If only they had a few people like you and I on staff, we would not be discussing this now.
dunno 'bout you, but i helped deliver my share of bugs to customers in my years. that's life in the big city. what drives me up the wall is developers who deny problems, "just tell the customer not to code that way," instead of just fixing them. as an industry we still undervalue qa folk, just look at their pay. but the rate of and nature of some of the problems we see in the net infrastrucure today do make one wonder about a lack of rigor or a failure to transfer the lessons from some years of software engineering. as our society relies more and more on the net (no accounting for taste), formal methods and the like become more crucial. whether they do remains to be seen. then again, in about '88, when asked why he, hoare, parnas, et al. were no longer beating the software methodology/engineering drums, klaus wirth replied "no one was listening." but, to keep this somewhat operational, would any large registrants be actually willing to use test.whois.internic.net and test.internec.net or whatever they would be called? i.e. is there anyone with significant data willing to be test victims? i doubt i could make a case for it in my company. what's the perceived gain i could use to sell it? randy