Hacker News Reader

Comment by vegadw | Hacker News Reader

On the one hand, this sounds like a lot of money, on the other, it's ~765k USD. If you just go off the rough assumption that 1 engineer/100k USD/yr then it's either a team of 7 for a bit over a year or a single engineer for a decade or so. I don't know if that's enough to make a significant impact? I'd like to think so, but a project that big, once it gets funding, is sure to have some tech debt that needs paid before anything new, so, will new things come of this?

View on HN

There are a few exceptions, but overall the BSDs have very well engineered and maintainable code. They don’t move as fast as Linux distros do partly due to less manpower and partly due to taking the time to do it the right way. There is much less churn and they don’t have the cruft of supporting obscure UNIX flavors that have faded into distant memory like GNU utilities do. Take a look at some FreeBSD code and their GNU and Linux equivalents and you’ll see what I’m talking about.

Interesting! Does that make it perform worse on some modern hardware due to having, say, worse scheduling for CPUs with multiple CCXs or worse graphics drivers? I'm obviously wandering a bit outside my knowledge here.

Yes FreeBSD does generally benchmark lower than Linux. The BSDs typically prefer correctness and maintainability to aggressive micro optimizations. There’s nothing wrong with either approach, but I can usually follow FreeBSDs code much easier than I can GNU coreutils or the Linux kernel for example.

The BSDs also have excellent documentation. Every utility, kernel module, and system configuration file have robust man pages.

Where are you getting this from? Most benchmarks I’ve seen put FreeBSD on par with linux.

https://www.phoronix.com/review/bsd-linux-threadripper-7980x...

Fascinating and I stand corrected!

Reply on news.ycombinator.com