Slashdot Mirror
Linux Gurus and OpenStep gurus collaborate
Anonymous Coward writes "www.DarwinLinux.com is a collaboration between the team at Infomagic (Linux gurus), and the team at TwinForces (WebObjects/OpenStep gurus). The goal is to create a complete distribution on a CD. Incidently, Apple will sync Darwin (Apple's BSD core under OS X) with FreeBSD later this year. "
They mentioned in an email to Omnigroup's MacOS-X admin & developer lists that they wanted to call the site "The Missing Link", missinglink.com, to go with the "Darwin" theme. However, the URL was taken (by a horseracing program, oddly enough).
So, I think they settled on "DarwinLinux" because they're a collaboration between Linux experts and Openstep (hence, Darwin) experts.
I agree with you that Apple is every bit as bad as Microsoft, and probably would have been worse if they had the position of dominance that Microsoft does.
That said, remember that pretty much the entire top management and technology team at Apple these days is from NeXt, not the Apple of yore. (Like 6 of the 8 VP's). Next pretty much did a takeover of Apple, not the other way around.
And Nextstep/Openstep is all about interoperability. This new Apple *did* release the source code to previously proprietary technologies, like NetInfo, the distributed user information database, and the Objective C language runtime. They *have* resubmitted improvements and adjustments they've made to things like Apache and Mach.
That said, I really wish they would move to a Mozilla-style development model for Darwin. As it is now, they work hard on the core OS, and every once in a while, release the code as Darwin, version whatever. This paralyzes third-party development on Darwin; nobody wants to touch the kernel or anything else, because Apple keeps saying, "we're going to release totally different veresions of this in a couple of months."
If they moved their own source code tree into the public, then people could actually collaborate and help and add value. As it is, Darwin will just be a pile of code that some curious Mac users download the binary of, and a couple of developers submit obvious bug fixes for.
IMHO, the whole concept of releasing Darwin core into open source is stupid unless Apple also releases the OpenStep/NextStep/WebObjects stuff
The OpenStep/WebObjects/Cocoa is Apple's market advantage: if they open this then Apple has nothing really left to sell. Such an act spells "shareholder lawsuit" - so don't be silly and rash.
Additionally, many Apple developers ASKED Apple to open what Apple now calls Darwin: Developers have specific use for it (developers can get some apps quicker to market because of Darwin - and this has already happened; some devs wanted to improve upon drivers, etc). So it clearly wasn't a stupid move by Apple.
All this is going to get us is yet ANOTHER X-based Unix clone
Actually, it's a varient of BSD, so it's not an all-together new type of Unix. And we all yell HURRAY when there's a new Linux varient. Why's another BSD varient bad?
And as somebody else wrote here: "Can't you see that this CD may become a huge success once Mac OS X client ships? It will be an easy way to add all the Unix stuff to Mac OS X [client/server] that Apple didn't bundle with the OS." In other words - this DarwinLinux CD allows OS X users to quickly get a large collection of standard unix utilities. This is a Big Deal for OS X power users.
And if its based on Mach, why the hell are they calling it Linux
Nobody called it Linux - not even the folk at LinuxDarwin.com. It's clearly a BSD varient. READ THINGS BEFORE YOU JUDGE THEM.
You make it sound as though Apple is the creator of DarwinLinux. They're not. They clearly stated that Darwin is simply a BSD/Mach UNIX. The "DarwinLinux," aside from the fact that it uses Darwin as a code base, has no direct relationship to Apple.
The current Mac OS can't be open-licenced, even if Apple wanted to. OpenStep in Mac OS X relies far too much on third party material, most notably Display PostScript. The GUI of Mac OS 8.x isn't based on UNIX at all, so open-sourcing that would be a nightmare. When the Apple-only Mac OS X Consumer is released, Apple will at least have the ability to open-source.
Further, Darwin included things which were *not* already available. These include HFS+ support, AppleTalk support, and several other items. In other words, it *does* contribute new things.
If it WERE Linux, there should be no need to "port" the software. All this is doing is diminishing the value of the Linux brand.
Posted by JPerlow:
IMHO, the whole concept of releasing Darwin core into open source is stupid unless Apple also releases the OpenStep/NextStep/WebObjects stuff (the gui and the libraries are missing, hello)they have for Intel (yes, you do remember NextStep/Openstep for intel, dont you?), because you wont be able to run any of the cool new MacOs applications or ported NextStep/OpenStep stuff without it.
All this is going to get us is yet ANOTHER X-based Unix clone (after they've ported all the good stuff we already have in Linux and FreeBSD)
and yet further balkanization. This is a total waste of time when open source porting efforts are already being made for Linux and FreeBSD. We have these things already, they are mature and we like them.
And if its based on Mach, why the hell are they calling it Linux??????
Looks like DarwinLinux is just a porting effort to get the missing Unix software onto Darwin. These are just generic Unix stuff and not Linux-specific. Also no Linux kernel resources are involved. So why use "Linux" in the name? Just to get attention?
Free Software: the software by the people, of the people and for the people. Develop! Share! Enhance! Enjoy!
Hmmm,
Apple can't release the GUI portions under a free license since they don't hold copyright on large portions of it...Display Postscript. There's some NeXT-specific code in there (interceptor, and some compositing operations) but most belongs to Adobe. Adobe refuses to even license DPS for YellowBox on NT. IMHO there are nefarious reasons for this.
Instead, Apple is developing Quartz, which is a derivative of Adobe PDF with the compositing stuff thrown in, as well as anti-aliasing and the like which were not part of DPS. I'd assert that this code will also not belong entirely to Apple, and will therefore never be opened until Adobe itself is forced to open code.
