My chemistry lab experience was a lot like software development, but not in the way the author means. We were often given open-ended or under-specified chemist

My chemistry lab experience was a lot like software development, but not in the way the author means. We were often given open-ended or under-specified chemistry problems that required some analysis work in the lab before you could develop a strategy toward whatever the assignment objective was. Also, in some instances each team was given different problems that required different approaches, so to some extent you were on your own if you got stuck. Our lab was unsupervised most of the time, though we were well-drilled on lab safety and procedure. The professors only reviewed our plans in case we were using unusually dangerous chemistries -- not to stop us but to provide specific additional safety information related to that chemistry. It was fun but I spent at least 25 hours per week in the lab for a few years. And the structured planning in the article really wouldn't have helped much because we didn't know what we didn't know, by design as a teaching tool. It wasn't paint-by-numbers. In hindsight, I appreciate that they made us learn chemistry this way. It was a struggle at times but it forced us to grow and learn how to deal with unknowns. This maps much more closely to my early years as a software developer -- being comfortable educating myself on the fly in order to tackle an unavoidably under-specified software problems. I appreciate the benefit of efficiently structured processes, but that won't help you be a successful software developer if you don't know how to systematically grapple with unknown unknowns effectively. Or perhaps more accurately, those skills are most useful if "software development" is limited to domains where the software problems are usually well-specified, like enterprise widget factories.