|
|
|
|
|
|
by Bekwnn
2197 days ago
|
|
|
I think I possibly fall into this category of developer. I'm currently a few years into game development in the AAA space. The people I listen to most in descending order are Casey Muratori, Andrew Kelley, and Jonathan Blow. I try to take their opinions with a grain of salt and find my own balance of considerations/tradeoffs. A common sentiment between them that I agree with is that software is a lot of commercial software is often an order of magnitude too slow in important places and that object oriented programming is almost always taken far too far. I also look up to a few senior programmers I work with who have in-the-same-ballpark values as them. On a whole, I think reading and listening to them has made me a better developer than if I just fumbled around to draw my own opinions completely from scratch. I think being that type of young dev is probably fine so long as you keep your dogmatism and humility in check, and ensure your mentors are a spread of different individuals. I think that's better than other young devs not seeking out information to try and grow. |
|
|
I have a fair amount of respect for game developers. However at the same time I feel they can make the assumption that all development should follow their practices. That everybody should be writing in C++/Zig/Jai/Rust, using data oriented programming and avoiding objects like the plague. That if other developers were just smart enough they'd make everything ridiculously fast and awesome. Not every game developer thinks this way, obviously, but there is a subset.
The problem with this line of thinking is that it assumes all developers have the same priorities. In actuality game development is an extremely rare area where there are close-to real time requirements. Not painting the screen every 1/24th of a second is a disaster in game dev. Meanwhile GitHub uses Rails in all of its object oriented, garbage collected, interpreted horror. And nobody cares!
I suspect if a troop of elite game developers were dropped into your average web development job, they'd spend an unreasonable amount of time optimizing perf, only to have their lunch eaten by someone who focused on user facing issues. Not that web development is hard or anything. It's just a different set of requirements.
Watching jblow et al is definitely great! But there might be a point where you're put on a team using OOP and asking them to refactor to DOP with bulk allocation because you gotta get that cache utilization will just earn you stares and no buy-in. Or you might be running a startup and realize that all the time you spent getting all requests under 100ms could have been spent writing features.