I don't know where you are trying to go with this logic. You still haven't assured that the 50MB is actually related to data structures used behind by bevy, lest you are insunating that the data structures used by bevy are solely responsible for that size, which is not what the post claims.
How small is a hello world in C on Linux? The last time I checked a few years ago, it was either 9Mb or 13Mb (memory is fuzzy on this one but one of them was right).
Edit: wow I’m way off… it must have been the 90s when I checked this lol
Yeah, testing here with GCC 13.3.0 led to 16K. Perhaps, when you made your test, the compiler was statically linking some libs, which would explain the massive increase in size.
I know the thread. Your point is about whether Bevy's "data structure" is worth 50MB. I pressing on the fact that you haven't shown what this 50MB actually are.
But, out of curiosity, do you also consider the Linux Kernel also a data structure?
> To be clear you think an ECS implementation is 50MB?
No.
> Do you think a spreadsheet the[sic] holds a function pointer is a kernel?
Do you think an entire engine, including its system scheduler, is a spreadsheet? If so, shouldn't be hard to apply the same argument to any kernel. Of course, it all relies on your definitions and sense of reductionism.
I think you're conflating other big features of a game engine with a way to store data. If you look at other ecs implementations they are data structures, they don't do "system scheduling" (is that different from regular scheduling?)