This wouldn't be a patch. The entire operating system and all the included applications would have to have been rebuilt and recompiled, with things like Rosetta stripped out.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
WHAT IF: Snow Leopard Were Universal?
- Thread starter MysticCow
- Start date
- Sort by reaction score
To be honest guys, the intel transition was inevitable. And they had been working on it since 10.2 I believe, from what Steve Jobs said in a video.
If PPC had it's **** together and had built a G5 processor that wasn't too power hungry and heat intensive they would have carried on the product line. The G4 line still would have had support until 10.7, then it would have died of
Intel would have still been used in the end, but it would have been a lot later and we would have still have had relatively up to date Macs.
Either way - I'm still using my PowerPC mac.
If PPC had it's **** together and had built a G5 processor that wasn't too power hungry and heat intensive they would have carried on the product line. The G4 line still would have had support until 10.7, then it would have died of
Intel would have still been used in the end, but it would have been a lot later and we would have still have had relatively up to date Macs.
Either way - I'm still using my PowerPC mac.
They didn't plan a transition until the G5 started becoming a huge disappointment, they had a team working on making OS X run on Intel, if it became necessary which it eventually did. However, Apple did toy with the idea of running it on both PowerPC and Intel. They once presented this idea to Sony, showing them OS X 10.2 (I think) running on a Vaio laptop. Sony however declined the idea, sticking with Windows. So while a move-over to Intel wasn't planned until the G5 era, Intel machines running OS X were always a possibility.
Hmm, not sure, basically Snow Leopard added a few features, I suspect the code could have been patched/ported into Leopard, we're looking at...
64-bit application support
Grand Central Dispatch for improved multi core processor support.
Reduced OS foot print (already discussed here)
Open GL graphics acceleration
Quicktime X
Expose Improvements
Dock scrollable stacks
Other minor finder tweaks
There were production problems with the G5 and the power/heat issues were also too much. It needed liquid cooling and there was no way to get a G5 into a laptop. The PPC had the potential to be an awesome way forward from the 68K world -- and was, for a while -- but the combination of G5 heat/power/production woes and an increasing demand for interoperability with the Windows world doomed it.
Not necessarily. The single and dual G5s had air cooling, only the quads had liquid cooling, although there might were some liquid cooled higher end dual(core) G5s, but I'm not sure.
Exactly!
I simply do not believe it's impossible, and TBH I am genuinely surprised that this has never been done. The multi core support would help all the later dual core core single chip G5's, not sure how much of an improvement it would be for machines with 2 single core chips.
Also given that this has enhanced 64 bit support, that might be the cut off, 64 bit processors only, no G4 support.
I simply do not believe it's impossible, and TBH I am genuinely surprised that this has never been done. The multi core support would help all the later dual core core single chip G5's, not sure how much of an improvement it would be for machines with 2 single core chips.
Also given that this has enhanced 64 bit support, that might be the cut off, 64 bit processors only, no G4 support.
Last edited:
The dual 2.5 and 2.7 GHz G5s were liquid cooled.
[doublepost=1461555233][/doublepost]
Apple may have compiled a version of 10.6 that ran on PPC, but we'll never know. Nonetheless, what I said remains true: moving 10.6 features into PowerPC isn't as simple as patching some code. The entire thing would have needed to be recompiled. Also a significant reason 10.6's footprint got smaller is because the PowerPC code was gone.
[doublepost=1461555446][/doublepost]
The move to Intel for portables was such a huge gain in performance, it can't be overstated.
Last edited:
There are some seriously talented coders out there, do not under estimate them!
A significant part of that performance gain was SL, rather than the Intel processor.
How does a mac pro compare with a G5 with similar specs and both are running Leopard? Or Tiger?
Faster, yes, but how much faster in the lab and in the real world?
A significant part of that performance gain was SL, rather than the Intel processor.
How does a mac pro compare with a G5 with similar specs and both are running Leopard? Or Tiger?
Faster, yes, but how much faster in the lab and in the real world?
