Sunday, July 09, 2006

MSRP: Managing Severely Retarded Protocol

The messaging and presence protocol space is a good miniature of human passions. It exhibit this exquisite duality you find in those good old duck dodger’s cartoons, between those wanting to master the world and those wanting to save the world.

Behind the smoke screen created by using words such as “open” or “simple”, the sipping working group has been actively busy coming up with yet another messaging protocol. And guess what, this protocol is so superior that “in other words you can build a skookum Instant Messaging application using this technology”. What a world savior!

What really struck me in this article is the blunt announcement that “SIMPLE is primarily a done deal from a standards perspective”. Anyone looking at the dependency list for the 3GPP mobile network will somewhat play down this affirmation. There are a lot of non standard drafts left in the list, with a dependency stated as critical, that are stalling many technical implementations. In particular, several of these dependencies relate to XCAP, the new form of XML query devised by another “would be master of the world” recently recuperated by Cisco. The article goes on saying 

The exceptions are to do with some ongoing work in putting together the last of the framework that’s needed for a user of SIMPLE to be able to tell a service provider what their preferences are on how people can reach them. Which are the last parts of the mechanism that let XCAP (XML Control Access Protocol) work. The XCAP data formats are all well defined and the protocol for the most part is well-defined. The one corner that we are working out right now is the modification of partial documents in XCAP.

There are a lot of conditionals in this explanation, don’t you think? To make it short, XCAP does not work yet. I believe its only “raison d’etre”, beyond the “NIH syndrome”, is the blatant inability of many in the SIP world to see beyond HTTP like request response protocols. To their defence, I would only say that, to those wanting to copy the POTS features, HTTP looks definitively like a very advanced protocol. But I maybe biased. In reality, disguising HTTP into XCAP may simply have been required by the inability of many mobile phone to support anything else than HTTP. If I remember well, in the beginning, J2ME and Symbian (two other world saviors) only had HTTP as their communication protocol. Not even a socket was exposed… One way or another, this is part of the ongoing refusal by most telecom players to consider a customer anything but a lucky “user” of the wonderful services they provide.

Going back to implementations, I have a compasionate thought for all the bright developers trying to make an instant messenger out of SIMPLE, XCAP, and now MSRP. But the gloom is somewhat brighten by the nice discovery that

… if we were in an IM session and I had a picture I wanted to show you. I could do that and we could share that image dynamically without interrupting the Text session plus associating that image with the session when we were done.

Don’t you think this alone is of a nature to create real value and reassure the developers about MSRP? Not long ago the excellent XTen client product team was producing sometimes more than a protocol build every week in order to keep up with the changing specifications. I am afraid this is not likely to change soon with the introduction of this new messaging transport.

Is creating new protocols really the way to promote open and inter-operable communication spaces? I frankly doubt it. But as I said at the beginning, there is no fun without would be “world masters” and would be “world saviors” whacking each other.

Technorati Tags: , , , ,