[QUOTE=Dick Applebaum;1765504]Isn't that a spec rather than a benefit?
Also, doesn't the smaller footprint result, to a large degree, from: a) the ARM architecture; b) unneeded hardware removed (drivers, HDD, OD); c) economies of SOC packaging (ala the A4); d) optimization and packaging of the OS for the target devices (ala iOS).
How about yes & no!
ARM in and of itself has little to do with a smaller foot print as you can run Linux and Unix there. You might make the arguement that SoC packaging results in smaller OSes due to less hardware to support but that is a mistake too. Again there are distributions of Linux that run on SoC hardware.
You do hit upon one thing though that is important that is tailoring the OS to support the target Hardware precisely. Realtime OSes can be tailored to support only the parts of the system needed to accomplish a task.
I am serious here -- A RTOS has advantages/benefits when monitoring/managing a lot of rapid realtime activity in an automobile, router, etc. But, what does that capability bring to an iPhone or a tablet?
For most user level stuff it brings nothing. QNX though brings it's long history with it without the baggage of Linux. QNX should allow them to quickly tailor the OS to the hardware which is a big advantage.
Due to it's realtime nature QNX may allow the main CPUs to take on the tasks often done with dedicated/embedded ARM chips say in the iPhone (Someone once indicated that there are in reality 6 ARM chips in iPhone. I don't know if that is accurate but it is possible. QNX, with the right hardware could allow the tablet to eliminate these support processors and deal with data directly.
Specifically, what can be done better, faster with a RTOS on the kinds of things a smart phone or a tablet can practically perform?
Well if RIM had the scratch to do it, they could do a custom SoC with much of the I/O managed directly by the OS. For example get rid of the baseband processor. From the user perspective they might be able to better respond to user inputs.
The interesting thing here with Apples A4 approach is this do they have the volume to really justify the development of A4 and the follow ons over the next few years. Especially when we talk consummer priced hardware. The current A4 is a trivial design in many respects with the potential for even more stuff on board. The question becomes wil they end up doing more with the on chip CPUs as opposed to embedded chips.
If you look back at the PC industry this is what happened to modems. Initially they where pretty smart devices. Then pretty quickly software based modems came into play. These lowered modem costs at the expense of CPU cycles. In any event QNX should have potential here with modern day analogs.
In the end QNX much like Linux can be transparent to the user. In a well design GUI you should not have to see the OSes dirty underthings. Ideally in a web browser you would not see any difference between that browser running in a QNX based OS, Windows or Linux. In the end what we are hopping for is a somewhat open system that leverages the stable QNX architecture.