So I’m pretty outside the *BSD ecosystem. From a user perspective, what makes DragonFlyBSD different from/nicer than the others? The homepage mentions virtual kernels (cool!), a new filesystem, dports, and the use of tokens as an internal synchronization mechanism.
Compared to Linux, most BSDs are rather conservative about changing their ecosystems.
DragonflyBSD is a fork of FreeBSD that is more experimental and bleeding edge. They have a strong thematic focus on parallelism, including better symmetric multiprocessing support in the kernel and a cool new filesystem “HAMMER”.
Culturally, DragonflyBSD seems to inherit a lot from the AmigaOS community.
Which is smart because both Amiga and BeOS kicked the crap out of UNIX, Mac, and Windows in those areas. BeOS being the overall winner on tech side as it also had a safer, concurrency scheme + advanced filesystem. Both turning into forgotten, but useful, parts of history. I immediately thought of them when I saw DragonflyBSD’s proposal and smiled when I found out at least one inspired it. They took notice of something in ancient history, decided those capabilities should be in BSD, and proved it could be done with a fork. A very useful capability on top of it given multicore era & CPU + HW accelerator SOC’s that are basically Amiga model evolved.
BeOS is not entirely forgotten. The Haiku project has achieved full binary compatibility with BeOS, and is working towards its first production release. The project includes an implementation of the BeFS, including all the relational features. They managed to leverage FreeBSD’s work on network drivers, so support all the same ethernet and wifi cards as FreeBSD. Other drivers are slow coming, and overall progress can drag a bit. But I’m still pretty excited about it overall.
Oh yeah, there’s that. I meant in the larger scheme of things. What I like about Haiku is… aside from building on BeOS legacy with new, fun OS as you said… is that they use the MIT license. With MIT or BSD, it might be easier to get support from various companies that steer clear of GPL. On top of potentially dual-licensed model where a company sponsors work on it while differentiating with a few, key features. Haiku could do that if it got stable enough for a corporate version that leveraged its benefits. Also, THEOS uses a lack of software for their platform as a differentiator said to boost productivity and security. Haiku, MorphOS, and others might try that in a corporate version.
I submitted this to HN if anyone would like to help dfly gain a bit more attention: https://news.ycombinator.com/item?id=12219924
Jumping over from here and voting on that might (rightfully) trip the voting ring detector.