DavidS wrote: ↑Tue Jul 23, 2019 6:30 pmYou omit to change the meaning all over again.Contrary to statements above, it is not easier to meet hard real-time constraints on a cooperative system. Although there may be other reasons to select a cooperative scheduler. I hope no one is misled by this.
I said that there is a reason that many Hard Real time systems are still cooperative multitasking. It is easier to predict the cycle count if you know all of the running code and the task schedular/switcher implementation, on a cooperative multitasking system, this is true.
Ok I should have been more clear that this IS NOT an advantage of RISC OS, do NOT ATTEMPT to use RISC OS for Hard Realtime (unless your time constraints allow for the tolorance needed).You lost what I said completely, and are attempting to make it sound like I said something different to what I said.Contrary to statements above, the issue of unresponsive GUIs is not down to the use of a preemptive multitasking system. The same situations, missing/ignoring GUI events, can be achieved with a cooperative scheduler.
I said that if software were to be unresponsive do to ignoring events on a cooperative multitasking system people would be less likely to use it. Thus it encourages good coding practices to help maintain user responciveness.
The same responsiveness is possible regardless of multitasking model. You are just likely to get away with worse responsiveness on a Preemptive Multitasking GUI system than on a Cooperative Multitasking one while still having users accept and continue to use the software.
I was extremely clear in my meaning. So as you are without question attempting to change the meaning of what is said to make it sound bad, I will not be surprised if this thread gets locked. Thanks a lot.
Yes locking your posts would be advantageous