Slashdot Mirror
Looking Back at MacOS on x86
nutt writes "The MacSpeedZone's Apple Confidential column has a good Article which looks back at what happened to Apple's Star Trek project, (which was to "boldly go where no Mac had gone before." ... Intel hardware.) Its a very good read, and makes one wonder where Star Trek is now? The Article says the NDA's on the engineers was lifted in late 1997. It would be _very_ interesting if something like this could get out to the OSS.
Note: Darwin currently compiles on Intel hardware."
Apple doesn't have to release a version that runs on *every* generic Intel box out there - they can make Apple-branded boxes that use Intel hardware.
One of the main reasons Mac OS is stable is that Apple controls the hardware environment. A PC box can have a near-infinite combination of random hardware of varying quality - half the time when my Windows box goes flakey it's the crappy sound card (or whatever) I got at a CompUSA free after rebate sale.
So, to compete with Windows in megahertz-hype marketing wars, Apple could bring out their own cool Intel or Athlon based machines. The best of both worlds, so to speak.
Soupwizard
---
Twelve Step "How to Mountain Bike" Program:
Step One: Falling and Hurting Yourself A Great Deal
Am I mistaken, or did Bill call Intel Chicken?
the boy named Griff asked "What's matter, McFly? Chicken?"
Actually, there is a special database (the Desktop DB). It's just updated automatically, so most people never need to know about it (especially since it doesn't tend to get corrupted like it used do in the earlier days of System 7). This is possible because of the underlying architectural decision to use unique application "creator" types and filetypes that aren't part of the filename, and the higher-level objects and tools that are built on those concepts (bundle resources etc).
It's also not true that there aren't absolute paths stored - there are. But they're a fallback, and they're not stuck with silly drive letters (or with changeable mount points, which is weakness of absolute symbolic links on Unix): Changes can be repaired automatically because file aliases and "alis" resources store multiple, redundant methods of finding a target, and the Alias Manager tries them successively when one fails. There are open file ID's (somewhat equivalent to inodes), folder IDs in the B*tree (equivalent to pretty much nothing on UFS or ext2fs), volume names, volume partition IDs and device types for when a volume has been renamed, and file names, type/creator info, and yes, paths for when the other methods fail. If something changes and a file can't be found the other methods are used to find it, and when that succeeds the alis resource is updated with the new, fresh information (so the path or whatever was used won't be needed next time and redundancy is restored).
All of this means that I can rename a file and the link doesn't break, because paths aren't the only method or even the first one used - but alternately I can trash a file or an entire application folder (say with an upgraded version of an app) and replace it with one of the same name *in the same path* and the link still doesn't break, even though the old file ID is gone (the path is used to find the new one). Only when all of these tricks are exhausted with no joy do you have to see the "delete/fix manually" dialog. It's a pretty good system, but it's not yet fully clear how some parts of it will work under Mac OS X and on UFS filesystems. NeXT-style app and folder bundles can provide a lot of it, and there was an interesting article by Fred Sanchez linked to from slashdot on the subject a while back.
The reason apple doesn't port their OS is because that would kill their main source of revenue: hardware sales. Technically there's no reason why we can't run Mac OS on a PC (just like why there's no technical reason why we can't run windows 2000 on a G4). So the reasons why we can't do that anyway are of political/economical nature.
I think it's a pitty mac os X will never run on intel. I'm unhappy with both linux (no decent UI) and windows (decent UI but unstable). Mac OS X seems like a winner in this area, stable, decent UI from the company that practically invented the concept of a UI, runs MS Office (killer app for any desktop environment), runs internet explorer (note I'm actually writing this in a mozilla nightly build) and runs unix apps and development tools. However, should apple ever port Mac Os X, there would be no technically sound reason to buy apple hardware anymore (at least not at the prices they currently sell it).
Jilles
That's easy to say that Apple lost "world domination" if they had only shipped the Mac OS on Intel. I'm sure Microsoft would have sat back and let them do it, too. If Apple had released something like that, Microsoft would have just pulled the plug on Microsoft Office for the Mac, and guaranteed that people would have stuck with them anyway. They've done that sort of thing before when they threatened to kill Office for the Mac if Apple didn't adopt Internet Explorer. Then Apple would have been left with having to compete their hardware platform with commodity hardware from Intel clones without Microsoft's tepid support.
