But they are kind of hypocrit, since feature creeps, complex or not, will break compilers via planned obsolescence on 5-10 years cycles.
From a reality point of view this 'backward compatibility' is actually a trojan horse for those feature creeps. This is this very toxic cycle which has to broken.
Compilers have already standard version switches, namely the migration path towards saner C syntax is to leverage such options: -std=c11 will allow syntax from "legacy C below c11", and for instance -std=c25 will start to enforce a simplified and saner C syntax. Adding feature creeps or fixing C syntax will have the same effect on real life code, but in the case of fixing C syntax, it is going to make compiler easier to write and real-life alternatives will spawns, which is is very good and sane thing.
From a reality point of view this 'backward compatibility' is actually a trojan horse for those feature creeps. This is this very toxic cycle which has to broken.
Compilers have already standard version switches, namely the migration path towards saner C syntax is to leverage such options: -std=c11 will allow syntax from "legacy C below c11", and for instance -std=c25 will start to enforce a simplified and saner C syntax. Adding feature creeps or fixing C syntax will have the same effect on real life code, but in the case of fixing C syntax, it is going to make compiler easier to write and real-life alternatives will spawns, which is is very good and sane thing.