Most engineering is based, not on novel applications of science, but on re-application of existing practice. Engineering applies pure science in ways that avoid unpredictability; science itself engages unpredictability, as in meteorology and geology. Part of what engineers do that scientists don’t is know where the areas of unpredictability are, and avoid them. Novel engineering is applied science research in itself, always a bit unpredictable, and the best of us can get caught by it. For instance, this from Arup, a well-respected architectural engineering firm:
The bridge opened to the public on 10 June 2000 when an estimated 80,000 to 100,000 people crossed it. As with all bridge structures, the Millennium Bridge is subject to a degree of movement. However, when large groups of people were crossing, greater than expected sideways movements occurred. The maximum sway of the deck was approximately 70mm. In order to fully investigate and resolve this phenomenon the decision was taken to close the bridge on 12 June. The movement and its effect on the crowds can be seen from the video footage.–http://info.arup.com/millenniumbridge/challenge/index.html
Note also the plea for publication of novel research at the end of the short essay–apparently the phenomenon had been noted before and not written up.
This leads to the following thought: perhaps one of the problems of having a weak distinction between pure and applied economics is that there is no disciplinary body of practice on the applied side: untested theories are put into practice on a grand scale, and when they fail, huge areas of the economy fail with them. This is rare in more mature disciplines; professional discipline makes large-scale failures unlikely. I find it interesting that I do not even have a name for the type of disciplines that are based on pure economics; neither “engineering” nor “design” seems to capture them.
Would there perhaps be some value in modeling an econometrics professional license on engineering or architectural licenses? Milton Friedman would be spinning in his grave if he could see this–he regarded professional licensing as a protection racket, pure and simple. But in fact licensing engineers and architects has worked out fairly well from the viewpoint of public safety, even when very large amounts of money are involved. Could that success perhaps be duplicated in econometrics?