[JHof]

Exploiting this aviation-related post to see if any aerospace folks will comment on my two-month old Ask - Ask HN: Why is the National Airspace System so dated? https://news.ycombinator.com/item?id=12651244

In regard to the safety of regionals vs. the major airlines that the article's author is writing about: I can see why the statistics are separate. Even though, according to the FAA, we're all under the "same standard", in practice, this isn't the case. As long as the paperwork looks the same, though, we're "safe". Also, within the industry, some pilots tend to think of themselves as shouldering the responsibility of safety, however, I sometimes think we're safe despite the pilots. Engineers of aircraft, airspace, arrival/departure/approach procedures, airports, ATC systems, etc. are silently keeping you safe. That the FAA is able to manage all of these elements is quite impressive. Still, its massive bureaucracy can be quite Vogon-like.

[rplst8]

I think a lot of software engineers would be blown away at how strict software standards are for avionic systems. Like 100% modified condition/decision coverage for all software tests, strict. No pointer math and no dynamic memory allocation, strict.

https://en.wikipedia.org/wiki/Modified_condition/decision_co...

[gurneyHaleck]

Being a party to lackluster software development on a regular basis, I somewhat pine for projects this strict.

As part of my daily routine, feigning surprise at the fact that, given a tree of decisions, I'm only permitted to code for the ones that make money, simply to find that "Quality Testing" reveals the 500 other permutations will inevitably result in thrown exceptions with hideous, glaring error messages.

"WHOOPS."

[rplst8]

This sort of thing is exactly what some of the software development processes are all about.

When I was in a CMMI training class 10 or more years ago, the instructor we had said basically that much. By instantiating a process, it takes the mental weight of knowing (or checking) if everything is correct.

No one builds cars that way - why should we build anything that way. The processes should catch the inevitable human errors.