Assuming all of that is true (which I doubt since there are many critical systems that are likely more complex than video game software) it still points to poor management. Video games have the luxury of not having to be precise in essentially any aspect. The margin of error in capturing the environment (mouse clicks, controller input) and producing output (small graphical artifacts are unnoticeable) is huge yet these guys are frequently blowing past it. If you know you’re building something complex then you need to allocate sufficient time and resources to do it right.
Imagine if every other commercial SpaceX rocket exploded. They couldn’t exactly brush it off by saying rocket science is hard.
1) I am not brushing it off by saying it's hard! Things have gotten bad and need to be fixed. The issue is
identifying where the problems lie and what fixes are possible. To understand that, you need to understand gamedev and I don't think any armchair analysts can have any idea of that without trying. There's a reason why after 30 years of software engineering theories and strategies, gamedev is still flaky, and it's not because the people involved are inherently lazy dumbasses.
2) In explaining gamedev, understand there's different types of hard. Software development is spectacularly hard because the systems are unpredictable and
uncontrollable.
If SpaceX designs around a part, and then that's not available, they source another part. They then have all the documentation and specs on that part, and the certification so they know it'll work within defined limits and tolerances. It's engineering; it's known quantities.
Now we all know nothing in life every goes that smoothly and Shit Happens, but AFAICS software operates on a whole other level.
In a game, you send code to the API with no idea what it's doing. You can't even know - it's too complex. You don't even know the lanugage you are using necessarily because the latest updates to c# or whatever have changed this, that and the other. You keep refering to the every-changing online dooucmentation and hope its comprehensive, which it isn't. You just trust the IHV to maintain a driver that can deal with your code. And they in part are writing around an OS they have no control over. And then there are constant updates that break things. Apple rolled out an update to Metal that plain worked differently and broke code. There's no way to solve these issues other than to Sherlock Holmes the shit out of what's happening. And the pace of change and cost of development means you cannot operate software at the same level of industrial engineering as other complex systems. If you stay outside of the complexities of game dev and stick to app development within the textbook industrial, you stand a chance. Gamdev always exceeds predictions and then the run away to try and pull it all together at the end, not unlike a theatrical production where the whole thing is falling apart at the Dress Rehearsal, hits impossible errors from nowhere that snowball issues.
eg. In a play, someone forgets their lines, you can
ad lib. You can recover on the fly. Many jobs, you can shoe-horn a fix in. Gamedev can't, and the hotfixes just pile junk and weigh the whole thing down.
It's not a pissing contest with gamedevs laying claim to "our job is the hardest in the world". It's trying to understand how epic problems arise, especially consistently recurring ones. "How can devs let this happen?" Well, let's actually find out instead of just saying they suck and should do better. Where do you even lay the blame - the game studio, the OS, the GPU company? All of them? Who has to make the fix and how? Education? What needs to be taught when? What needs financing? What's the cost and where's the money coming from?
www.ea.com
