There is a familiar discussion in project management forums about the problems of IT projects, and why information technology projects cannot be more like construction projects.
First of all, I'm not convinced that the Standish Report and other statistics account adequately for projects that are stopped early. If you start a project and stop it the next day, is that a failure or a success? If you stop it after 20% of the work and save 80%, is the project a success or a failure? Was US Airways flight 1549 landed by Sullenberger on the Hudson a failure or a success? When your flight gets cancelled due to weather conditions, do you count it as a failure?
People often tell me that IT projects should be built like bridges. Check how many bridges have the same 'too-thin gusset plates' as in the collapsed I-35W Mississippi River bridge. Did the contractor enjoy the rewards? Henceforth, if future bridges are built more like agile IT projects, there will be more early testing and early prototyping (for example to test the design of the gusset plates) - design of experiments, conjoint analysis, stress loads, ease of assembly, ease of maintenance, risk factors - using physical modelling and computer-aided design - more cooperative and informed decision-making, less nonsense. There will be fewer costly, disastrous and fatal bridge designs. And future failed projects will be corrected, or stopped. Will that be failure or success? Then maybe we will be building cathedrals, not bridges.
If you know the Sir John Egan report, "Rethinking Construction", please convince me that it isn't aligned with agile thinking.