On 28 Jul 2005 at 21:38, John Howell wrote: > At 5:48 PM -0400 7/28/05, David W. Fenton wrote: > > >I can understand exactly how an error like this could be missed. The > >test machines they use are problem machines dedicated to this kind of > >testing, and they probably image them to a particular state before > >running the installer. But since they aren't someone's actual > >machine, they don't notice that the address book and Firefox add-ins > >and so forth have disappeared, since the machines are not actually > >used for anything other than testing Finale. > > Yeah. And Thalidomide wasn't tested on pregnant women because you > just don't test unproven drugs on pregnant women. Only it turned out > to have horrible and unpredicted side effects on the fetuses. There > are worse things than losing a few files, frustrating as it certainly > is!
Er, Thalidomide was never approved for use by the FDA in the US -- the problem was in European countries which did not, at that time, have the same kind of regulatory framework for drug trials and drug approval as the US had. By now, I'm certain that the European countries have vastly superior drug approval mechanisms, designed with improving the health of people in mind rather than with maintaining the health of the balance sheets of big drug companies. But I digress... The main point is: you have to do the right testing to catch bugs, and you have to evaluate the results appropriately. Again, my assumption is that they *did* test the installer, but just didn't check for the problem that occurred. So, they tested the software, just didn't evaluate all its results. That's *very* easy to do, especially under deadline. -- David W. Fenton http://www.bway.net/~dfenton David Fenton Associates http://www.bway.net/~dfassoc _______________________________________________ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale