|
Peter Cochrane's Hard Drive 2001 Thinking it through Peter Cochrane despairs at the current knee-jerk decision-making mechanisms for issues that have national and/or global implications I AM at a conference and the speaker has just completed a tour de force on revolutionising the global IT (information technology) infrastructure with new optical technologies. Many people in the audience are reacting negatively and debates are raging on the pros and cons. I could chip in with observations and facts, but instead choose to abstain. The reason? As usual, the presenter and delegates alike seem blind-sided by their own work, their experiences, company policies, established wisdoms and other tribal pressures. Without doubt, they are all wrong in one or more aspect, not to mention the detail. The issue under consideration involves at least the following parameters: technology, civil engineering, installation, operation, maintenance, training and skills, reliability, functionality, serviceability, efficiency, quality of service, cost, lifetime, service, support, customers, availability, integration, management, futures and service migration. This is a modest 20-plus parameters, and there are probably more. This morning, the speaker is considering only three to five issues at most, which is like playing a game of chess by looking at only nine squares. This problem is not uniquely high-tech - it is universal. All significant human activity now entails more choices of parameter than we have evolved to visualise and cope with. Most of us can handle three or four dimensions, but even five can be overwhelming. A common strategy is to choose two or three parameters as the most significant and optimise/make decisions on an approximation. This works well when those chosen are really dominant, but it totally fails if one or more of the omissions are significant. As a result, many analyses are not only questionable, they are erroneous and very misleading. It is bad enough participating in this continuation of errors within the engineering and scientific community, where at least there is some inkling that the problem exists. But living in a society unable to recognise the problem even exists is not easy; many of our governmental, public and company policy decisions are wrong by virtue of the above. For example, smoking and cancer risk seems well established empirically, but the efficacy of health foods, natural farming, organic crops, speed cameras, traffic lights at traffic islands, public transport policies, bus lanes, immunisation, contraception scares, global warming forecasts and more are presented with about as much credibility as astrology. Where is the evidence and the models? At worst, they do not exist, at best a failure to include a comprehensive range of parameters is questionable. For sure, we ought not to ignore real risks, and we should analyse all perceived risks, but we ought not to draw conclusions and make decisions on the basis of inadequate models. Our challenge is to tap the increasing computational, networking and visualisation ability of information technology to educate and inform decisionmakers. We should discourage the use of the knee (as in knee jerk) as a decision-making mechanism for issues that have national and/or global implications. |