Good questions.   The fundamental issue is that we have more projects than resources, so not everything is progressing at an even pace.  The top priority at the moment is to get the new processor up and running with full support for the existing SRV1_Protocol so that the new processor can benefit from existing software support with MSRS, Myro, RoboRealm, SRV1Console, Webots, Player, etc.  So in large part, the ARM7 and BF537 processors will be running the same firmware source code.  Once this is achieved, we should be able to move firmware enhancements back and forth between the two processors.  
 
There's a limit to how far we can extend the ARM7, as it has only 128kb flash and 64kb ram vs. 4MB flash and 32MB ram on the BF537, but minimally, we should be extending the ARM7 with the back EMF motor control, laser pointer range finding, neural net primitives and further extensions to the built-in C interpreter.  Longer term, there are limits to how far we can extend the ARM7 firmware due to size and speed limits, but that's the current plan.  
 
We definitely plan to offer the possibility of a complete processor/camera/radio retrofit.  Because we haven't yet finished the layout of the new board that interfaces to the processor module + radio + motor controls, we don't yet know what the costs will be, but we aren't making any changes to the SRV-1 robot base or power source, so the physical mounting will be unchanged for the new controller.  
 
We have a fair number of SRV-1's in the field, with the majority in use by universities and research labs.  We have done a lot to make the existing SRV-1's useful, and we will continue that effort with a common codebase, though at some point, there will be a divergence as we take advantage of the performance differences between the processors.