OSX coming to X86?!

[jaguar]

Get a load of this !!!
Sounds interesting indeed.

[Undertoad]

Project Marklar? Heh.

Even if they never plan to release it on the world, it would make great sense for them to keep slowly maintaining it, and then pepper announcements to the press.

Why: it's an on MS tactic. Announce that you have vaporware and force your competitors to compete with it before it's actually a product. If you force your competition into enough of a tizzy, you'll win without having to develop or market anything.

[MaggieL]

Originally posted by Undertoad

Why: it's an on MS tactic. Announce that you have vaporware and force your competitors to compete with it before it's actually a product.

Which of Apple's competitors would care about an X86-ported OS X? Microsoft?

I'm guessing the code isn't vaporware at all; it's probably not that big a deal to maintain the port. What strikes me as vaporous is how it could be productized without creating chaos in their buisiness model, developer community and channels.

[Slight]

I hate to say it but this is old news. It has been widely know that Rhapsody & then OS X has been running on x86 since it's inception (not that I 100% believe this story). Darwin, the open source, base OS of OSX has openly been ported to x86 (But why do we need another *BSD on x86?) I doubt apple intentionally leaked this as Steve is a secret freak. Its just that the idea is so sensational to the mac public that once someone finds out everyone knows about it. You can also see evidence of it in cruft left around different releases of OSX. Speculation that it is being used as a vaporware scare tactic is weak at best. It is more likely that it is being maintained specifically for a rainy day. That day is when Motorolla bites it. Currently IBM does not build the PPC chips that go in macintoshes because apple likes Moto's AltiVec vector unit on the G4 chip.

One look at Apple's yearly earnings report will tell you that Apple currently makes most of its revenue from hardware sales. While it is not out of the question that Apple could release a general purpose port of OS X to supplant Windoze, the driver support of all the millions of varying harware setups would instantly kill Apple. They could do 1 of 3 things: Support only their x86 hardware or certain hardware of their choosing. 2 they could release it and die of a Be type death or, 3 They could rely upon the open source community much as linux has to write drivers. I think apple would go with option 1, their hardware because Steve hated the clones. But I find option 3 the most intriguing.

It would be fairly easy for apple to get developers to re-release software recompiled for x86, because I am pretty sure gcc and CodeWurks can cross compile. The hard part would be recompiling drivers, code written in asm and code optimized for the G4 vector unit.

I see this as a long term decision, developed as an out. It would require mucho time to both redesign current hardware and convince the faithful that the /switch/ is good. Btw, I think project marklar used to be called project enterprise

[mbpark]

Apple's had this for years.

Darwin for x86 has been downloadable for a while (www.apple.com/darwin).

It's not that hard to port Aqua over to x86, apparently. From what I've heard, Nvidia's provided the video drivers. This means, if Apple did this, it would be done on THEIR x86 hardware. There's no way you're going to run this on your generic x86 box.

However, the best choice, due to the extra registers, would be the AMD "Hammer" architecture. They've apparently done ports to the P3, P4, and Athlon (someone did a Strings dump of their DVD player and found directives for compiling for those three architectures), and the Hammer has the extra regs they could use to make something like Altivec work somewhat half-assed.

Remember back in the 68K days, when they made products like Emplant for the Amiga, which actually used Mac ROM chips and a lot of custom hardware to make Mac OS run on an Amiga 2000, 3000, or 4000? It actually was faster than the 68K Macs, due to the fact that Amigas had a better architecture. However, it was a pretty large add-in card with a lot of extra logic, plus custom software to run the Mac in a "sandbox" on Workbench. It had to emulate darn near everything to make MacOS boot, and by placing most of that code in ROM, it worked really well. It's still a testament to good code that the Emplant developers got it to work as fast as they did.

People STILL use them to run Netscape on their Amigas, since that platform has everything but a decent web browser (amIRC, YAM, ixemul [UNIX emulator library like Cygwin], NewsRog, and a ton of other utils).

The only way you'll get this to run on an X86 is via something like an Emplant for PC. Since PC hardware is that much more complicated these days, I foresee it being pretty messy.

