Bill-E said:
Apparantly you failed to note the true nature of my statement which is that no complex system designed and built by man will be 100.000000000000.......% reliable.
I got that. Given that I design fault-tolerant and fault-resilant computer-based systems I think I can appreciate that notion.
However, I don't think that dmoore's problem is the typical "computer boot" problem, at least not of the "1 in x Million (or trillion)" type you describe above.
Consider this. If his problem truly is the 1 in x Million (or trillion) type that simply occurs randomly with no root cause then what is the liklihood it would happen between two different restarts? I ask that because he had issues across two restarts. It's very unlikely that two successive restarts encounter this 1 in x Million (trillion) problem, no?
Or maybe the poor performance upon the 2nd restart was the computer "relearning" from being totally f'ed up due to the poor boot prior? And, if that were the case, would it likely have re-learned by the next startup (when things started working)?
Bill-E, don't get me wrong, this might be a computer glitch...I'm just not convinced that it is a boot up issue. For example, does the main carputer even boot upon engine start or is it constantly on as long as the battery is connected and has power (the difference between cold boot and hibernating, so to speak)? Anyone know the answer to that?
I don't know the internals or the design of the Ford car computer, but I do understand that it "learns" things about your driving. Learning requires storage to be able to trend data from various sensors. That storage is either volatile (lost when the power goes off and therefore powered all the time), or non-volatile (retains across power cycles). Now, I know that disconnecting the battery wipes that storage, so it is very possible that the computer is actually running all the time, even if in a limbo/hibernated mode while the car has power. If that is the case, there is no "boot" associated with each restart of the engine. As I said, I don't know.
I'm interested in finding out what dmoores problem is/was. I suspect it might have been a glitch in the computer that worked itself out over restarts and relearning.
Oh, and as for your wish that PCs be as reliable as cars...I doubt it will ever happen. The tolerance for a failure in a PC is simply too high when compared to the same tolerance for a car. People aren't likely to die when their PC crashes. That's not true with a carputer, or a computer running within a medical device, etc. Systems can be made to be more fault tolerant and actually fault resiliant, but for a PC the cost would be prohibitive for most.
Later,
TJR
