This might be off topic, but the illustrations in the article are nice! Crux really did a good job there.

In regards to the actual article: as someone who primarily works in higher abstraction level languages (Java, .NET, Ruby, Python, ...), i find it really nice that i can spend more time thinking about domain/business concepts and problems, instead of worrying about the lower level concerns. My hat's off to the people who valiantly work on the lower levels of the stack.

The combination of the following seems like a key problem:

1. Refusal to break ABI compatibility (so no changes to existing APIs if they cause breakage)

2. Refusal to introduce replacement APIs which can exist alongside the old ones (so you also can't have, to use the article's example, "regex2")

In isolation, either of these policies is fine, but having both together seems like a refusal to improve.

Perhaps true, but not the main topic of this post. C and C++ are at the foundation of a great deal of software, and that creates a lot of pressure for stability. This is blocking evolution of both languages. Stability is great, but stagnation (sooner or later) will get you.

Sooner or later we're going to have to accept that C family languages are fundamentally broken and unsafe at any speed for new development.

Halfway through that horror story I found myself thinking, "Egad! I like rust now."

Finally we’ll be able to rewrite everything in Rust and Haskell and Ocaml and Ada and ATS with built in theorem provers to enforce correctness.

It helps when fed up with the economy loss caused by security exploits, governments finally start making the companies liable for data losses and ransonware attacks.

Then management will magically have the bugdets for security, or close doors, just like many dirty kitchens have to, after an health inspection visit.

But that's not a true statement, is it?

What about non-binary banshees?