dtw2hyd wrote:brindabella wrote:Also a significant additive to successful software production:
1. You write Program "A", and I test it.
2. I write Program "B", and you test it.
Early-on after the ET accident I was appalled when a member here described the MCAS software getting major changes in parameters very late in testing.
One would like to think that a separate programmer doing the testing would have rapidly come to a "what the f... is this!!" stop point before throwing the junk back
to the originating programmer.
cheers
You seem to have very high expectations. Modern day software design, development and testing methodology is sketchy at best. Not in just aviation.
The problem is not the software design nor where in the world it was written. What is the problem is that most probably non aviation programmers were hired (what is legit) and they do not care if something activated repeatetly if not stated it should not. The dealbreaker is that the engineers that wrote the spevifications or the people reviewing were just not up to the job.