Bitrig, a fork of OpenBSD, has surfaced. It aims to be less conservative than OpenBSD by being less loose with new features and only supporting modern architectures. Bitrig has also converted over to using LLVM‘s clang rather than GNU gcc for it’s toolchain/compiler. A roadmap for Bitrig is also available.
Very interesting to see this!
Sounds like a promising project, so it’ll be great to see what they come up with. PC-BSD has done some great things using FreeBSD as a base, so maybe this will eventually become somewhat similar (but for OpenBSD).
It sounds like they still want to have a small codebase, so that’s good IMO. The one thing that I’d love to see with PC-BSD is a small variant of it – CD-sized rather than DVD-sized.
You are comparing derivates (forks) with distributions. From what I see this project appears to be an independent operating system which usually is frowned upon by BSD people, unless there is a good reason – else it’s just a waste of man power.
Well, it’s these people’s life, so they should do what they consider right, but supporting fewer platforms and else just wanting to enhance the system doesn’t really sound like a good reason for forking. See DragonFly for example. They actually headed in a different direction and the main developer’s commit bit has unfairly been removed. That’s a good reason.
I know, OpenBSD has Theo, who is like a feature and a curse sometimes, because he has very strong opinion. This made OpenBSD conservative, which usually is something that OpenBSD users know/expect/love. If not then there still are projects like NetBSD (OpenBSD is a fork of NetBSD, in case someone here doesn’t know), which actually would love new developers and have very similar goals, except not supporting fewer hardware, which again really isn’t something I’d consider a feature. The maintenance “problem” has been fixed by the NetBSD project with recent developments and they have some great new features, no other OS has oh and by a more sane measurement (actual number of vulnerabilities instead of that default install thingy) they have fewer vulnerabilities in base. Also they don’t have a ports/package tree with tons of outdated, vulnerable packages.
Sorry, if that sounds too negative. OpenBSD is great. I have it running on a PA-RISC where the only alternatives are HP-UX, which I don’t really like and Linux, which doesn’t really support it. OpenBSD has a security hype and is a lot into politics. I like that they are true to their philosophy in regard to being open (more open in fact than pretty much any Linux distribution, since they are strictly against blobs and don’t enforce the whole GPL upon you, but use the simple short, easily understandable ISC license).
Again, I think it’s great that people are doing these things, but were they really unable to get their stuff into OpenBSD and why didn’t they join MirOS? What’s their target group? Why not use another BSD?
I think answering these questions more precisely than they do now would be a good idea if they want to get both users and developers.
Edited 2012-06-13 14:43 UTC
My thoughts exactly. I don’t really know what’s and who’s behind the fork, but yes, their goals are hardly a valid reason to make a fork. Removing support for legacy platform is hardly modernization and the switch to Clang/LLVM is hardly going to bring any advantage to the end user… it is mostly done due to the license and the ones who would gain the most from the switch are companies.
However, there is a lot wrong with OpenBSD when it doesn’t come to security (first and foremost, unsustainable release and updating scheme and very poor backward/forward compatibility).
By the way I am a NetBSD user. The last stable RELEASE really doesn’t do justice to NetBSD, because 6.0-BETA and CURRENT have a lot of nice and useful features that are just missing in 5.1.2. I like NetBSD and DragonFly because they don’t force you to follow them in everything. FreeBSD and OpenBSD on the other hand are obsessed that the end user has to follow them, and often do things that have no advantage but are just BSD-licensed rewrites.
How is limiting of options (say, blobs) more open, and how does Linux enforce the whole GPL upon me? (I do whatever I want with it…)
Yes, their stance on specs and drivers is great (as are some other things, like OpenSSL or PF), but…
BTW, thanks for that info about NetBSD (also vs OpenBSD & vulnerabilities), from the outside it sometimes seems stagnating (though OTOH, NetBSD – not ~OSX – powering Apple AirPort Extreme and Time Capsule is pure win ;p )
I’d rather see Anonym.OS resurrected.
Interesting news indeed. Curious about the names on the “Recent Changes” page, Marco and Dhill. Marco Peereboom (of Conformal/Cyphertite) and David Hill? I can’t really check further as I’m surfing from the phone right now, but if it is those 2, then I guess the project is in good hands. They’re pretty known on OpenBSD mailing lists, and are also responsible for my favourite tiling window manager, Spectrwm.
Will take a look at the snapshots later in the day when I get home. Looking forward to it.
I do hope they are competent. A system is only as secure as its least secure component (like a chain is only as strong as its weakest link), so it would very easy to degrade OpenBSD’s main (and almost only – sorry) selling point.
Been studying the sources and visited their IRC.
It’s the same Marco Peereboom.
https://opensource.conformal.com/wiki/Main_Page
Pleasantly surprised their LLVM/Clang support is already so complete.
Mm, I have mixed feelings about this…
While the people working on this seem to be more competent and sane than the people behind MirOS or AerieBSD, the kind of fork and all the name whoring means we won’t see any more contributions from them to OpenBSD and this is sad.
I for one rely on scrotwm and xxxterm. In prespective, all the silly name changing was likely in preparation for this project. The real names made it hard to sell a product.
I wish them luck. On one hand, the OpenBSD codebase is very good, on the other, the “embedded” support will require some pretty big structural changes if it has to compete with real embedded OSes. At this point it can’t even stand a fight with the massive non-RT Linux market. And what will the PHBs think of security when they learn they need to upgrade to a smartphone processor to get the same performance they had with an SH4.
…
Reading the FAQ, I wonder if something as simple as a VAX backend for LLVM could have avoided this fracture.
I’m the only one who see that the servers (at the time of writing this) of this fork of OpenBSD are located in U.S.A.?
Nobody remember why OpenBSD main servers are located in Canada?
If not remember the t-shirt.
http://www.openbsd.org/tshirts.html#7
I hope they made a great O.S…
Yet another BSD fork…
At least this time they’re using a good base and have useful goals.
I have read up a little bit on LLVM. It has so much potential but all I see here is that they’re using it to compile, as a GCC replacement. Not really utilizing the JIT or platform independent LLVM IR.
Basically, they have switched compilers. Nothing spectacular.
I’d like to see a distro (BSD, Linux, whatever) with a package manager that serves LLVM bitcode and the source is JIT compiled / cached upon installation.
When a new LLVM is released you could recompile everything on your system to take advantages of new optimizations.
Where would I get my bi-annual dose of Stockholm syndrome?
JK, love OpenBSD on older laptops.
OpenBSD is pretty solid *nix OS. What they say works, and what they say doesn’t work … doesn’t work.
Unlike Linux where due to the distro framentation.
I hope they get rid of the big kernel lock. I am not sure how packages and ports will work.
Edited 2012-06-13 18:12 UTC