Hacker News Reader

Comment by stefco_ | Hacker News Reader

> a zillion NASA engineers didn't really think it through when lives were on the line

The in-depth investigation following the Challenger disaster [0] actually reached a slightly different conclusion: engineers tended to estimate ~1/100 launches would end in failure, while management estimates were about 3-6 orders of magnitude lower. The engineers did think it through; it was management that failed utterly and completely in their assigned task.

Anyway, I think your point is correct, but in this case it was management who learned that they could keep ignoring engineering expertise and get away with it until it literally blew up in their faces. Anyone who's had their boss urge them to rush bad code out the door (only to get chewed out when it fails as catastrophically as they knew it would) is familiar with this.

[0] https://en.wikipedia.org/wiki/Rogers_Commission_Report

View on HN

Thanks for the correction.

But I'm not sure I agree with your diagnosis. I suppose it depends on how one thinks of engineers and software developers. Are we professionals, like doctors, with a bedrock responsibility to society and our users? Or are we just very well-paid gofers, who do whatever the bosses say?

In my view, responsibility isn't a zero-sum game. Sure, the managers failed. But the engineers also failed, and in the same way the Amazon thief did: they had reason to suspect consequences were coming, but they hung around because it was rewarding.

> Several of the [...] engineers stated their concerns about the O-rings and urged a delay, but the concerns didn't get communicated beyond the Level III Flight Readiness Review (FRR).

I think this is partially a fault of engineers, and/or the development/project model. They were not able to clearly state that no, it's a no-go, or yes, it's a go, they had "concerns". But that is probably useless. Therefore the problem is they lacked the conceptualization framework, a risk model, to state the severity of their concern, what is it based on, and someone should have been able to aggregate it and determine how this estimated risk influences the mission's overall risk.

That said, it is entirely the management's fault that they lacked this critical expertise/competence to manage risk and the project, and the engineers. (And naturally some of those managers themselves were engineers.)

I don't see any evidence in the Rogers report summary that management believed the launch failure rate would be 1 in 10^8. It seems like the report suggests a management belief of 1 in 10^5 (which is still fantastically and unjustifiably optimistic given the complexity and historical track record of space missions, of course).

The typical reliable car fails to about as frequently as 1 in 10,000.

Reply on news.ycombinator.com