Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
But don't expect me to believe that nobody even at Apple knew or at least logically deduced that they had continued to maintain the Intel port that had already existed for seven years and been sent to developers in pre-release versions of OS X. I also get annoyed that the story is presented as if they started from nowhere and accomplished the impossible (or at least improbable) because again the Intel port was already there.

Unsure what the context is here regarding the "Intel port".

The earlier Star Trek project was to get Classic OS running on Intel.

The underpinnings of OS X come from the NeXT acquisition and are completely unrelated to Star Trek. Later versions of NeXTstep did run on Intel.
 
Ah

IBM can blame themselves for losing a big partner like Apple on this. They failed to deliver a G5 (or any processor faster than Motorolla's G4) for the laptop series. Apple laptops were stuck with aged and underpowered G4 for a long time waiting for IBM to deliver something new. Apple had to move on eventually, or lose the laptop market entirely.

I'm not glad to say this, though. PowerPC processors seemed to have a better architecture, leading to a great performance on lower clock speeds than Intel's cpus.

In retrospect I loved my PowerMac G5 2.5 dual. At the time its performance was without equal in certain areas.

But it was a ridiculous machine when it comes right down to it: under heavy load it could double as space heater - the temperature monitoring app once recorded a high temperature of 212 degrees on the backside of the memory controller. We'd actually close the doors to our office and warm up with it in the winter.

It used enormous amounts of electricity too. And ultimately the liquid cooling system was a boondoggle as that failed more than once.

The Intel change was a smart decision. The i7 chip in my current Mac has more cores, is faster, has more cache, uses less electricity and simply runs the operating system better than the G5 from five years previous.

Plus the fact that we can run windows if we want is a plus not a negative.

At the time though it seemed like the end of the world. I felt like a girl had dumped me or something.
 
I tried a MacBook for two days, running Panther. It just wasn't ready for me and I wasn't ready for Apple. That would have been about the time all of this change was going on inside Apple to change over to Intel.

I remember Steve's announcement that Apple would move to Intel platforms. The buzz said it would take three years and that it would be a bumpy road. Then, a few weeks later, the Intel Macs were available. It was a master stroke, executed flawlessly. By 2006 I was using an Intel MacBook Pro and never looked back.

Thanks to all who made this huge step happen. :D
 
I wonder if there is a team or person doing the same for ARM processors??

The vast majority of that work is already done -- it's just that they don't call it OS X when it runs on an ARM. They give it a different skin and call it iOS.
 
I tried a MacBook for two days, running Panther. It just wasn't ready for me and I wasn't ready for Apple. That would have been about the time all of this change was going on inside Apple to change over to Intel.

I remember Steve's announcement that Apple would move to Intel platforms. The buzz said it would take three years and that it would be a bumpy road. Then, a few weeks later, the Intel Macs were available. It was a master stroke, executed flawlessly. By 2006 I was using an Intel MacBook Pro and never looked back.

Thanks to all who made this huge step happen. :D

Panther didn’t run on MacBooks.
 
But Intel was also working on a completely different chip based on the P3, and that eventually became Centrino and wiped out AMD. AMD had nothing - all they knew how to do is compete with the P4.

Ok, you're the second person who has referred to Centrino as a 'chip' or CPU 'architecture'. I suppose that just goes to show how confusing, pointless, and prominent Intel's Centrino branding was, because it certainly wasn't a chip or CPU architecture. Centrino was a package of CPU, North/South bridge, and wireless network chipset. The Centrino branding has it's roots back in the day of the PII/Celeron, but it is now relegated to just Intel's Wi-Fi and WiMax chipsets.

The PII and PIII CPUs are, architecturally, descendents of the old Pentium Pro CPU. You can find their descendents in the Core/Core Duo/i3/i5/i7 lines. The Pentium IV was a dead-end based on the idea that longer pipelines enabled faster clock speeds. The PIV was, simply put, a dead-end requiring far too much power to actually achieve the goals they had laid out for it. They had to ditch the line and move back to the older PII/PIII architecture to drag themselves out of the holes they dug with the PIV and Itanium. (And, yes, that pair of mistakes are what gave AMD their heyday, but AMD is still competitive at the lower end of the x86-64 market, and are the ones that brought us 64-bit x86-compatbile CPUs. Those weren't *ever* on Intel's road map until well after AMD produced their 64-bit Athlons.)
 
