Re:After Spike in Deaths, New York to Get 250 Miles of Prote...
by bluGill on Monday October 28, 2019 @03:19PM
(#21380509)
They don't. If I make a change that prevents a death you will never know about it. If I make that change in response to a death you will hear about it, thus biasing you into thinking I don't care until there are no deaths. In reality though sometimes you need an actual situation to happen before you can imagine that failure most. Trying to guess how your widget could kill something before their are any deaths is hard. (in some cases you can look at existing widgets, but there is always something new)
[ Parent
| on HN ]
Re:After Spike in Deaths, New York to Get 250 Miles of Prote...
by bdamm on Monday October 28, 2019 @03:33PM
(#21380663)
Quite right. Also, trying to prevent harm through prediction is vastly more expensive than trying a product out in the market, and gets more expensive the more harm you try to prevent. It also causes product development paralysis, as deeper predictive scenario analysis results in longer design times and design lock-down ahead of product release, reducing agility.
While we might want all products and designs to be safe, in reality, we can only economically do so much before the cost of predictably assuring safety becomes exponential. See NASA.