HN Explorer

4 comments

[ 3.7 ms ] story [ 23.3 ms ] thread
gnat
tl;dr: "I don't know the specifics of the 737Max but I do have prejudice against a methodology of software development that mistakenly write like it's an acronym. Let me share that with you."
virken
i'm a private pilot and a software product manager - and i similarly don't know for sure - but think your conjecture is completely plausible - agile is fine and does have cost/productivity advantages - but i rather doubt you'll see the textbook version of it implemented by NASA - but everyone in commercial software is falling all over themselves to reduce software development costs and Agile is supposed to be the holy grail - I've seen it work just fine for line of business software - but I similarly share the concern that manufacturers of airplanes, medical electronics, power infrastructure are adopting it for cost-savings purposes and are not sufficiently concerned about the edge cases where it goes bad with devastating consequences - but that is the hallmark of capitalism - shareholders come first...
chungleong
An iterative process makes it impossible to hold the line on safety, I think. In each cycle it'd get compromised a bit, given how both the producer's and the client's overriding concern is cost. Waterfall is the natural model for software that deals with regulated activities but the word has somehow become an epithet.
NikkiA
There has been, IMHO, a gentle decline in the quality of software from an engineering point of view in the last 20 years, and while I think agile may have contributed to this, I also feel agile is a symptom of this, and that the root cause is that the 'war' between 'software engineering' and 'coding' as the act of production of software that happened in the 90s, and was largely lost to those that see it as 'coding' and 'an artform'.

That the aviation software engineering community has moved from Ada towards MISRA-C is also another hybrid symptom-cause.