Porting Wine to amd64 on NetBSD, first evaluation report
blog.netbsd.org
blog.netbsd.org
I hope they learn skills that will be helpful down the road.
Are there many users of NetBSD on amd64? What's your use case?
I absolutely applaud people sinking their minds into interesting niche projects like these. You never know the macro results 10 years down the road.
Would you be able to speak to the differences between OpenBSD, NetBSD, And FreeBSD? So far as I can tell, FreeBSD is generally focused on getting the most software and the most drivers, etc, in order to compete with Linux, while OpenBSD focuses on implementing everything in the "correct" way, and NetBSD focuses on being a lightweight OS for routers and the like. Is this about right?
1) people scratch their own itch; if something is important or interesting, they do it
2) the engineers are smart (though which project is going to say “I think engineering is our weak spot”?) The decisions that are made happen to mostly resonate with what I’d say are my values - pushing ahead with a semblance of respect for Unix.
3) NetBSD is (and has been) a research operating system. Microsoft used it to test out a new hardware arch they developed[0][1]. It’s (twice) held the “Internet Landspeed Record”[2]. It’s cutting edge in the unikernel/anykernel[3] space.
4) re: portability (“of course it runs NetBSD!”[4]), I like to think there’s a feedback loop where correctness and good design support porting and porting demands that correctness and design. Example when AMD first unleashed the now ubiquitous 64 bit architecture we all enjoy[5].
I don’t know what kind of development you do, but it supports gcc, clang, and PCC, Java, probably any dynamic language you care to name... it’s an SMP POSIX system running elf binaries on big and little endian architectures, all from a single codebase. I see why people would say it’s a stripped-down OS, or built to run in weird places like routers, but I see it as an effective, coherent, non-offensive Unix. I could go on, but let me close this with tongue planted in cheek, stealing from mutt[6]: All operating systems suck. NetBSD just sucks less.
[0] http://wiki.netbsd.org/ports/emips/
[1] https://www.microsoft.com/en-us/download/details.aspx?id=524...
[2] https://m.slashdot.org/story/45653
[3] https://en.wikipedia.org/wiki/Rump_kernel
[4] http://www.informit.com/articles/article.aspx?p=655422
[5] https://www.usenix.org/legacy/publications/library/proceedin...
The issues show up because I'm a hyperpolyglot when it comes to programming languages. So for instance I'm thinking about trying J (which I've messed with in the past) and Idris. As you can probably guess, having a common OS is nice for this.
You sound like me. These days I’m C and Tcl w [ex]ctags and cscope in nvi (in-tree default, but vim available in pkgsrc/pkg), using dtrace, gdb, fossil/git, etc, etc, in tmux, either in the terminal or an xterm managed by dwm. Firefox and Rust (a dependency!) are available. I used to build them from scratch but recently Rust has turned into a real bear to build... like building Linux from Ye Olden Days where you set aside a few days and hoped... but I digress - [X|GNU]emacs are there for you too.
If OpenBSD seems attractive to you, jump in :)
Give it a fair few weeks and evaluate. They’ve got some neat architecture features to play with (e.g. unveil(2), pledge(2)[0]) to exercise your software design/implementation. NetBSD and FreeBSD have dtrace[1] (NetBSD impl. is incomplete but still interesting and useful) and virtualization like bhyve[2] and nvmm[3]. Jump in!
[0] https://news.ycombinator.com/item?id=17277067
[1] https://en.wikipedia.org/wiki/DTrace
[2] https://wiki.freebsd.org/bhyve
[3] https://www.phoronix.com/scan.php?page=news_item&px=NVMM-Net...
Good luck!
Not even close.
They all are general purpose BSD unix operating systems. All of them are interested in correctness, performance, portability and security (but not in the same order).
The reason why BSD forked has more to do with politics (the seek of power and control) than technical differences.
In order to justify why there so many forks of the BSD kernel it seems to be an urge for exalting their differences when the main reason is simple politics.
What's your use case? Use my computer.
It's a good solid unix system without pretending to be something else.