But don't expect me to believe that nobody even at Apple knew or at least logically deduced that they had continued to maintain the Intel port that had already existed for seven years and been sent to developers in pre-release versions of OS X. I also get annoyed that the story is presented as if they started from nowhere and accomplished the impossible (or at least improbable) because again the Intel port was already there.

As I said, I understand why they were publicly silent. They were probably also super-careful about who had access to the codebase and machines running the port, because one installable image of the Intel port in the wild and everyone would be trying to install it on their home PC.
I may have not made this very clear in the earlier post, so apologies for any confusion. There was never a version of OS X for Intel sent to developers prior to the 10.4 version sent on the P4 dev machines in 2005. There was versions of Rhapsody, and NeXTSTEP/OpenSTEP for Intel. But when Apple ended Rhapsody client plans and started work on what became the OS X we actually run, all developer builds accessible to people outside Apple were PPC only. OS X removed some major chunks of Rhapsody, and added many new components. The only Intel component visible to the public that remained was Darwin. And that is only one of many pieces you need to make a full OS X system.

With how Apple runs and their system of secrecy, I can believe they kept this pretty close to the chest, and may have had it on ice for a while. It's hard to really put into perspective how dire things were in 98 for Apple these days with their current success. When the simple stick was being swung around hard to get back to a sustainable model, it's easy for me to believe that impacted the Intel work internally for a period of two years. That long of a time gap when an OS is under construction is huge, and would take some time to port to another platform if only one CPU family was being targeted. To put into perspective 2 years, the actual code that shipped as Windows Vista was only in development for 1.5 years, started from the Server 2003 SP1 codebase.
 
Friend still has a prototype Intel machine that was blue binned (send to the crusher) when Jobs returned to Apple. Multiple internal interfaces including SCSI, IDE, etc inside a 8500 case.
 
It was such a please to stop explaining the megahertz myth over and over and over and ...

Yes I remember them doing that presentation over and over. The fact is they should have moved to x86 years before. That was the performance standard and by staying on PowerPC as long as they did only hurt sales compared to Intel PCs which were clocking so much faster. The other thing is updates came so slow with PowerPC. Intel has new chips available all the time with real speed improvements.
 
But this story and the related user posts raises another rhetorical question. Why didn't the copy machines at Redmond take the open source stuff and morph it into a Apple compatible run time utility on PC's?
I would venture to guess that's because most of the really important stuff required to actually differentiate a real Mac-compatible run-time from something generic such as X11, is all hidden away inside APIs that are not open source.

Also, what incentive would Microsoft have had to expend the effort to do so? Most of the major software vendors were already using the Windows API as their tier-1 platform, so adding such a compatibility layer likely wouldn't have resulted in a significantly more diverse software ecosystem for their users, and therefore likely wouldn't have resulted in significantly more sales for Microsoft.
 
Last edited:
This story seems to indicate that OS X was "suddenly" developed to be able to run on generic Intel based PCs.

Like others have said Rhapsody ran on Intel as did OPENSTEP.

And, let's not forget that "Darwin" has always been an open source project, freely downloadable from Apple's website, which ran on most generic PCs at the time, too.

And let's not forget Project Star Trek. During the early 1990s Apple (partnered with Novell) was actively developing MacOS for Intel. The thinking at the time was it was well underway, perhaps even all but completed. The project was terminated for reasons that are not totally clear, but at least some part of the cause is buried in U.S. v. Microsoft. In that timeframe Apple could not find hardware partners at the big OEMs willing to buck Microsoft. It was common knowledge that Microsoft would punish any OEM that didn't play according to their rules. So while x86 Darwin and Marklar were obviously big deals, and changed the face of computing, Star Trek came long before both of them and could have had an even larger impact.
 
