I've mostly been an observer in this thread but I feel I must make an observation.

The definitions I've seen of RTOS are along the lines of "An OS that provides a certain capability within a specified time contraint" or "An OS that operates in an environment whereby the failure to respond within a given time constraint results in critical failure."

Notably absent is any specific time constraint. In fact, given these vague and non-specific definitions, even Windows could be considered "real-time" by these definitions if the time constraints are loose enough.

As such, I'd say that the term "real-time" in this context is pretty much useless and virtually any system could arguably be "real-time." If any system could potentially be considered real-time, the definition of RTOS is too broad or too non-specific to be useful. You might as well just call it an OS or a mini-OS that happens to meet the time constraints of a given application.

The use of the words "real-time" in the term does convey the idea that the OS will respond quickly, but that is not implied by the technical definitions I've seen of RTOS. "Real-time" would seem to imply something that happens in real time, i.e., almost instantly--but that's not the definition used in RTOS. So, again, the "RT" seems redundant and perhaps even confusing since it doesn't provide any qualifiers beyond just an "OS".

It seems unnecessary to create new terms that mean the same thing as RTOS but it also does seems reasonable to make sure everyone knows that there's nothing inherently "now" about an RTOS.

Ironically, whether an OS is an "RTOS" is entirely dependent on the specifications of the application, not the OS itself.

Regards,
Craig Steiner