BC>I think Craig and Peter covered this well, but to add, think of OS/2
BC>"real time" tasks with a grain of salt.  In any true real-time
BC>system,  you handle real-time events with an interrupt handler.

Not necesarilly, it depends on the architecture of your RTOS. If you're
using QNX for instance it's pretty commonplace to simply have the
interrupt routine set a proxy which then activates a user-task.  This
makes the creation of device-dependent code *so much* easier since such
code runs just like any other app on the system and thus debugging
becomes a case of using existing tools and techniques.  For testing
and debugging you can simulate the interrupts by generating the proxy
from another app.

You *can* write realtime apps using DOS and the basis (even if it's just
the bootstrap to your own RT executive) but using a purpose-designed
RTOS makes a developer's life so much easier that it's a false economy
to try and force-fit your app into a less-suitable (but perhaps cheaper)
operating system.

--- Enigmax v1.04a