The precursor to MacOS 10 was the NeXTOS which had been successfully ported to Intel-based PCs IN THE LATE 90's and called the MachOS. So the experience of porting the eventual MacOS 10 to Intel-based PCs is a lot deeper than many realize. Dual-boot PCs with MachOS and Windows were commercially available and often found in university computer labs. An early form of BootCamp allowed one to choose the OS. When Apple purchased NeXT and Steve Jobs returned to Apple, he brought all of that EXTENSIVE Intel-porting experience over with him. The jump to Intel was therefore a no-brainer -- and a snap -- once IBM started dragging its feet in extending the PowerPC platform.

I think that Mach came first and that NextStep and OS X are built on a version of Mach. Mach was a development for a kernel to replace the kernel in BSD Unix. Apple bought XNU and worked with that version of Mach to create OS X according to an article...

IBM did eventually come up with a version of their processor fit for notebook use, but it was too late. I remember people hypothesizing that Apple would then transition back to PowerPC chips because they hated being tied to a lethargic Intel who also had a number of issues with quality control at the time. Well, and people thinking that Apple would use both processors to somehow hedge the market or something, which made less than zero sense...

----------

I agree. It is very interesting Jobs immediately went to see Sony.

I'm amazed that Sony Viao's still have some kind of 'reality distortion field' surrounding them. They really aren't that great. Usually slower speeds, and higher prices than comparable systems from other vendors... Maybe that's why Steve thought Sony? Of course back then Sony probably still had their image of also being a quality producing company...
 
Last edited:
I'm amazed that Sony Viao's still have some kind of 'reality distortion field' surrounding them. They really aren't that great. Usually slower speeds, and higher prices than comparable systems from other vendors... Maybe that's why Steve thought Sony? Of course back then Sony probably still had their image of also being a quality producing company...

Sony had only just launched the VAIO range around that time. The quality and support for the first three years was vastly superior to anything since. I worked in a PC store at the time and the buzz around the product was second only to the iMac/iBook. They were *the* Windows notebook to own at the time.
 
I may have not made this very clear in the earlier post, so apologies for any confusion. There was never a version of OS X for Intel sent to developers prior to the 10.4 version sent on the P4 dev machines in 2005. There was versions of Rhapsody, and NeXTSTEP/OpenSTEP for Intel. But when Apple ended Rhapsody client plans and started work on what became the OS X we actually run, all developer builds accessible to people outside Apple were PPC only. OS X removed some major chunks of Rhapsody, and added many new components. The only Intel component visible to the public that remained was Darwin. And that is only one of many pieces you need to make a full OS X system.

With how Apple runs and their system of secrecy, I can believe they kept this pretty close to the chest, and may have had it on ice for a while. It's hard to really put into perspective how dire things were in 98 for Apple these days with their current success. When the simple stick was being swung around hard to get back to a sustainable model, it's easy for me to believe that impacted the Intel work internally for a period of two years. That long of a time gap when an OS is under construction is huge, and would take some time to port to another platform if only one CPU family was being targeted. To put into perspective 2 years, the actual code that shipped as Windows Vista was only in development for 1.5 years, started from the Server 2003 SP1 codebase.

I think we just have a difference of opinion.

I consider Rhapsody and today's Mac OS X to be the same operating system. I know there were some major evolutionary differences between what was shipped as Mac OS X Server 1.0 and what became Mac OS X 10.0 for consumer release, but I see it all as part of a continuum, the same codebase that they were adding to over time. The kernel was already there. They transitioned from Display Postscript to Aqua, which is still based on PDF which is very similar to PostScript. The developer tools and APIs were basically carried over from OPENSTEP, which was rebranded as Cocoa but it was largely the same. It was still Project Builder and Interface Builder as the IDE until 10.3, which went all the way back to NEXTSTEP 0.8!

Indeed, a *lot* of stuff was added between Rhapsody DR2 Intel and the point at which this guy began working full time on the Intel version of the then-current codebase, and there may have been a lot of loose ends. I'm sure getting Carbon running took some effort, as that had surely grown close to Apple's legacy hardware. But this was still a case of perhaps an 80%+ complete port existing at all times, not something that was conjured out of whole cloth as is frequently implied.

So I don't necessarily disagree with your take entirely, but I personally (from a developer's perspective FWIW) tend to give a lot of weight to how much stuff was already there (and publicly known to be there) and see this effort as maintenance rather than a Manhattan Project style effort to create an Intel version from nothing.
 