If Apple has it, it's running on an Nvidia chipset, more than likely with an AMD Athlon MP, with a lot of extra information there, and a specialized BIOS and bootloader. In other words, you'll be able to get it to run on a PC, but it will take serious work as they've probably put a lot of chipset and video-card specific code in there.

They did do it with MacOS 9 on LinuxPPC, and SheepShaver, but it wasn't easy.

Mitch

[tw]

Before Apple had committed to the PowerPC, they had already successfully ported their OS to the Intel chipset. Intel was fully involved in providing direct support for the effort - I believe with the direct and personal cooperation of Andy Grove. But these were the days of Michael Spindler. Innovation would be routinely quashed. Apple's OS was fully functional on Intel hardware and still Apple could not commit to marketing the product.

[mbpark]

Tom,

They had finder working, and a few apps.

However, it was because of the fact that a lot of Apple software used ASM shortcuts all over the place for performance (most old Apples did not have fast processors, either), that they canned it.

Since most of the apps would have required massive porting efforts, it was decided to drop it.

Today, most apps except games are written mainly in C/C++, and the port time is much quicker. That, and Apple's centralized their app development around the most common compiler out there next to Microsoft Visual Studio, GCC. They've also used a lot of Open Source/GPL/BSD software that's already undergone the X86 to PPC conversion, such as Apache, mySQL, Postgres, and the GNU tools. Hence Darwin, which is OS X without the GUI, and already fully operational on X86. All that's needed is the Aqua conversion, and they've got an environment where the port time can be measured in days, not months/years like with Star Trek.

The only thing that killed Star Trek was the fact that most of the Apple apps were using ASM. Therefore, the developers would have freaked and left the platform. Now, it's a different story, since they're using a common available codebase and development suite.

Mitch

[juju]

Wow, it appears that open source has given the company a lot of options. Even if they choose not to go with x86, it's great for a company to have adaptability. Especially in a market as competitive as this one.

[jaguar]

mbpark - the article talks about all that.

and most of the possabilties. I doubt it'll ever be released. If it is it'll be apple only hardware and i've got a feeling it'll be as inaccessable as it currantly is. I'd say its more likely an issue of who is going ot be making hte processors, i've heards mumours that motorola want out.

[mbpark]

Jaguar,

Doesn't mention that this has happened before with 68K boxes and the Amiga, and that the company that made Emplant got around it by offering a card that mimicked the Mobo .

I'm sure Apple knows, which is why IBM may be providing their next CPU, a derivative of a FAST cpu known as the Power4. Look at www.spec.org's submitted 2001 results. Compare SPEC results from the Power4 with the Pentium 4 or Athlon XP.

Now imagine that chip at 2Ghz.

I think we've got the "G5" right there .

Mitch

[jaguar]

True, true (amiga)
And yes, G5.....Does look nice, guess we'll see.

[mbpark]

Jag,

Now you understand why Apple would not want to go x86. Who is to say some Chinese or Southeast Asian company, which does not honor US copyright law, would not build "unauthorized" clones like Emplant? Emplant barely skirted the law by making you buy Mac ROMs. Seeing as how the Mac OS rom now is contained on the OS installation media, there's nothing stopping you from running MacOS on another machine without a card.

In related news, people do have MacOnLinux running on LinuxPPC on the "new" Amiga boxes, the Pegasos from bPlan, and apparently the AmigaOne. These machines are G3's, and I think someone has a G4 in the Pegasos, and they are booting OS 9.2 now. It will not be long before they boot 10, and have MacOS running on a clone platform again, albeit one that just doesn't have the Mac experience.

There are clones again. I'm waiting for Apple to smack down Eyetech and bPlan for making them if they advertise MacOnLinux. Knowing the people behind these companies, and the lunacy defined as advocacy by the Amiga "community", I wouldn't be surprised if one of the companies was that boneheaded to do so and got smacked by Apple for it.

