Something this paper touches on is that ultimately, we’ve made it so that compilers don’t work for us, because we’ve set them toxic goals.
By making compiled code size and speed the be all and end all of how we compare compilers, we’ve driven compiler developers to USB/IDP/UB hackery that means they can chase the benchmarks we pretend to care about, when really we should be encouraging compiler devs to instead develop for correct-in-practice.
Something this paper touches on is that ultimately, we’ve made it so that compilers don’t work for us, because we’ve set them toxic goals.
By making compiled code size and speed the be all and end all of how we compare compilers, we’ve driven compiler developers to USB/IDP/UB hackery that means they can chase the benchmarks we pretend to care about, when really we should be encouraging compiler devs to instead develop for correct-in-practice.