Mac OS X Snow Leopard Drops PowerPC Support

[PurrBall]

I though with Vista, they dropped support for 16 bit apps?

Unfortunately they didn't, I can still run apps from Windows 1.0 just fine.

While dropping PPC would be a great move from a developing standpoint, I think the number of customers that still run PPC that wouldnt be able to buy Snow Leopard would still result in less profit.

 

[compuguy1088]

Consider Apple was developing Intel builds of OSX from almost the beginning along with the actual release builds, that probably wouldn't be a major issue for them.

True...though having just one team focusing on x86/x86-64 would allow for more progress on future versions of OSX. Though currently Orchard Spy mentioned that they are developing PPC builds of Snow Leopard. There may be the possibility of there being two different versions of 10.6, one for Intel, and one for PPC.

Unfortunately they didn't, I can still run apps from Windows 1.0 just fine.

While dropping PPC would be a great move from a developing standpoint, I think the number of customers that still run PPC that wouldnt be able to buy Snow Leopard would still result in less profit.

I recall that certain 16 bit installers for certain games (X-wing alliance) would not work under vista for exactly this reason...unless this is only true with 64-bit Vista (which may be the case).

 

[logaut]

All this discussion just gives me feeling that almost all of 'just Intel' users are actually very latent Windows users.

Maybe they should go 'home' and fight for better support.
(almost is mentioned)

 

[PurrBall]

True...though having just one team focusing on x86/x86-64 would allow for more progress on future versions of OSX. Though currently Orchard Spy mentioned that they are developing PPC builds of Snow Leopard. There may be the possibility of there being two different versions of 10.6, one for Intel, and one for PPC.



I recall that certain 16 bit installers for certain games (X-wing alliance) would not work under vista for exactly this reason...unless this is only true with 64-bit Vista (which may be the case).

64-bit Vista dropped 16-bit support, 32-bit still has it.

tbh it's way to early to make assumptions on what will be supported.

 

[dsharits]

This is just like the american civil war

Mac(USA)

Intel-North

PPC-South

We all know who won the civil war

anyways whats the big deal about using leopard? not snow leopard, if you have a ppc enjoy it and leopard, it could always be worse you could be running vista

Let's not forget the North's unwarranted invasion of the South being another parallel

 

[MagnusVonMagnum]

you don't have to upgrade to 10.6 when it comes out. apple will support 10.5 for a long time. heck, they still release updates for panther 10.3.9

Perhaps, but application programmers by far do NOT support older OS revisions. There were a LOT of apps that supported Tiger but not Panther and now we're already seeing some new apps that do not support Tiger anymore despite the fact Tiger is vastly more stable than Leopard (at least for some of us) and a LOT faster.

If PowerPC is as optimized as it can get already as someone said, why is Leopard so dog slow compared to Tiger? Leopard added some new toys, but overall, it's definitely more evolutionary than revolutionary. Some of us have wondered for some time why Leopard is so much more resource hungry, slower and more bloated in general. It feels more like they were rushed to get it finished and didn't bother optmizing ANYTHING in it. With all the iPhone-mania going on and delaying it to begin with, it's not such a huge stretch for some of us to believe that (or at least something close to it). So now they announcing OSX isn't going to get more features for years to come and they're going to put their proverbial 6 programmers working on cleaning up the crap they COULD have cleaned up last year if they weren't stretching their OSX team so thin to push iPhone forward.

The thing is, Apple has a LOT of free operating cash. They COULD just hire more programmers and keep BOTH moving on the fast track, but they seem rather obtuse in that area or are worried about their bottom-line appearing to stock holders or something. I can't fathom any other reason why they would delay OSX to push iPhone stuff forward except they don't have enough employees to do both at the same time.

As for dropping PowerPC, I see no reason for them to do it even if they don't intend to optimize for it. Dropping it won't make Intel machines run faster. The PowerPC code isn't used on Intel machines! Seriously, I think some people on here don't understand how operating systems work. They've been maintaining an Intel kernel for OSX all along folks! It isn't like it appeared out of nowhere in the past two years. They simply didn't release it as it wasn't in their hardware plans until then, but they kept the option there. Maybe Rosetta wasn't ready, but internally they had an Intel OSX the entire life of OSX. You guys had no idea it existed. Now you assume PowerPC support is dead because Apple didn't include YOU in a developer beta. Did they give you an ARM beta of OSX to play with in that build??? No? Does that mean Snow Leopard won't support current iPhones using ARM processors? They ARE running OSX, you know.