Looking at this, moving MacOS X to x86 would require a wholly different hardware architecture and chipset, and many other things that enterprising crackers would break. Apple's got the "clone problem" again because of enterprising Amiga users who want to run MacOS on their "flash" new Amiga HW.

Even then, who's to say someone's not going to set up an Emplant for X86.

Moving to Power4 is a good decision for them, if they've made it, for one important reason. They'll be the lowest-priced Power4 machines, and emulating them will reduce the user experience to the point where it's not worth it anyway. That, and no one's going to spend $8,000 or up on a similarly configured RS/6000 to run MacOS just to say they're not running it on an $2,500 Apple.

If they put it on x86, there will more than likely be a "sandbox" mod for Linux within 2 weeks to enable MacOnLinuxIntel, and the "Apple Experience" that Steve Jobs is so big on will be diminished because every script kiddie and "l33t hax0r" will be running MacOS on their White Box and Dell PC's. This would have the net effect of knocking down their main way of making money, which is packaging HW and SW together in one tested and reliable system package, and charging a lot of cash for it.

In other words, it would kill Apple worse than the clones did because this time, it will open MacOS to every PC without a licensing fee. Everyone knows how Linux users are. Most develop it "because they can", and combined with the outright encouragement of piracy to "share information" that places like Slashdot profess daily, the P2P networks where people trade XP corporate serial numbers and OS ISOs for every OS out there, the outright misuse of good intentions by opportunists for piracy gains (such as the pirate Dreamcast CD's running DCLinux, XMAME, and 1,500 arcade games), and you've got yourself a morass that would kill a smaller HW/SW company like Apple.

I hope that Apple's people aren't that stupid.

Mitch

[tw]

Originally posted by mbpark
The only thing that killed Star Trek was the fact that most of the Apple apps were using ASM. Therefore, the developers would have freaked and left the platform. Now, it's a different story, since they're using a common available codebase and development suite.

Developer freaked anyway. The same reasons cited for not porting to x86 architecture are the same problems that Apple encountered when going to PowerPC. In fact, Apple gave developers almost nothing to make the transistion (a minicomputer based program that took days to execute) - mostly because no one in Apple apparently even thought about the transistion problem. It was a Toronto based company that made the transistion possible - discovered by Apple in a trade show, I believe.

No real reasons are provided for why Apple did or did not do many things in those days because most decisions made in the days of Spindler were quite muddled. Maybe it was corporate fear of Intel. Maybe it was fear that the Mac OS would be pirated. Maybe it was fear of clones. Maybe the OS was really only a reason to buy the hardware. Maybe - and this is one of the most likely reasons - it was fear of change.

At any rate, Mac OSes could have executed on Intel hardware and did so in a small team hidden from Apple's own people - in an office so secret that it did not even have signs. The x86 transistion would have been no more painful than shifting to PowerPC - another completely difference architecture from a failing 68x00 series. One common fact appears to be consistent in Apple decisions of those days - no one in top management could make a decision.

One of those Apple executives is a fraternty brother - a very calm and deliberate individual. Also, BTW, an MBA. However his futile attempt to get Apple to innovate (that right - the MBA was more innovative than the VP of engineering) actually resulted in a violent action. He threw his clipboard (?) down and stomped out of the room. Knowing him personally, that is extreme behavior especially for him. But Apple was that anti-innovative then. My friend left to become President of some other Silicon Vally companies. At one point, he was suppose to be selling Apple clones until Jobs foolishly terminated that effort.

[russotto]

Yeah, Apple is anti-innovative all right. Apple farts and Dell moves to copy the smell, and it's Apple who is anti-innovative.

Being an actual Apple developer (since late 1983), I know that Apple did in fact provide developers with tools for making the transition to PowerPC. Most obviously, the 68K emulator included in the system; some 68K applications STILL execute under 10.2 (in Classic mode).

Certainly it would have been possible at that time to move the the x86 architecture instead of PowerPC. But it wouldn't have made sense to do so; the x86 architecture was at the time and remains a creaky mess -- AMD and Intel have managaed to make the pig fly, but it's still a pig.