Comment by kunley
2 days ago
Maybe Linus by writing kernels, not compilers, has even more to say, because his use-case is much more practical than anything that compiler designers could imagine.
2 days ago
Maybe Linus by writing kernels, not compilers, has even more to say, because his use-case is much more practical than anything that compiler designers could imagine.
That's kind of a rude thing to say, suggesting that compilers aren't real programmers. Nevertheless, it is the case that most optimizations in the compiler are motivated by real issues with somebody's code, as opposed to compiler engineers thinking up optimizations in their heads.
So, if my sentence is rude, and "take anything Linus says about compilers with a grain of salt" isn't, then I definitely quit this discussion.
I'm suggesting that Linus isn't the best person to be used as an expert in compilers because his field isn't in compilers.
You're suggesting that someone who works on compilers shouldn't be used as an expert in compilers because their field is in compilers.
There is a difference between those two statements, and that difference is what makes one rude where the other isn't.
Your statement
> by writing kernels... his use-case is much more practical than anything that compiler designers could imagine.
States that compiler designers cannot even imagine the practicalities needed to write kernels. Which is quite a blanket statement to try to make and defend.
Compilers are practical too, software tools used all the time. They're just as fundamental to building software as an operating system kernel is to running software. I don't know what point you're trying to make.