Simulated code is about as beneficial as simulated sex.

Your thoughts on simulators are very well-known here, but I still believe they have a place in the development process.

Since I wrote two (8052 Simulator and Pinnacle 52) I obviously use them in my own development. Almost all my development occurs within Pinnacle. I immediately assemble and run the simulation. I can observe that the logic works and quickly make adjustments when I find bugs.

When it comes to working out timing issues or dealing with external I/O issues then the simulator isn't going to cut it; you need to test that stuff on the real thing. But I've found that I can develop and test the vast majority of my code in the simulator.

In fact, on my last project I was able to write and simulate about 90% of the code before they even told me what 8052 derivative we'd be using and what external parts I would be communicating with. I was able to write it, simulate it, and be very comfortable that it worked. When they defined the hardware I was able to drop in the hardware-specific subroutines and test them on the real thing. They were quite surprised when I had the firmware done one day after they had defined the hardware.

Simulators aren't the right tool for every job or every phase of the project, but to completely ignore their speed and ease of use is to ignore what can be a very useful tool in many cases.

Regards,
Craig Steiner