|
|
|
|
|
by taeric
4287 days ago
|
|
I think this response actually came later. Apologies for that to the other poster. This is pretty much my point. This "embedded systems use microkernels" is easy to paint in such a way that it supports either argument. Comes down almost to the scottsman falacy. More to the point, if I were to claim that my phone is really running a microkernel, I would be ignoring the majority of the software on my phone. I mean, yes it is true. But it is laughable in that the majority of the programming on the device is not in L4. To a large margin. Consider, you are at this point merely claiming that a state machine in your device is running in a microkernel. And, to the non-microkernel that runs the majority of the phone, it is but a device. So, we have a non-microkernel running a microkernel as a standard device. |
|
> I would be ignoring the majority of the software on my phone
I'd like to argue that the baseband processor is the biggest part of the phone, save for relaying and converting the microphone input. The rest is just a handheld computer. But that's the True Scottsman you made out. In fact, Android might be a good example, just one that doesn't support your point too well, because the Scottsman Fallacy is working both ways. When we are talking true OS with services and drivers et al, android is relevant, ofc. Still, it might be chosen because of any other reason as well. The reference to Android is akin to an argument of authority.
You can run Linux on L4 as a hypervisor, by the way. I have no clue how that works, but maybe we have a microkernel running the Linux-kernel as a service. :)