OK let's not chew more on how I percept your first answer to the OP.



I don't think I would not know how the watchdog I would use works. I never used one - for reasons I believe were/are valid for the designs I was involved in. However, I recommended to use some watchdog here and there, where I thought it appropriate, overviewing the result with the designer, and giving hints on its usage. You might have read my view on watchdogs here and in that (still unfinished) document I advertise so often.

I also don't think you don't know how the watchdog you are using works. You might have your doubts whether an IC which internals are not quite open to the public might be resilient enough to EMI/conducted interference, which would throw your - as you say, in this respect, perfect - design out of normal operation; but then it might be questionable whether your design needed a watchdog at all - for example, if your designs are operating in environment where you CAN easily demand "correct" operational environment, legally or through a contract (e.g. in laboratories, hospitals).



You, as a proud engineer, would undoubtedly do everything possible to achieve this.

But, remember, engineering is a balancing act, and you should never over-engineer your product. If your product is supposed to meed a certain level of immunity, either because of some compliance testing (CE, FCC, automotive, zillions of other) or because you KNOW it is going to be operate in a certain level of EM noise; if you make it an order of magnitude more resilient at a cost an order of magnitude higher than a less resilient design would cost, you might very easily got out of that particular business.

Again YMMV; and, again, THAT was my point.

JW