As for the comments about ARM being used for iPhone forever, that's just showing pure unadulterated ignorance. It's akin to fortune telling. Many have speculated that Apple's purchase of a semi-conductor company that's specialty is indeed PowerPC chips may very well be because they want to use it in a future iPhone or other gizmo. They ALREADY have PowerPC support for OSX so it would be no big deal to use such a chip if they so chose. Dropping all support internally for PowerPC would be shortsighted on their part because it would limit their future hardware options (even if not for desktops but for mobile devices) and it would be simple to keep it now rather than have to play catch-up in the future should a newer generation PowerPC chip prove to be useful.

 

[MagnusVonMagnum]

What?!?!? I can't install Snow Leopard on my Mac 512k???? This is an outrage! I only get 25 years of use out of it! I'm switching to Vista! I'm getting a Zune!

Oh if only people like you would, some of us would be a lot happier.

 

[MagnusVonMagnum]

Correction; to run faster than Leopard on multicore, Intel processors. I don't think G4's qualify...

I'm getting sick of reading erroneous information in this thread. You seem to think because they mentioned multi-core enhancements, that's the ONLY ones they plan????? Give me a break. Steve Jobs mentioned how little memory it requires compared to Leopard. That smacks more of Tiger efficiency than Leopard BLOAT. Indeed, Snow Leopard SHOULD run AT LEAST as fast as Tiger given previous MacOS updates were always faster, now slower. Leopard is bloatware defined (ok maybe Vista is bloatware defined; Leopard is just a footnote in bloat).

Really, this thread has degraded into a barrage of opinions with no real basis in fact. People in this thread no more know what Snow Leopard will do or support or how fast it'll be in the end than the groundhog in Punxsutawney, PA.

 

[MagnusVonMagnum]

Not if the means of making the code faster is by removing PPC bloat and optimizing for features in Intel hardware

Once again, I'm bewildered by a comment here. *WHAT* "bloat" are you referring to, exactly? Have you run Tiger before? It's WAY faster and less bloated than Leopard and it was released for PowerPC first. Leopard is bloated on PPC just as much as it is on Intel. That tells me the bloat is now processor specific, but rather Apple put OSX on the backburner to further iPhone development (they had to make an ARM version of OSX after all as well which some of you forget or don't even realize since you don't see it on the shelf at Best Buy by itself, but only in the iPhone).

Really, what most of the people on the "North" side of this argument keep saying is PowerPC is old, crappy and is slowing my Intel machine down! Good riddence! That has about as much factual and scientific information in it as a Milky Way bar has healthful ingredients.

(i.e. moving software away from Intel approximations of Altivec into full

Intel approximations of Altivec!?!? What the heck are you talking about??? Do you just make stuff up that sounds good to you or what?

 

[matticus008]

Once again, I'm bewildered by a comment here. *WHAT* "bloat" are you referring to, exactly?

Having to run non-native code on Intel through Rosetta, having to maintain fat binaries with extra code, having to maintain two separate architectural code bases. It shouldn't be particularly bewildering.

I've been a proponent of RISC processors for over 15 years, but this isn't a platform war. It's reality. The decision has been made. PowerPC Macs are over. This day has been coming for a while. Everyone expected PPC to run out of steam around 2010. Most people anticipated dropping support by 10.7. It seems it might happen a bit sooner. If 10.6 provides no new features, there's nothing really lost. PPC machines will continue running Leopard until 10.7 leaves them behind.

Have you run Tiger before? It's WAY faster and less bloated than Leopard and it was released for PowerPC first.

I don't accept your premise, and the history lesson, apart from a statement of the obvious, has nothing to do with future performance optimizations.

That tells me the bloat is now processor specific, but rather Apple put OSX on the backburner

What on earth are you talking about? 'Bloat' is no one, specific thing, and whether the iPhone platform delayed development or not is immaterial.

Really, what most of the people on the "North" side of this argument keep saying is PowerPC is old, crappy and is slowing my Intel machine down!

It is old, and it is dead weight on OS X. Concentrating on hardware-specific optimizations and code improvements on the Intel platform doesn't have anything to do with the PowerPC platform's viability or with its impact on existing machines.

It has to do with letting the past go and concentrating efforts elsewhere.

