Was! It was fun. But at that time Airbus was not paying that much so I changed industry after some time.
True! That's interesting to see the similitude with trading automation. I still fly manual. But I know the huge potential of flying automation. I admire people who reached that stage and succeeded.
In Flash Boys, Michael Lewis describe the Goldman Sachs complex trading system. For each error that happens, the part concerned is patched. Sub-systems are added when new functions are required.
By accretion of all those modifications during all those years, it becomes a system incredibly complex, redundant, messy and slow, that no IT guy really understands in its entirety.
Therefore nobody is able to modify it in its core to make it simpler, more efficient and competitive against HFT firms.
I believe the more people are involved, the more unnecessary complex it sometimes tend to be . Let's hope that it won't be the same case for airplanes automation systems.
I've never worked as a programmer exclusively, so I cannot comment what it's like in a team but I can't imagine how difficult it must be. Few people I know in the industry have commented that there's always this one guy who writes code that no-one else understands, I suspect I'd be that guy as I don't use testing/assertions and I rarely comment.
There are many approaches to doing the same thing, sometimes they are equally valid. Now if you have a team of coders each writing functions according to their preference and philosophy, it can become really bad. Then again, as Github projects are proof of, it can and will work. I've participated in a few and as long as there are only a few main maintainers to keep the structure in check, it's not as bad.