Hugh Pickens writes writes: "When it comes to infrastructure, politicians usually prefer shiny new projects over humdrum repairs. A brand-new highway is exciting: There’s a ribbon-cutting, and there’s less need to clog up existing lanes with orange cones and repair crews. So it’s not surprising that 57 percent of all state highway funding goes toward new construction, often stretching out to the suburbs, even though new roads represent just 1.3 percent of the overall system. Now Brad Plumer writes in the Washington Post that many transportation reformers think this is a wrong-headed approach and that we should focus our dollars on fixing and upgrading existing infrastructure rather than continuing to build sprawling new roads). UCLA economist Matthew Kahn and the University of Minnesota’s David Levinson made a more detailed case for a “fix-it first” strategy. They noted that, at the moment, federal highway spending doesn’t get subjected to strict cost-benefit analysis, and governments often build new roads when they arguably shouldn’t (PDF). And that’s to say nothing of data suggesting that poor road conditions are a “significant factor” in one-third of all fatal crashes, and cause extra wear and tear on cars. "When a highway gets clogged, states find it more palatable to simply build new lanes rather than, say, put in place congestion fees — even though research has found that widening highways does little to alleviate traffic jams," concludes Plumer. "There’s a strong policy case that we could stand to build fewer new highways out to the suburbs, at least for the time being.""
Computers are unreliable, but humans are even more unreliable.
Any system which depends on human reliability is unreliable.