I'll admit that putting the Mac OS on Intel would have gone a long way towards acceptance of the Mac because people wouldn't have to invest in hardware to try out the system. However, Apple would have suffered the same problem as they did with the later clones. The Macintosh is an integration between hardware and software, and running the software on generic hardware waters down the Mac quite a bit. With that, the Mac would lose a lot of its distinction, and I don't that would have helped Apple's business any.
Insert simplistic political, ideological, or personal proselytization here.
MacOS was clearly superior to Windows 3.x. Hell, MacOS 7 was superior to Windows 95/98 in just about every way imaginable. That it was never ported and sold as a product may be a success and notch in Microsoft's belt, but it's one of the best examples of how Microsoft's (mis)behavior hurt consumers, and in so doing damaged the American economy. When your clueless office-mate asks you "how did Microsoft's practices hurt me?" you can point to the MacOS on Intel that never was and say "here's something you might have wanted that Microsoft made damn certain you couldn't get."
And that's NOT a level playing field...
[Replying to my own postings .. pfeh!]
I'd almost forgotten about MAE - the Mac Application Environment. This was a MacOS 7.5.3 emulator for HP-UX, and was around in the early-mid '90s. Apple officially dropped it in 1998. OK - it's not x86, but it *did* run on PA-RISC
Alison
"It is a miracle that curiosity survives formal education." - Albert Einstein
As the article hints at, it would make no sense for Apple to release an Intel version of the MacOS, because all of the existing Mac software has been compiled for a PowerPC (or a Motorola 68K series). The Star Trek project mentioned in the article is source-compatible, not binary-compatible, so every software manufacturer would have to recompile all of their code for the x86, and some would have problems, and many just wouldn't bother. The biggest issue would not be with the major applications like Photoshop and Excel (I have no doubt Apple could talk Adobe and Microsoft into supporting any new type of MacOS they come out with) but the thousands of free, shareware, or small commercial applications that people wouldn't bother to recompile.
When Apple switched from the 68K to the PowerPC a few years ago, not only was the PowerPC many times faster and able to emulate the 68K in real time, but it was the only significant change to the hardware, so many programs which accessed low-level MacOS hardware on the 68K still worked on the PowerPC. Also, Intel/PC hardware is too varied. Remember how long it's taken Linux to get support for all of the different PC hardware out there?
If Apple had continued to develop this, improving stability and porting more software to the platform, then M$ may not be in the dominant position it is today. This would also mean that the whole personal computing world could be focused on cheap x86 technology, rather than being fragmented betwen two different architectures
We would also have a lack of real choice. The beauty of nice, standardized Mac hardware is that it is far more likely to just work then hardware in the x86 world. I like to have that option available to me rather than being stuck in a situation where you have to buy x86 if you want a computer.
Besides, look at all the standards that Apple pioneered and eventually brought to the x86 marketplace: 3.5" floppy drives, built-in ethernet, SCSI, affordable/practical wireless networking, software power control, FireWire. Heck, it even did a lot to popularize USB for Intel. It's amazing to me that even to this day that most x86 machines don't come with ethernet built in, yet almost all Apple machines (even laptops) have had that for years and years. Additionally, Apple's now shipping gigabit ethernet standard on the MP G4s.
No, I contend that it's a good thing we have the Mac hardware platform. It's brought real choice and innovation to the market.
- Scott
------
Scott Stevenson
Scott Stevenson
Tree House Ideas
Apple's past maps directly to their fututre, it seems. When they bought NeXT, they got a nice multi-platform system, not just at the OS layer, but at the application layer as well. Since then, the have reworked it into Mac OS X, which has become increasingly Mac hardware specific. Darwin might run on Intel, but that's like having the Linux kernel cross platform but no libraries or applications. Just like its NEXTSTEP predecessor, the first developer preview of Mac OS X had an Intel version, but Apple dropped it after that. Now they say they're no longer supporting application-level cross compilation to Windows (aka, Yellow Box). They're also dropping Objective-C, their most useful foundation technology in my opinion, for future versions of WebObjects.
In the early 90's I ditched Apple for Linux because I needed a base OS that actually worked well. In the late 90's I went back to Apple (but not for my server! :-) by way of NeXT with every hope that Apple would have the resources to take the NeXT technology in the right direction. Here I sit in the early 00's looking again at Linux and being pleased with how far GNUstep has come.
The nature of the application market requires cross platform support these days. Apple continues to snub their developers when they make these kinds of decisions. Unless they start making better decisions, they may well end up as the "Also Ran" that some people have been calling for the last 15 years. Sad but true.