I don't think we were ever very far apart. If you look back through our posts I think you will find I have said an RTOS on the 51 is a waste of time on more than one occasion.

With regard to Keil and RTX, there is an interesting web site that covers the Japanese 'standard' embedded RTOS called ITRON. It is interesting because it publishes regular market research data on the use of an RTOS. Their last published data (1999) shows that over 70% of 8 bit designs did not use any RTOS and 27% made their own. Only 3% used a commercial RTOS. Mind you given the number of 8 bit designs, that is still a fair size market.


I am talking about at the design stage before code is cut when you are deciding how you are going to achieve the design goals.

I think we are in danger of misunderstanding each other because we do not have a common understanding of concurrent. To me concurrent simply means (appearing) to do more than one thing at once. I say appearing because a single 8051 con only execute one piece of code at once so you need a mechanism to share the processor time between tasks. There are several ways to do this. A superloop is one, interrupts are another. Perhaps the worst is to save the whole processor state in X-RAM. One method you mentioned in a previous post was using more than one microcontroller i.e. multi-processing.

I strongly suspect that many 8051 project designers are using techniques 'stolen' from the world of the RTOS without realising it. The fact that a lot of people have spent a lot of time developing RTOSs, not to mention a whole bunch of academics who study them, means that we, as engineers could learn a few useful tricks from them.

Ian