@SwiftOnSecurity No such thing as human error in safety critical systems.

@SwiftOnSecurity the quality and safety differences between control system software and application software are a vast chasm. It's depressing, frankly, to know how secure things could be versus how they are.

@SwiftOnSecurity bag of chemicals with electricity inside do the stupid. Responsibility of all bags of chemicals to mitigate stupid.

@SwiftOnSecurity Airplane pilots are (highly) skilled professionals who know that their lives are on the line, and yet the sad history of airplane accidents (especially "controlled flight into terrain" ones) demonstrates that even they can't get it right all of the time.

@SwiftOnSecurity Survivorship bias. We know that the precautions we have taken have always proven sufficient. Others may have had a final horrific moment of realizing their best efforts had not been adequate but are not commenting on their experience.

@SwiftOnSecurity the work on the USCSB on youtube is great to understand more about industrial incidents, i'm not sure yet how it'll influence my work on software systems, but i'm sure it's helpful somehow. https://www.youtube.com/user/USCSB

@SwiftOnSecurity "Set Phasers to Stun" vol 2 but it's just Swift and corn

@SwiftOnSecurity One of the main causes of absolute shit design in virtually everything is its isolation in a stovepiped corporate design bureaucracy that's insulated from user feedback. It's exacerbated by the very large share of products that are made to be sold to the unaccountable procurement department of another corporate bureaucracy, for a captive clientele of users who have no ability to reject shoddy stuff. That's why the proprietary "productivity software" used in most corporations is so godawful compared to the browser- and desktop-based stuff people use in their own homes.
Or to take the example of a physical appliance, just look at the cooling pads used with CPM (continuous passive motion) machines used for range-of-motion excercise on knee patients. The rehab unit where I worked used unglamorous and plain-looking, but entirely serviceable models, for several years. Then management brought in a new model. Instead of a well-insulated ice compartment like that of an ice chest, the new model's ice compartment was a single layer of clear plastic -- which meant our understaffed unit was having to take all the ice machines to the ice machine to refill them every hour or two, instead of the ice lasting an entire shift as previously. Worse, while the old model lasted for years, we had patients who'd gone through as many as three of the new models because of breakdowns -- and got billed for them, at a more expensive price, of course. But because of the clear plastic they LOOKED nifty. And since there were several layers of bureaucracy between the patient and the firm that manufactured them, and they were most likely paid for by a third party, the transaction costs of passing along feedback and holding them accountable were more than the savings were worth.
This is why, as Paul Goodman once said, it typically costs 300-400% more to make or do anything in America than it should. And why there's a 40-hour work week: most of the hours go to waste production, unnecessary overhead, embedded monopoly rents, and bullshit jobs whose main purpose is to compensate for the irrationality.

@SwiftOnSecurity Wow. This is profound.

@enigcryptist @SwiftOnSecurity This book is a slightly dated but great read on the subject of design failures. https://www.goodreads.com/book/show/13517691-to-forgive-design