This. The goals for Snow Leopard were improved performance and greater efficiency. The addition of new features as with 10.2-10.5 were very much a secondary consideration. The memory footprint was much lower, and much of the legacy PPC code was purged.
I appreciate the PowerPC Macs for what they are. The suggestion that some talented home coders can put the PowerPC code back in to Snow Leopard is fanciful, bordering on delusional. Apologies for being blunt!
Regardless of how efficient any new code written for PowerPC is, there's no getting away from the fact the hardware is over 10 years old and performance is now dwarved by modern CPUs.
However, as shown with Coreplayer, there's potentially still plenty of untapped power lurking within the PPC architecture - I'm surprised no one has done a Kickstarter for an Altivec Optimised PPC web browser?
Let's not forget MorphOS - that's written specifically for PPC and works wonders on the old hardware.
For those who want to stick with PPC in the 'now' era, there's the Amiga X5000, currently running a dual core Freescale PPC but moving to quad soon and also the Talos Power8 desktop with talk of OSX virtualisation.
However, as shown with Coreplayer, there's potentially still plenty of untapped power lurking within the PPC architecture - I'm surprised no one has done a Kickstarter for an Altivec Optimised PPC web browser?
Let's not forget MorphOS - that's written specifically for PPC and works wonders on the old hardware.
For those who want to stick with PPC in the 'now' era, there's the Amiga X5000, currently running a dual core Freescale PPC but moving to quad soon and also the Talos Power8 desktop with talk of OSX virtualisation.
Last edited:
I think you're underestimating the talent out there, and not all coders are necessarily amateurs.
Are there people out there who have the skills to do it?
I have reason to believe there are.
Are they actually going to do it?
I have reason to believe they aren't*.
*I've asked about much smaller projects in the past, and the answer was a firm but polite 'No.'
Are there people out there who have the skills to do it?
I have reason to believe there are.
Are they actually going to do it?
I have reason to believe they aren't*.
*I've asked about much smaller projects in the past, and the answer was a firm but polite 'No.'
I'm not certain you understand the limitations of closed source development.
It doesn't matter how good a person is at coding if they don't have access to Snow Leopard's source code. They could be the best developer in the world, but without access to this code, nothing could ever happen. From what I've read, you seem to believe that someone who is talented at coding can just try and take features from the released version of Snow Leopard and sort of stitch them onto PPC Leopard, but it doesn't work like that. The reason that it has not been done is because it can't be done. Only Apple could do something like this, but they opted to not do so.
TenFourFox uses Altivec for some purposes, but I agree, an Altivec browser would be spectacular.
Also, the biggest reason WHY Apple jumped from PowerPC to Intel was mainly because it was a dead-end. The G5 quad was reasonably powerful in the day, but it required liquid cooling. For portables, this was especially bad. They couldn't exponentially release newer and newer laptops with the same CPU, or one that was under clocked to deal with the thermals, to the point where it was similar to the past generation. It didn't make fiscal sense.
Now, I believe there was still a good deal of deliberation. The untapped power mentioned above was probably one of the factors against. I personally believe the changes they applied in Snow Leopard (performance wise) could've been applied to PowerPC, but they didn't want to steal the sales from their newer (more faster, more better) Intel line of portables, desktops, and all in ones. That business model is increasingly more common. Why do you think the RAM is soldered in now? Sure, you are saving space, but you now either have to buy a new Mac or swap the logic board (of one with higher memory) to get more RAM.
As others have said, it's clear you don't understand how the development process works. Without access to Apple's source code, a "coder" would have to develop 10.6 for PowerPC from scratch. They'd be writing an entirely new operating system. It's simply not worth any developer's time to write an operating system for a platform that's been dead for 10 years.
OK, so first up, I'm not a coder, and you're right I am not intimately familiar with the process of software development, or hacking into an operating system.
I do strongly suspect that some perhaps even all of SL's features could (theoretically) be patched. maybe even copied and pasted into a decompiled Leopard.
