFAICT, the only Jabber-related Perl modules that exist or have
been proposed deal with "the Jabber protocol", from the network level
(moving bits over the TCP) to the application level (building and
parsing the XML). Ryan Eatmon's excellent and substantial offerings,
for example, live under Net:: and XML::.
If you're planning on doing something more integrated into Jabber's
core, then I endorse a Jabber:: top-level namespace. If you just want to
make the point that Jabber is more than IM, then I don't think the PAUSE
namespace is the best forum for that.
- Kurt
