All persistence will be managed by the JBoss CMP engine, therefore, there is no specific work to do concerning JMS/JBoss persistence. Additionally, JMS/JBoss is not simply being refactored--it is being completely rewritten, and therefore, the code in head is irrelevant.
As for the remainder of the project, the overall theme is removing JBossMQ specific invocation layers, persistence, transports, etc. in favor of their generalized system-wide constructs. In short, this means there should be significantly less code required to implement JMS/JBoss 4.0.
I'm finishing up a consulting job, and will begin getting very specific concerning the JMS/JBoss project plan near the end of February. At that point, I invite you to get involved. Watch this forum for details.
I am also really eager to contribute to the re-write, specifically to the new invocation layer, wire protocols and non-java clients. Now that it is March (I believe my calendar despite the snow that is currently falling), has the project plan been drafted?
I plan on releasing a source and binary version of the XIL for jboss3.2 sometime this week or next, but I am anxious to see what the new IL architecture will look like, so I can move XIL to the latest and greatest.
Right now, the invocation stuff is the most in flux. There seems to be several competing sub-projects and I don't think we have it all fleshed out just yet. However, Bill Burke has told me that the I should look at the jmx-remoting project to form the basis of the JMS invocation layer. I think right now, it supports RMI and SOAP, but I'm not completely sure. That is where you'd want to direct your attention. Take a look at that project and see how your stuff might fit in.
Sorry, the module name is jboss-remoting not jmx-remoting.
I am also interested to contribute to the new implementation. I will watch this topic ;)