Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

It's not insane, the author has been bitten by their poor experiences with dependencies in other languages and is misapplying that experience to Rust out of hand.

Listen, I'd be as happy as anyone to have random numbers in the Rust standard library. Compared to the Rust developers, I'm a believer in stdlib maximalism, downsides be damned. But all this recent hand-wringing about dependencies is a tiresome moral panic.



"moral panic" is a bit of a reach don't you think? Increasing dependencies is a real problem with real downsides. There are plenty of characters expressing unreasonable things, but that doesn't mean everyone expressing concern about dependencies is indulging in a moral panic. There is nuance!

If there weren't real costs to dependencies then I personally never would have published regex-lite.


> If there weren't real costs to dependencies

The OP isn't addressing the real costs of dependencies, the moral panic in question is the automatic assertion that more dependencies is worse than fewer dependencies, which implies that e.g. all the work you have done to cleanly separate regex out into reusable regex-syntax and regex-automata crates has done a disservice to your users. There are real arguments to be made about wrangling one's trusted computing base, but this isn't making that argument, and by throwing the baby out with the bathwater it sets us back as a profession.


> The OP isn't addressing the real costs of dependencies

Sorry, what? The section on "Compilation Times" is absolutely a real cost!

> the moral panic in question is the automatic assertion that more dependencies is worse than fewer dependencies

I agree that is a moral panic, but I disagree that Armin is indulging in that assertion.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: