Apple's Darwin Runs XFree4

Mneme writes "Looks like Apple finally has the open source core of MacOS X up and running on Intel boxen. We'll have to wait until the WWDC is over before we'll get our hands on it, but it's still a very pleasing development. Check out a story about the demonstration, or click below to read the message from the Darwin Developer's list.

Here's the message from the Darwin Development mailing list:

Date: Tue, 16 May 2000 01:11:51 -0700 (PDT)
From: Dave Zarzycki
To: darwin-development@public.lists.apple.com
Subject: FYI: Monday's Darwin demos at WWDC
Message-ID:

On Monday, Fred [Sanchez] and I had the pleasure of demonstrating two very interesting Darwin developments.

First, we demonstrated Darwin running on Intel. Now I know that many of you are interested in this development, and I can probably predict many of your questions (where do I get it, how do I install it, where do I get the source, etc). Please exercise patience as we sort out the details. As you might imagine, we're very busy with Apple's WWDC.

Second, we were able to demonstrate X11 (XFree86 4.0 to be exact) running on Darwin's native I/O architecture IOKit. I had a lot of fun doing this port, which was loosely based off of John Carmack's port of XFree86 3.3 to AppKit on MacOS X Server. What I don't think most people realize, is that it only took a few days to do the port. X11 isn't as hard as some of you might think! ;-)

Please exercise patience with the above, we're very busy with WWDC. We'll try and push the necessary information out as soon as I can.

davez

--
Dave Zarzycki
MacOS X & Darwin
Apple Computer, Inc.

Comments

How about the other way around?

[ceswiedler]

...can I run the Mac GUI instead of XFree on my Linux box?

Yeah, yeah, free software and all, long live KDE and Gnome, but Mac does GUIs well. If they sold their GUI alone as a windowmanager, I'd snap it up.

Re:What is the comparative performance?

[jht]

I can recall a story from long ago and far away that once upon a time apple had an earlier version of one of their OS recompiled and running on and intel platform.

they dropped the project because it ran a bit faster than their own official Apple hardware, and they didn't want to shoot themselves in the hardware department.

There was a project to port MacOS to Intel, called "Star Trek" internally if I recall. And from what I remember of it, they had amazing initial success getting the OS up and working, and most of the functionality implemented. "Most" is key here. It was pretty fast (though PowerPC itself had a big performance lead at the time, Apple's supporting architecture was dog-slow), but there were still a good amount of key features not implemented yet at the demo point. Apple's good at getting an OS project most of the way there - it's the last 25% or so that kills them.

Ultimately, Apple decided to stick with PowerPC, and they have since based their hardware on faster stuff that's comparable to the state of the art in PC hardware (100 MHz bus, AGP, ATA-66, etc). So an Intel port for the "classic" MacOS (which is what Star Trek was) wouldn't be relevant and a waste of time and resources. That said, OS X on Intel would be a different story, and if Apple ultimately supported OS X native and Carbonized MacOS (through emulation - a recompile would be a killer) apps on an OS X Intel port, it would probably be a Good Thing. But they need to concentrate on their own platform before they give serious thought to a port. The fact that Darwin (the core of OS X) runs on Intel helps show that it's not too far from their minds.

- -Josh Turiel

Apple Reality Check

[maggard]

OK, lots of Slashdotters are really out of touch with the Apple & the Mac market. Here's a few points to keep folks on the reality track:

• Apple is a hardware company. MacOS is simply another way of selling their boxes.
• Apple is not interested in becoming a OS vendor. They have a closed platform to develop for and know their limitations: the wild-'n'wooly world of Intel x86 boxes is not a profitable place for them. They've stated repeatedly and has the numbers to back it up: they're not going to become an OS vendor. They're more likely to start selling hairdryers (or the fabled iBrator) then shrink-wrapped MacOS-X-for-Intel-x86 at Wal-Mart.
• The chance of Apple opening up Quartz, Carbon, Cocoa, QuickTime, or Aqua to Open Source or GPL is something approaching nil. These are the technologies Apple has paid a pretty penny to develop and their customers pay a premium to use. If you were an Apple stock-holder would you want them to do so?
• Apple never had "clones" (Apple's proprietary "Toolbox" embedded in a ROM stopped that.) They did have a "licensing program" with which they attempted to reach markets that Apple itself couldn't. They killed the program when the licensees started to savage Apple's core markets. Apple was loosing money on both the licenses and on the lost sales - not a smart move for a company and one they eventually stopped. Oh - and Power Computing? it was about to collapse anyway as it's financials proved out.
• Apple is again over 5% of the market and growing quickly. That doesn't mean they're about to challenge MS's OS dominance but they're going strong and expanding their market. That's not bad for a company that was written off a few years ago.
• Darwin is not the same as MacOS X - it's simply the Next-derived Mach kernel and support services. Yes, it's the base layer, but it's not the goodies. Getting it to run in Intel x86 is no great feat - NextStep was ported to 5 CPU's already including Intel x86 (PowerPC, Sparc, Alpha, MIPS.) All Apple's done is maintained that Intel x86 portability for the kernel. The big question would be if Apple could do the same for the other layers such as the Quartz rendering layer or the Aqua interface. Without those you've just got a very nice, very mature formerly-commercial Mach implementation - hot but not MacOS X.
• MacOS X has been shipping for well over a year. MacOS XServer is available at your local Apple Vendor or direct from Apple. What everyone is talking about is the MacOS X Workstation implementation. While everyone's been staring at that hand though Apple's been busy learning and tweaking with the Server version so when it finally ships the rest they'll have it nailed.
• Porting to Darwin won't be any more difficult then porting to any other BSD. With X now in place it'll be that much easier to use an interface. The big question will be how easy it'll be to make those same apps work through the Quartz/Aqua layers and become "native". If it's as straightforward as it now appears then Apple may have a tremendous position sitting at the crossroads of the market.
• Don't ignore the contributions Apple can make to the market as a whole. They've already pioneered the use of XML for putting a uniform graphical interface on the notoriously idiosyncratic *nix config files.^1 They also provide a great transition market for ISV's heading towards the *nix market but wanting to go through a more traditional platform. This could well be the *nix for "the rest of them" (the non-geek population.)
• As to behind schedule - well the beta is indeed going to be a few months behind schedule. How does this affect the rest of the timeline? Well, Apple hasn't changed their "will ship pre-installed in December 2001" commitment. Furthermore with Apple's Unified Motherboard Architecture finally in place (gone are the days of 50 models a year all with unique slots and ROMs and bus variations!) this looks like a legitimate possibility.
• Finally, remember Apple has been a massive developer of cool ahead-of-their-time technologies. Apparently Jobs has been going through the warehouse and pulling some of these off of the shelf now that they've a strong OS and a massively powerful CPU to ride upon. Expect to see some surprises coming out from Apple - Jobs is ever the showman and no slouch when it comes to hot tech.

-- Michael

1. Yes, some invariably CS Sophomore will pop up saying *they* thought of whatever, in this case XML-based forms for providing a uniform interface to the various config files years ago and mebbe they even have a few lines of code somewhere - well it didn't really happen then and Apple has now made it so. That alone is "A Good Thing".

Make use of legacy hardware?

[Croaker]

How about this... say Apple is making a play for the server market (go ahead, say it!). Let's suppose an IS administrator is sold on the Apple OS X server platform but... he has a server room full of x86 legacy hardware. Ripping out all of those machines and replacing them with new Mac hardware, while great for Apple, would put the cost of moving over to the Mac platform out of reach.

Along comes Darwin, which for many server-level tasks looks like OS X. The Apple folks can sell the IS folks on upgrading their exisitng x86 hardware to Darwin, making it interoperate with the new OS X servers, and down the line the customer will very likely replace the x86 hardware with more capable and more compatible Macs.
</wildassguess>