Intel approximations of Altivec!?!? What the heck are you talking about??? Do you just make stuff up that sounds good to you or what?

SSE2, 3, and 4, components of MMX, and other technologies that accomplish the same thing Altivec does. It's fairly obvious, and your rant notwithstanding, not a controversial statement.

Optimizing speed and efficiency doesn't mean that it magically works on all platforms. If they're making adjustments and refinements for their targeted Intel hardware, those changes do not automatically translate to benefits for PPC users. If Apple chooses not to bother with the effort of making any changes to PPC code and releases 10.6 as Intel-only, so be it. This is like the melodramatic rehashing of the PowerPC switch all over again.

No one shot your dog.

 

[Sage Harupyuia]

This PPC-b!+ching is getting ridiculous.

Look, there IS NOTHING in Snow Leopard that is going to change the way you work. All the bling under the hood is meant to completely tighten up the code for the newer Intel machines so that they can run much faster and more efficiently.

WTF will a PPC Mac user run Snow Leopard for? Will it kill you to NOT spend $120 for an OS X upgrade?

Your PPC Macs will run FINE in Leopard. There WILL be security patches and application updates released for older versions of OS X. Panther received the Quicktime 7.5 update just 2 days ago. Tiger got another security patch a month ago. Apple is NOT leaving behind the older OSes.

You all bitvh so much about what Apple doesnt offer but do you EVEN make use of what Apple has to offer in the first place?

Read through the 20+ pages and see how many people are running PPC Macs on Tiger/Panther and ignoring Leopard. These people swear that Tiger/Panther is superior to Leopard and never made the jump. Do you think those people will give a hoot if PPC is dropped for 10.6? NO.

In the end, what are you PPC Mac users concerned about? Updates. Security patches. Nothing else. you will STILL be getting your precious patches and updates.

You tell me, what is it about Snow Leopard that puts PPC Macs at a disadvantage? Speed? Intel Macs already outclassed PPC Macs in speed since the Core Duo days. Functionality? PPC Macs and Intel Macs can do essentially the same thing.

I'll just be honest: you just want Apple to support your old stuff because you are to stingy with your money.. I have an old iBook G4 and im not complaining about the drop in PPC support. Even if Apple ceases patches and updates for Leopard i'm all for it. Because i know that Leopard runs fine in my iBook and i dont see why i should be bothered that Intel Mac owners are getting better stuff. You want up-to-date stuff, you pay for it.

 

[Manic Mouse]

(just joking guys, in case anyone takes this the wrong way!!)

 

[PCMacUser]

All this talk of RISC CPUs is making me reminisce...

I bought my first RISC computer in 1990 - an Acorn Archimedes with an ARM2 processor. That was back in the day when ARM stood for 'Acorn RISC Machine'... ahhh good times, good times...

 

[GITANAJAVA]

Inevitable, just as Classic support was dropped for Leopard. The OS has to move forwards, and if lack of PPC support is the price for lean, mean, fast code with a small footprint, I'm all for it.

Matt

To Matt and other kindred writers, I'm all for fast code, too, but not at the expense of leaving so many folks behind the eight ball. Not everyone can afford to dump their Powerbook and run out to buy the newest Mac iteration. Sure, we'd love to, but all pocketbooks are not created equal. As of when I'm writing this, the 700+ responses in this thread are running in the neighbourhood of a close 50/50 for/against: a not insignificant representation of those who would like Apple to support Power Macs just a little longer, please.

 

[PurrBall]

This PPC-b!+ching is getting ridiculous.

Look, there IS NOTHING in Snow Leopard that is going to change the way you work. All the bling under the hood is meant to completely tighten up the code for the newer Intel machines so that they can run much faster and more efficiently.

WTF will a PPC Mac user run Snow Leopard for? Will it kill you to NOT spend $120 for an OS X upgrade?

Your PPC Macs will run FINE in Leopard. There WILL be security patches and application updates released for older versions of OS X. Panther received the Quicktime 7.5 update just 2 days ago. Tiger got another security patch a month ago. Apple is NOT leaving behind the older OSes.

You all bitvh so much about what Apple doesnt offer but do you EVEN make use of what Apple has to offer in the first place?

Read through the 20+ pages and see how many people are running PPC Macs on Tiger/Panther and ignoring Leopard. These people swear that Tiger/Panther is superior to Leopard and never made the jump. Do you think those people will give a hoot if PPC is dropped for 10.6? NO.

