In the past I have cited the construction and auto industries as examples of more mature industries. In my view, these industries have better established practices and are more likely to achieve consistent results. Nevertheless, this example shows that they are not immune to problems when those practices are abandoned in favor of untested "innovative" techniques.
In this particular episode, the Department of Transportation decided to go with a concrete overlay procedure that had never been tested with such big traffic volume (100.000 cars a day). The DOT engineers lifted the procedure from a small road in the Raleigh area, made a few assumptions, and sold the idea to the state officials.
At a cost of U$18.6 million dollars, the reconstruction effort will involve the removal of 50.000 tons of concrete and subsequent resurfacing with regular asphalt. For the innovative employees, the consequences are expected to be just as dire, in the words of the Secretary of Transportation:
"My investigation revealed that a number of state DOT employees were at fault from 1999 through the completion of the project in 2004," he said. "Many of these employees have since retired or otherwise left the department. Others who still remain with the department will be subject to suspension, demotion or firing. "Interesting and alien aspects to software engineers; ranging from a failure that cannot be explained away in terms of unrealistic expectations (the asphalt should last 30 years and is crumbling after only 2) to documented links between the failures and the people responsible for it.
In the presence of real liabilities, it seems that innovation should be restricted to the labs until proven. At least if one is building cars or roads.