> Software is also moving into critical spaces like aerospace, medicine, banking, etc. The thrust of the article is that we're trusting more-and-more critical infrastructure to a discipline that anyone who's worked in knows is untrustworthy.
"Anyone" who's worked in those industries knows SW can be done in a trustworthy way.
At least not less than other engineering disciplines.
"hidden corner cases, unfortunate tradeoffs, rushed deadlines" in uncontrolled proportions are a symptom of lack of discipline, either originating directly at low level (even if maybe mainly because of cultural influences, but I mean, what is not?), or under pressure from the hierarchy. The same conditions can led to critical failures of other kind of engineering realisations. One key point of critical failures resulting from hierarchy pressure is that it does not absolves the engineers doing the work, and some engineering culture actually recognize and teach that. Other cultures mixe everything in the same pot without even an once of ethics nor serious reliability thinking, and you get people maintaining the myth that software just can't be reliable, that the whole industry - without exception - is in an eternal crisis, and that that's even normal because the field is "young". None of that is true; you even have plenty examples around you, and decades of history to study. And of course, we must remain exigent so that the quality does not decline just because of a kind of self prophecy.
"Anyone" who's worked in those industries knows SW can be done in a trustworthy way.
At least not less than other engineering disciplines.
"hidden corner cases, unfortunate tradeoffs, rushed deadlines" in uncontrolled proportions are a symptom of lack of discipline, either originating directly at low level (even if maybe mainly because of cultural influences, but I mean, what is not?), or under pressure from the hierarchy. The same conditions can led to critical failures of other kind of engineering realisations. One key point of critical failures resulting from hierarchy pressure is that it does not absolves the engineers doing the work, and some engineering culture actually recognize and teach that. Other cultures mixe everything in the same pot without even an once of ethics nor serious reliability thinking, and you get people maintaining the myth that software just can't be reliable, that the whole industry - without exception - is in an eternal crisis, and that that's even normal because the field is "young". None of that is true; you even have plenty examples around you, and decades of history to study. And of course, we must remain exigent so that the quality does not decline just because of a kind of self prophecy.