The author of this article <https://spectrum.ieee.org/aerospace/aviation/how-the-boeing-737-max-disaster-looks-to-a-software-developer.html> (archive.org copy at <https://web.archive.org/web/20190418200057/https://spectrum.ieee.org/aerospace/aviation/how-the-boeing-737-max-disaster-looks-to-a-software-developer>) is both a pilot and a software developer. He looks into the root causes of the Boeing 737 MAX crashes in Indonesia and Ethiopia. Basic points: * The new aircraft has sufficiently different airframe handling characteristics to be classified as a new model. * Boeing didn’t want it to be treated as a new model, because of the extra regulatory costs, pilot training etc this would entail. * So they implemented software (“MCAS”) that actively managed the handling characteristics to make it behave to pilots like previous 737 models. * They neglected to tell the pilots about the software. * It also turns out that correct operation of this software depends on a hardware component that is known to be unreliable. Commentary from Bruce Schneier <https://www.schneier.com/blog/archives/2019/04/excellent_analy.html> and Glyn Moody <https://www.techdirt.com/articles/20190424/06571442075/shoddy-software-is-eating-world-people-are-dying-as-result.shtml>.