I had a lot of tabs to close, if you can’t tell.
- ClonOS 18.12 BETA1. I didn’t know this existed.
- Next NYCBUG meeting: January 2, with a presentation on OpenBSD from Brian Callahan.
- FOSSJobs. I don’t like the acronym, but I’m not going to complain if it gets someone good work. (via)
- Running NetBSD on DEC VAX on SIMH on Raspberry Pi. Video. (via)
- The Future of ZFS in FreeBSD. (Via many places) I don’t like the notion that the BSD implementation of ZFS is now going to be at the very best, only equal to what’s on Linux. That makes it no longer a compelling choice.
- OpenBSD song 6.2: A 3 line diff. (via)
- OpenRC on FreeBSD. (via)
- FreeNAS 11.2 out. (via)
- Related: December FreeNAS Plugins Updates.
- SoloBSD 12-STABLE.
- Stable release: HardenedBSD-stable 12-STABLE v1200058. (via)
- OpenBSD Errata: December 22nd, 2018 (pcbopts). (via)
- Grafana + InfluxDB fun on DragonFly BSD. Okay okay not other BSD.
- Problem booting after upgrading to [DragonFly] 5.4. The story has a happy ending.
- The Super Capsicumizer 9000. (via)
- What are you using for a simple GUI TextEdit/Notepad? The usual suspects of nvi/emacs show up, but there’s a lot more options listed.
- Reducing the delta with upstream version of sanitizers. (via)
- Python Development on BSD.
- Using the GOG.com installers for Linux, on NetBSD. (via)
- Torchlight 2 on NetBSD *almost* works!
- NetBSD desktop pt.5: automounting with Berkeley am-utils. (via)
- Valuable News – 2018/12/14, Valuable News – 2018/12/21.
- Standards. “There is no standard version of the tar program.” Linked with early BSD history, of course. (via)
- telnet, still being a problem. (via)
- protectli router.
- de facto vs de jure maintenance.
- The NetBSD support update before the LLVM-8.0 branching point. (via)
Justin,
I believe that FreeBSD’s version of ZFS now being derived from the ZOL project is probably a necessary step, but it is likely to be downhill from there onwards for FreeBSD. The sheer amount of manpower available in the Linux community (compared to FreeBSD) to implement features in ZFS cannot be overstated. Because of this there is likely to be a lot of Linuxisms and other features that would need to be reworked to support FreeBSD resulting in additional challenges. Its like with DragonFly and graphics support. It might make more sense to eventually make some of FreeBSD’s subsystems more Linux-like to not have to constantly reinvent the wheel. At that point what is the need for FreeBSD? FreeBSD’s “killer feature” for the last several years has been ZFS. I wonder what others think. I imagine there is quite a lot of discussion going on about this throughout the BSD community. Even more scary, what is NetBSD going to do with their ZFS port?
Can Linux boot from a ZFS partition yet? Maybe it doesn’t matter. I suspect that as the protected UEFIs come on that everyone will be booting from an MS-DOS partition anyway, the way Raspberry Pis do.
I still rate FreeBSD’s init/rc framework and lack of cgroups as my favourite features. The vague notion that it’s at least possible for my feeble mind to understand what’s going on, all the way back to the start. Oh, and that every running program knows where its source code is, so that I can point a debugger at it, should that prove necessary.
Oh, and the way that all of the Linux distributions that I know make it so hard to stay within cooee of up-stream application development, thanks to their one-shot notion of “stable”.
There’s a tricky balance between having unique features and having enough manpower to implement something. Unique features need coding time, which every volunteer project is short on. (I was going to say “open source project”, but Linux is not a volunteer program at this point, and that’s important.)
When it’s software that’s on every system and doesn’t drive adoption, you don’t need to roll your own version. Everyone has xterm, for instance. ZFS was a competitive advantage for FreeBSD.
On the other hand, there’s already porting work happening on Linux, so it may not have stayed a competitive advantage, and therefore those development hours would have been wasted duplicating what happened with ZoL. It’s not an easy choice to make, but I can see where it’s coming from.
Very fair assessment Justin, thanks!
If Linux won, then which distribution should be used?
BSD of course.