??? 06/01/08 13:25 Read: times |
#155345 - the fallacy of 'testing' Responding to: ???'s previous message |
I have already mentioned the case of a box, otherwise operating flawlessly many hours in many installations, failing when placed in a room next to powerful transmitters used in car racing.
I do not have enough fingers and toes to count the cases I have seen/heard of where a 'flawless design' failed in the environment it was supposed to work in. If the OP has a power problem, the the cure is NOT a watchdog, but a couple of NiCads (or supercaps). Two cases can be seen a) you know what the problem is (the OP indicates he does) so the solution is simple handle it. b) the problem does not have a known cause so the solution is find out why and handle it. Any design that depends on the watchdog is, in my book, faulty. That the watchdog serves a purpose is illustarted by Jans example when, as it turned out, a new powerful radar has been installed on the nearby airport but once that has been found out a redesign making the puppy sleep peacefully is required. Erik |