Yes, a good source of random errors, true--but only good for acid testing. For verifying error handling coverage you really want repeatable test cases. For that tools such as software network error injectors are really what you want--then you can specifically cover every error and know you've covered it.
Testing hardware is indeed more difficult--you'd need more expensive tools and known faulty equipment can help you there. It all just depends on what you're testing. At the time I wasn't testing--I was trying to get stuff done, and this hardware snafu cost us a lot of time--so I wanted to guarantee it didn't bite us again. Wire wrap? No thanks. So tedious. But on the plus, you could prototype fairly quickly with it. Now a lot of stuff is so high frequency you gotta have it on a pcb w/carefully planned traces, impedance matched, etc. /* PLUG: http://plug.org, #utah on irc.freenode.net Unsubscribe: http://plug.org/mailman/options/plug Don't fear the penguin. */
