[QUOTE="Dillon"] The combination? Let's say we have air temperature, humidity, air speed (all at some position in the inlet system), throttle position, vehicle speed, gear selected, engine speed. Temperature, humidity and speed all are used to calculate air mass. Okay, we have 7 inputs. Small number. How many potential combinations of input numbers do we have? Air temp = 0C. Humidity = 50%. Air speed = 100 m/s. Throttle position = 35%. Gear = 4 (out of 5). Engine speed = 3500 rpm. That's one. Air temp = 35C. Humidity = 75%. Air speed = 100 m/s. Throttle position = 55%. Gear = 2. Engine speed = 4800 rpm. Two, shall we go on? Any selection of the above may result in an out of bounds situation whcih the software may handle in any one of a variety of well defined conditions. Or it may fail. I've been involved in the development of a number of "fault tolerant" systems where we had an out that said that the system might fail or behave erratically in the event of "multiple simultaneous failures". In the one of these systems, the number of inputs was only 12 and the number of permitted failures was 3 (more than 3 = MSF). Customer throw 800 different scenarios with more than 3 failures. We found another 1500. Customer wanted to go to 5 permitted failures. We told customer "write us a check for another 5 million". Said system currently lives on 3 failures. Oh, of the 2300 potentilal failures, none have occured. But the system has collapesed three times due to other combinations of events that none of us could have predicted. .... Yes. But was that considered to be one of the testable parameters? Here's the magical part of this. The software guys can only test for situations that are in the realistic to marginally realistic range. The hardware guys had a very, very high level of confidence that the event that he demonstrated would not happen. And his effort required significant work on his part to effect. Have you ever been involved in complex systems development?[/QUOTE] I was glad to see your post. You demonstrated the complexities of failure analysis that the average person has no concept of. They think : "What's the problem - you only have 7 parameters to take into account." They have no concept of permutations, or interactions of complex systems in general, or the unforeseen scenarios that have to be thought of for a meaningful analysis. I was involved in manufacturing FMEA's with one of the big three over 10 years ago. The starting point, by the nominal manufaturer's requirement imposed on us for us to be able to do business with them, for a manufacturing FMEA is to be the design FMEA. The nominal manufacturer often, with a wink, told us they couldn't afford the time it takes to do the design FMEA's,and so we would have to fake that part as the input to our manufacturing FMEA. What a joke the whole thing was. *They* couldn't afford to do what, by their own procedural and contractual requirements, was required on their part, but we, their supplier who they were raping on the per-unit price they were paying us, were not only forced to do our part before we could start manufacturing, but incur extra cost and risk by faking the input to start our process that they were supposed to have provide to us.