My reason is, I strongly suspect that the released version of SL is essentially a reduced version of an SL Apple developed for the planned but canned 3.0ghz mac G5 that would have run those nice cool running, low watt 907GX chips. (the ones that would have made G5 laptops viable)
I strongly suspect that much of the optimization work was done prior to the decision to switch Intel processors. Because of this, there would be scope for this to be done with SL, but possibly not than later OS's.
The ability to get into the OS using the version you and I have on our hard drives is key to this. You are very clearly very sure this flat out cannot be done, I'm not.
Bottom line is that I have always suspected the descision to ditch PPC support was taken relatively late in the development cycle, and the PPC code strip out would have been relatively hurried. (I have a sneaking suspicion that it's actually a bit of a mess in there)
The big uncertainty in my mind, is how easily lines of code written in different programing languages could be made to work together.
[doublepost=1461656564][/doublepost]Ultimately, SL may have been written in Cocoa, but that's a developer tool, not a programing language.
Leopard and Snow Leopard are both a mix of mostly C and C++ code. As are Windows and Linux.
I do strongly suspect that some perhaps even all of SL's features could (theoretically) be patched. maybe even copied and pasted into a decompiled Leopard.
My reason is, I strongly suspect that the released version of SL is essentially a reduced version of an SL Apple developed for the planned but canned 3.0ghz mac G5 that would have run those nice cool running, low watt 907GX chips. (the ones that would have made G5 laptops viable)
I strongly suspect that much of the optimization work was done prior to the decision to switch Intel processors. Because of this, there would be scope for this to be done with SL, but possibly not than later OS's.
The ability to get into the OS using the version you and I have on our hard drives is key to this. You are very clearly very sure this flat out cannot be done, I'm not.
Bottom line is that I have always suspected the descision to ditch PPC support was taken relatively late in the development cycle, and the PPC code strip out would have been relatively hurried. (I have a sneaking suspicion that it's actually a bit of a mess in there)
The big uncertainty in my mind, is how easily lines of code written in different programing languages could be made to work together.
[doublepost=1461656564][/doublepost]Ultimately, SL may have been written in Cocoa, but that's a developer tool, not a programing language.
Leopard and Snow Leopard are both a mix of mostly C and C++ code. As are Windows and Linux.
Bear in mind that Leopard was released 2 years after the Intel transition was announced - I suspect by then, any loyalty to PPC in the OS was a necessary evil rather than a core value - we know Steve Job's opinion on "laggards" who wouldn't keep up! Just look at the hardware requirement shift that came with Leopard as opposed to earlier releases - a sign that it was written and designed with Intel in mind.
So, by the time of Snow Leopard, I think any vestige of PPC compatibility would have been long gone or just coincidental.
EDIT:
To remind myself, those hardware requirements again:
10.0 Cheetah 128Mb G3
10.1 Puma 128Mb G3
10.2 Jaguar 128Mb G3
10.3 Panther 128Mb G3
10.4 Tiger 256Mb G3
10.5 Leopard 512Mb G4 867Mhz
So, by the time of Snow Leopard, I think any vestige of PPC compatibility would have been long gone or just coincidental.
EDIT:
To remind myself, those hardware requirements again:
10.0 Cheetah 128Mb G3
10.1 Puma 128Mb G3
10.2 Jaguar 128Mb G3
10.3 Panther 128Mb G3
10.4 Tiger 256Mb G3
10.5 Leopard 512Mb G4 867Mhz
Last edited:
You can't decompile OS X. Think of an operating system as being like a cake. You can't unbake it, end up with raw ingredients, and bake those ingredients again into cupcakes.
Certainly, Apple could have put features for PowerPC into Snow Leopard, because they have the source code. The source code in the cake analogy equates to the recipe and the raw ingredients. (Yes, this analogy is overly simplified.)
Let's continue with the cake analogy:
Call Intel support a cake, and PowerPC support cupcakes. Basically the same ingredients, but they're cooked differently. The baking process for the cake equates to the compilation process for the operating system. Apple used the ingredients to bake a cake, not cupcakes. There's no way to go back from the cake you have and turn it into freshly baked cupcakes. (And no, cutting up the cake into round shapes does not give you cupcakes either.)
