A large part of the cost of fixing a bug, especially late in the dev cycle
after testing is complete, is the cost of regression testing. The cost of
regression testing of a patch for commercial software is much higher than the
cost of a custom web application. Think of an Oracle bug that span
Jon,
I think you're getting out of the scope of the costing exercise. The
research and estimates around "time to fix" are based on the cost
associated with developing the patch, not with deploying it. One could
argue that the cost of fixing bugs - particularly major ones - is much
higher for web a