This is the primary argument I have for corporations to endorse open code. In no other business that I can think of can a company pull the rug out from a (potential) industry simply by withdrawing a product. If Ford decides to stop selling cars with electronic ignition systems, your existing Ford will keep on running (well, as well as any other Ford). Adobe pulled DPS, which forced DEC to drop it on their products, and will eventually force Sun, IBM, and others to as well. Anything that depends on it (luckily, not much) will die with it next time you upgrade your commercial X server. DEC has done the same with Pathworks/Mac (VMS File Services for Macintosh) on V7.2. You upgrade, your Mac file services are no longer supported. This stinks.
I personally haven't agreed to the APSL, since I have no hardware that it will (yet) run on. I have browsed the DarwinLinux directories though and have found that there are i386 directories and even some EISA driver code out there. I think Apple has provided the Intel stuff, it is just broken and out of sync right now. Someone will no doubt get it going on Intel soon. I hope that their effort pays off for both Apple and their customers better than any of them could have expected...this would represent the best possible turn of events as it will build pressure inside and out of Apple to open more code.
Lastly, I'd like to point out that the lack of GUI on Darwin represents fertile ground for some of the interesting GUI projects that have been in the works. Maybe the Berlin people would be interested? A native DGS would be nice for GNUStep... Hmmm. Some unclaimed territory...
Being scared away because something is BSD-derived or has a license other than the GPL is pretty closed-minded. Go check your favorite Linux distribution and you'll find a lot of stuff with some "Regents" copyrights, lifted from BSD.
It is quite true -- most of the kernel of OS X comes from Mach (although as I noted the networking code has been taken from NetBSD, has has the userland). However, accuracy is rarely /.'s strongpoint. Reporters are supposed to always make sure they get the details right, but unfortunately, /. often skimps a bit too much in this regard. Fact checking would help a lot.
The OS X userland and TCP stack come from NETBSD, not FREEBSD.
If you don't believe me, look at the Darwin source code yourself. We've been working with folks from Apple for a long time, and we've been importing most of the improvements and bug fixes they've made so the source bases stay in sync.
I'm sure Apple would love to time travel back to 1985 and take up Bill Gates offer to make the MacOS the IBM compatible GUI of choice.
However, now their stuck in the position that they have to stay in business by selling hardware. Even when they were allowing clones, Apple was doing most of the hardware R+D, and they would have had to charge $300 for MacOS to even break even on the lost Mac sales.
Of course, Apple doesn't make anything which even remotely resembles server class hardware, so OS-X Server/Intel would be nice option that wouldn't hurt their sales much.
--
Business. Numbers. Money. People. Computer World.
Apple, like Sun, IBM, Oracle, SGI, and Microsoft, are the enemy. They pollute the world with more non-freed software aimed at the horizontal market. If Apple is really committed to GNU/Linux, they will release their Mac OS GUI under a freed software licence such as the GPL. So far, they've only displayed a willingness to release code that was already freed (e.g. Mach) under a more-restrictive licence.
Just because Apple is anti-Microsoft does not mean they are our friends. Yes, they have produced nice hardware, but so has Microsoft--I like the Microsoft IntelliMouse. That doesn't mean I embrace Microsoft's vision of one operating system for the entire world.
Apple is more proprietary than Microsoft--Microsoft has usually shown a willingness towards open hardware designes. Apple won't even come that far. As far as Apple is concerned, they should be the only hardware vendor. Standards-based specifications such as SCSI or USB help this situation somewhat, but I recall in recent history a great difficulty on the part of the LinuxPPC team in getting decent specs on the iMac.
Hopefully, anal freed-software fanatics such as I will be able to steer the DarwinLinux group in the right direction. I certainly wouldn't mind any improvements in FreeBSD, as I've found it a wonderful kernel (I just want Debian GNU/FreeBSD ;þ). I am not writing this to show my hate for Apple; I'm simply pointing out that they are yet another huge corporation, such as my beloved IBM, that tends not to have any respect for the lowly hacker such as you and I.
--jon. Postel is dead. May we all mourn his, and our, loss.
hey, I think you can be more assine about this than you are being....come on, express your opinion....dont hold back.
really, you just managed to make yourself look like a real prick...people make mistakes, deal...its a free site that makes no claims of accuracy
I've always refused to get into the whole "name debate" thing. But doesn't it seem strange that they would call it "Darwin Linux" when the actual Linux kernel is conspicuously missing?
Thanks for the info, BTW.
--- Tao
I remember those good days I had with my little Mac. Now wouldn't it be cool if I could just get that nice little interface on top of a free unix operating system AND run Escape Velocity. . hmmm posix compliant Macos gaming machine, drool. Too bad those Apple people are a bunch of pricks and expect me to buy a Power Mac and an expensive operating syetem. Oh well.
Though I would like to see how fast Apple would pull out there termination clause if someone tried to add Macos style functionality and compatibility to Darwin.
Once Darwin is further along, I wouldn't be surprised to see Apple simply recompile the remaining layers,(ie Quartz and Cocoa) and release OSX/x86 ala Rhapsody DR2/x86. I am unaware of any technical reason they couldn't do this. With this approach, they can provide a larger market for their developers without dooming the project by attempting to do too much initially.
(For those of you who don't know, Quartz is the replacement for the DisplayPostScript imaging model found in OSXServer and Openstep. Cocoa is the new name for the Obj.C and Java API's formerly known as YellowBox)
Openstep is Dead! Long Live Openstep!