Last edited:
Unsure what the context is here regarding the "Intel port".

The earlier Star Trek project was to get Classic OS running on Intel.

The underpinnings of OS X come from the NeXT acquisition and are completely unrelated to Star Trek. Later versions of NeXTstep did run on Intel.

The context is clear from this entire thread being about the Intel port of Mac OS X. The numerous mentions of NeXT technology in my posts are also a clue that I'm talking about the technology acquired from the NeXT merger, that became OS X. When I talk about the existing Intel port that had been there for seven years before this guy started working on it in 2000, I am referring to the fact that NEXTSTEP was released for Intel in 1993 (I think, it may have been a little earlier). Since NEXSTEP became OPENSTEP became Rhapsody became OS X Server became OS X and iOS, I conclude that in a general sense "OS X" had been running on Intel hardware since 1993.
 
that's what I was thinking.

arn

We have to remember that Jobs was still trying hard then to restore Apple's financial stability. May be, he thought licensing it to top tier companies would do that.

I loved the article alright, especially the heartwarming aspect of how they wanted their son to be close to his grandparent which started this whole thing. There is a bit of the unavoidable romanticizing that goes with such narrations. Nothing bad. We have to view the sequence of events in that context. One can easily get the following impression:

JK asks for permission to work on Intel. Granted
1.5 years passes, no one knows or cares about what JK does.
Including his boss, because he asks 'I need to justify your salary, tell me what you are working on'
JK shows OS X running on a intel PC
Everyone is impressed.
Jobs is told that night
Jobs leaves for Japan the next morning

Probably not exactly how it happened. May be his boss new what was going on but did not share, or Jobs knew what was going on or he went to Japan after a few days ;)

I actually heard he went to Hawaii to meet with the Sony Chief.
 
Last edited:
I mean that this story indicates that development of OS X for Intel was "suddenly" being developed once it already run on PPC.
....

Bet you a dollar that there's a build of OS X running on an ARM-based MacBook Air right now. Hobbled because of the 32-bit data path. Patiently waiting to be released upon an unsuspecting world when 64-bit AX chips are available.

By the way, ARM released their 64-bit instruction set for the ARMv8 architecture last year.

----------

Hello? You do realize that ARM chips are currently way slower than even the slowest Intel chips? That's kind of the issue there.
[...]

Hello. I do realize that ARM chips are currently way slower than some Intel chips.
(But probably faster and more energy-efficient than Intel's legacy-friendly Atom chip.)

The key word is "currently." Things change. Fast.
Just ask Microsoft.
 
The context is clear from this entire thread being about the Intel port of Mac OS X. The numerous mentions of NeXT technology in my posts are also a clue that I'm talking about the technology acquired from the NeXT merger, that became OS X. When I talk about the existing Intel port that had been there for seven years before this guy started working on it in 2000, I am referring to the fact that NEXTSTEP was released for Intel in 1993 (I think, it may have been a little earlier). Since NEXSTEP became OPENSTEP became Rhapsody became OS X Server became OS X and iOS, I conclude that in a general sense "OS X" had been running on Intel hardware since 1993.

No it wasn't, but thanks for your input. :rolleyes:
 
Last edited:
Bet you a dollar that there's a build of OS X running on an ARM-based MacBook Air right now. Hobbled because of the 32-bit data path. Patiently waiting to be released upon an unsuspecting world when 64-bit AX chips are available.

OS X has been shipping on ARM devices for 5 years. Only the uppermost veneer of iOS is different from OS X. It's the same operating system with two GUI toolkits.
 
OS X has been shipping on ARM devices for 5 years. Only the uppermost veneer of iOS is different from OS X. It's the same operating system with two GUI toolkits.

Agree that the basic OS X kernel (derived from the Mach kernel developed at Carnegie-Mellon) is essentially the same as that of iOS. But all the "fat" has been trimmed away from the higher layers. Especially, as you point out, the GUI layer.

Looking forward to 64-bit multi-core ARM-based AX SoCs on iPhone and iPad. I think the 64-bit data path is the gating factor in using said chips on Macs.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.