Edited by Leslie GordonAnd we trust CAD software, within boundaries. But when what seemed trustworthy fails, we might find ourselves “trussed” by severely constraining circumstances. What can free us from these bonds?
To start, we must constantly update our trust. Each new experience adds to or detracts from it. When someone we trust lets us down, we are hurt and disappointed — and quickly recalibrate our opinion of their ability to keep a commitment.
It’s not too different with software. Whether it’s an O/S, FEA program, or spreadsheet, software is expected to work as advertised. The more we rely on a program’s results for engineering decisions, the more keenly we note its failings — especially because we often feel clueless how it gets from input to output.
For example, there is still significant controversy surrounding the trustworthiness of analysis software. It might be entirely reliable because it does exactly what the developer claims. However, nonexperts might apply improper loads or constraints and then rely on bad results. Still, developers promote the software’s use by nonexperts such as design engineers to help get first approximations. But those well versed in CAE maintain that the education required to properly use analysis software renders nonexpert use questionable — even dangerous. A solution might come from better training for casual users and software that is better at flagging improbable results.
Years ago, my friend Dick Tango said, “Don’t expect what you don’t inspect.” (In science and engineering this is loosely called “an appeal to first principles.”) In other words, when working with people, seek confirmation that their tasks are proceeding as planned. And when using CAE, keep a sense of reasonability by, for instance, performing manual calculations to ensure results are in the ballpark. And in any project, always have “plan B” ready should anything fail. This is a powerful project-management tactic well worth cultivating.
Of course, observations are not exactly first principles, in the formal scientific sense. When we double-check with Joe or Mary to see if their deliverables are on track, we take them at their word. We might request proof, such as an initial draft. This is not an incontrovertible approach, but it does parallel the idea of determining the certainty of a future event by using all available information.
Demanding proof might seem cynical, but it is not necessarily so. You can maintain a positive view of life and people — but still count your change. People, software, and systems are fallible and subject to life’s vagaries. Stuff happens.
So to avoid being trussed by circumstances, measure and qualify your trust. Go as far as possible in inspecting what you expect. Always have a fallback plan, so it’s clear what to do when the expected does not happen. That’s what will make you free.
— Joel Orr
Joel Orr is Chief Visionary at Cyon Research Corp. in Bethesda, Md. Got a question or a comment? Reach Joel at [email protected]