[standards-jig] Core Tool Protocols

Iain Shigeoka iain.shigeoka at messaginglogic.com
Wed Aug 7 15:49:27 UTC 2002


Sorry Adam.  I think to start talking about JNG needs to address other
issues first.  So I'll put them inline and see what you think.  :)

On 8/6/02 2:20 PM, "Adam Theo" <theo at theoretic.com> wrote:

> I had planned to carefully build them as the start to Jabber 2.0, but
> after some good advice from dizzyd, I have realized that Jabber 2.0
> should not be a hard break from the current protocols. It should instead
> be a gradual, evolutionary progress that results in a whole new
> generation for the Jabber protocol.

I hear this a lot.  My problem with the smooth evolution theory is this:

We (er, well at least "I") want to make two fundamental changes to Jabber to
enable next generation features:

1) Transport independence - free the Jabber protocols from the underlying
network transport.  (raw persistent TCP/IP, HTTP, SMS, etc).

2) Framing - it is pretty widely accepted that the current lack of framing
makes designing and implementing the protocol difficult.  Each packet
(chunk) should be a separate and complete XML documents.  This helps to
handle the network traffic (routing), parsing, namespaces, etc.

I still don't see how you could make a smooth transition between what we
have now and a JNG based around these two major changes.  Perhaps a smooth
transition isn't as important as a compatible transition.  JNG should be
able to run in parallel with Jabber in the same server and messages should
be able to be passed between clients using JNG and Jabber (and servers
should be able to S2S with Jabber and JNG implementations).

Sort of like HTML --> XHTML (transitional) --> XML + CSS

Perhaps we should be thinking more of:

J1.0 --> Jtransition --> JNG

With J1.0 and JNG being necessarily incompatible...  I think it is also
significant that the order of development of the standards was: HTML, XML,
XHTML.  They made an incompatible new protocol then created a transitional
bridge to move from the old to the new...

-iain




More information about the Standards mailing list