Why are we so bad at software engineering? | www.bitlog.com

Interesting essay that highlights the challenges of building software, and making sure that it works all the time. I would highlight that as an industry we do know how to make software that does work with very high quality, however it is expensive. Understanding that expense and comparing it to the cost of failure is a practice that doesn't happen often, if ever.

This essay helped me develop a personal takeaway: I need to formalize the cost of a redo when planning a project. I’ve handled this intuitively in the past, but it should be explicit. This formalization makes it easier to determine which tasks cannot be compromised on. This matches my past behavior; I used to work in mobile robotics, which had long implementation cycles and the damage of failure can be high. We spent a lot of time adding observability and making foolproof ways to throttle and terminate out-of-control systems. I’ve also worked on consumer websites for a decade, where the consequences of failure are lower. I’ve been more willing to take on short-term debt and push forward in the face of temporary failure, especially when rollback is cheap and data loss isn’t likely. After all, I’m incentivized to do this. Our industry also has techniques for teasing out these questions. “Premortems” are one example. I should do more of those.

I wonder if there are any formal models for determining this. 🤔