The lack of PowerPC support within Snow Leopard isn't a bug or something that can be fixed easily. You can't just develop a "patch" and get it to work properly on those machines. Here's an example: Leopard was only officially available for G4 processors above a certain MHz threshold (800 MHz?), but Macs with processors that didn't meet these requirements could still run Leopard through a hack like LeopardAssist. This is because the threshold Apple set was arbitrarily created, and the resources to get Leopard running on these systems were still there. So, it all worked, and it wasn't hard for a developer to come up with a way to fool the system requirement checked during the installation.
However, this thought process can't be applied for Snow Leopard and PPC machines. A whole, separate version of Snow Leopard would have had to be created because PowerPC is a completely different processor architecture. Snow Leopard was released for Intel only, with PPC code stripped out, so there would never be any way to get it to run on PowerPC. Apple didn't provide a way for it to work, and they are the only ones who can because Mac OS X in it's finished form is a proprietary, closed source OS. The resources to get Snow Leopard working on PPC Mac's simply were not there. PowerPC systems didn't meet the requirements and there was nothing that could be done this time. No "patch" could EVER be developed.
You cannot decompile an operating system. (The cake analogy above is just perfect!)
Snow Leopard is NOT a reduced version of an OS that Apple developed for a 3.0 GHz G5 Mac. First off, Apple never ever got any G5 chip that ran at 3.0 GHz. The furthest they ever were able to take the G5 was 2.7 GHz. This processor didn't exist, and as time went on, looked as if it might never exist. Mac OS X already ran on G5's anyways, so why would they have needed to use development resources creating a special OS X version for it? The 3.0 GHz used the PPC architecture just like the rest of the Mac's at the time, anyways. Also, since when does Apple develop an operating system for one specific processor, running at only one clock speed? Snow Leopard was released in 2009, so presumably development on it commenced in earnest sometime in 2007, I would imagine. By that time, Apple had already completed the Intel transition and had been selling only Intel based Mac's for about a year. However, if what you are saying is true, then that means that Apple must have had a working version of Snow Leopard by late 2004 to early 2005? (If they did have that, then why release Leopard in the first place?) Do you really, truly think that Apple had a working version of PowerPC Snow Leopard at that time? That it was complete enough that they were doing optimizations? Reality shows us that Snow Leopard was most likely developed from scratch to be Intel only; that Apple wanted to cut ties with their old, legacy CPU architecture. Dropping support for PPC earlier than they may have had to, earlier than their customers probably would have liked is a classic Apple move. Did they work on a version of Snow Leopard for PowerPC? Perhaps, but I think it was decided early on that it was going to be Intel only. Now, it is certainly possible that they only decided to axe PPC late during the development cycle, but I think that most evidence shows that this is very unlikely.
Why are you talking about uncertainty with respect to different programming languages? There is no uncertainty. Mac OS X has always used different programming languages, and they've always worked together just fine regardless of what architecture they are compiled for. The Cocoa API that developers use is primarily Objective-C (but you can use C and C++), while OS X's kernel is mostly written in C, and there are a few other areas that use C++. Cocoa was derived from NeXTSTEP and has been built into OS X since it's inception. These doesn't mean that Snow Leopard will run on PPC, however.
We are all very clearly sure that nothing can be done because it can't be.
I'm not going to explain why I'm so sure of what I say, so I'll just say it again, and make this clear, feel free to read between the lines.
I have very good reasons to believe that it is possible to decompile C and C++.
The description I was given, was that it is extremely time consuming (and mildly soul destroying) to go through vast amounts of decompiled code to find specific bits for nodding or deletion.
I have very good reasons to believe that it is possible to decompile C and C++.
The description I was given, was that it is extremely time consuming (and mildly soul destroying) to go through vast amounts of decompiled code to find specific bits for nodding or deletion.