> I am even more frustrated with the moral crusaders from languages like Rust, one of whom went as far as to suggest that I should personally be criminally pros

> I am even more frustrated with the moral crusaders from languages like Rust, one of whom went as far as to suggest that I should personally be criminally prosecuted if some downstream Hare software has a use-after-free bug. My goal is not to force anyone who doesn’t like Hare to use it, or issue judgements upon projects which choose another language. In return, I will be pleased if members of other language communities refrain from flaming too much on Hare. This might be referring to comments elsewhere, but I thought there was a pretty thought-provoking debate about safety tradeoffs in the Hare intro thread.[1] Which I'd summarize as: let's say we now know how to prevent, say, 70 out of 100 security bugs in C codebases, without performance compromise, by statically ruling out things like buffer overflows and use-after-free; and we also have good evidence that bugs your language ecosystem is bad at detecting are hard to backport detection for. Is it a good idea to make a language that prevents _most_ of those 70 mistakes, but not all that we know how to prevent, in exchange for being simpler, and therefore reducing the other 30 mistakes and getting more software done that helps people? Or would it be better to avoid investing in or relying on new languages with that tradeoff for infrastructure code, and focus on seeing how simple a language can be that prevent all 70? Which isn't a logic question, but an engineering question: does the mostly-safe language prevent 65 out of 70 memory bugs, or 30 out of 70? Does it let you get twice as much done as the safer language, or 10% more done? Does it result in fewer logic bugs than the more complex language, or the same number? I don't know, but I'm interested, because I want the next billion lines of code that affect me to do useful stuff and not break. "My goal is not to force anyone who doesn’t like Hare to use it" isn't really an option; I'll be impacted by all the code people write in every language. So: I'm happy to see people make new things that test a new point in the design space! But I'm _also_ happy to see other people say, wait, before I end up with a ton of this code tucked into the lower levels of my machine and the other hundred billion machines wired up to it, what mix of features would convince me that "less safe than we know how to make new languages" is still safe enough for a new language in this case? [1] https://news.ycombinator.com/item?id=31151937 https://news.ycombinator.com/item?id=31151937