Why I love FreeBSD

A personal reflection on my first encounter with FreeBSD in 2002, how it shaped the way I design and run systems, and why its philosophy, stability, and community still matter to me more than twenty years later.

https://it-notes.dragas.net/2026/03/16/why-i-love-freebsd/

#FreeBSD #ITNotes #RunBSD #IT #SysAdmin #OS #OpenSource

Why I Love FreeBSD

A personal reflection on my first encounter with FreeBSD in 2002, how it shaped the way I design and run systems, and why its philosophy, stability, and community still matter to me more than twenty years later.

IT Notes

@stefano I run Linux but it is because I have since the mid 1990s. I know darn well there is a much quicker way to say what makes FreeBSD attractive compared to a GNU/Linux.

It is nothing to do with GNU, in my opinion, except as a matter of taste. Some prefer more Unix-like tools, I prefer the more full-featured GNU tools.

But kernels...

Linux is an absolutely HORRIBLE kernel. It is simply atrocious. And the closely kernel-related infrastructure is even worse and getting worse.

@chemoelectric @stefano That's exactly why I'm still sad about the premature end of the GNU/kFreeBSD (https://www.debian.org/ports/kfreebsd-gnu/) project.
Debian -- Debian GNU/kFreeBSD

@catavz @stefano

The HURD would have been very nice, but I gather Mach was not really suitable for the job. I have run Debian GNU/HURD and it wasn’t even reliable in a virtual machine, but it was nice.

There was a story a little while ago about Linus blowing his top over some people wanting PowerPC code in the kernel. And Linus considered bringing in big-endian code now unreasonable. I was like, ‘HE LETS PEOPLE COMMIT ENDIAN-SENSITIVE CODE INTO THE KERNEL?!’ But of course he does.

@catavz @stefano I roll my eyes when I so much as see hash code that is sensitive to endianness or alignment, even if it has a warning on it.

@catavz @stefano

My Unix Users club at the time handed us out the first edition FreeBSD CDs that were withdrawn from sale because they still had some AT&T code in them. Those discs weren’t destroyed, they were handed out to people like me. But I never installed it. I was actually still running DOS on an old 286 then.

RE: https://masto.ai/@chemoelectric/116238827313991924

Who knows what code that was?

Presumably not kernel code, but rather library and/or commands code.

For instance, ex/vi was likely based on ed. And perhaps they still had the old nroff instead of groff. These are two examples that are too obviously fixable, however, because plainly good alternatives already existed.