I feel that the embedded industry is always looking for a silver bullet that they can simply pull from the desktop market and will solve all the issues of the embedded world. A while ago it was Java, and everyone thought java would solve all issues. However everyone soon learned that it didn't and introduced a whole bunch of its own.
Is it possible Adobe's flash is being look at to be the new silver bullet?
On Linkedin in one of the settop box groups they've been discussing the merits of using flash for the UI and it has brought up many interesting issues and experiences. Here is just a couple excerpts :
Based on my team's experience with flash on a professional product for its user interface I am against the use of flash for embedded devices. The flash engine was very unfriendly for the kind of low impact and real-time system response needed, and we were constantly making trade offs between the cycles given to the flash engine to redraw the screen (and keep animations smooth) and other processing requirements in the system.
if people keep pouring money into flash the environments are likely to get there. But if I was king and had my project to do over again there's no way I'd start out using Flash. Our machine had two PowerPC's in it and the engineers talked about enabling a third just to run the Flash UI. But the problem (IMHO) was the environmental disconnect in the architecture between the Flash world and the rest of the system, not CPU power.
I think Flash is more a marketing term than an architecture decision. The output of the tools doesn't readily translate into embedded environments.