In the end, what are you PPC Mac users concerned about? Updates. Security patches. Nothing else. you will STILL be getting your precious patches and updates.

You tell me, what is it about Snow Leopard that puts PPC Macs at a disadvantage? Speed? Intel Macs already outclassed PPC Macs in speed since the Core Duo days. Functionality? PPC Macs and Intel Macs can do essentially the same thing.

I'll just be honest: you just want Apple to support your old stuff because you are to stingy with your money.. I have an old iBook G4 and im not complaining about the drop in PPC support. Even if Apple ceases patches and updates for Leopard i'm all for it. Because i know that Leopard runs fine in my iBook and i dont see why i should be bothered that Intel Mac owners are getting better stuff. You want up-to-date stuff, you pay for it.

Our PPC systems are running Leopard as fine as an Intel mac does, why should we not be able to get the latest and greatest on a processor that works absolutely fine?

 

[Imhotep397]

It's too damn early to drop PPC support. Probably half of all Mac users have PowerPC machines and as soon as Apple drops PPC support developers are going to drop Universal binary creation...Which means that in probably less than a year the half of Mac users that have PPC will have no new software...which means that Apple will have less power in software sales because they are selling significantly less software. Dropping PPC support just isn't a smart thing to do.

 

[Manic Mouse]

It's too damn early to drop PPC support. Probably half of all Mac users have PowerPC machines and as soon as Apple drops PPC support developers are going to drop Universal binary creation...Which means that in probably less than a year the half of Mac users that have PPC will have no new software...which means that Apple will have less power in software sales because they are selling significantly less software. Dropping PPC support just isn't a smart thing to do.

Do you think it would have a big impact on software sales? How much software will people will 4 year old computers be buying?

 

[atszyman]

Having to run non-native code on Intel through Rosetta, having to maintain fat binaries with extra code, having to maintain two separate architectural code bases. It shouldn't be particularly bewildering.

And yet they did it for the first 4 versions of OSX with no software sales or revenue generated by the Intel version. No matter what I'd expect them to be running alternative processor versions of OSX internally for years to come. The PPC market may look dead now, but who knows what the future will bring. It's not like IBM has abandoned their POWER processor line.

It is old, and it is dead weight on OS X. Concentrating on hardware-specific optimizations and code improvements on the Intel platform doesn't have anything to do with the PowerPC platform's viability or with its impact on existing machines.

PPC code cannot execute on Intel, Intel code cannot execute on PPC the machine languages are not compatible. Optimizations can be made to the internal functons and compilers to take advantage of each processor and neither provides bloat or instability on the other processor if done properly.

SSE2, 3, and 4, components of MMX, and other technologies that accomplish the same thing Altivec does. It's fairly obvious, and your rant notwithstanding, not a controversial statement.

If they're using any software or machine language to mimic Altivec on Intel I'd be amazed. That would have been the first bottleneck to address when working on Intel and any code that was intended to run on Altivec would have been modified to use MMX or SSE very early on. If they didn't and those are the optimizations they're doing then there's no reason to not support PPC in Snow Leopard. It might not see as much of a performance boost as Intel machines but if they enhance stability it would probably be a worthwhile update for many, and any new features (contrary to reports here nowhere has it been said that it will have zero new features, just not as many as previous releases and more focus on speed/stability).

Optimizing speed and efficiency doesn't mean that it magically works on all platforms. If they're making adjustments and refinements for their targeted Intel hardware, those changes do not automatically translate to benefits for PPC users. If Apple chooses not to bother with the effort of making any changes to PPC code and releases 10.6 as Intel-only, so be it. This is like the melodramatic rehashing of the PowerPC switch all over again.

And just because they said they were optimizing it for speed and efficiency does not mean they are dropping support for a large percentage of the Mac install base.

Until a developer or Apple shows the official announcement of PPC being dropped it's safe to assume that PPC is still there. Developers would want to know so that if they develop 10.6 only apps they don't have to waste all the time and money with optimizing and testing PPC versions of the applications. Since no one has produced any evidence other than the requirements of the Developer Preview I'd put money on PPC support being in the final release, just not quite ready for developers to play with.

 

[Life is Sweet]

Options For SL

My thoughts are that SL could be one of 3 things:

1: The last PPC OS, which ties up the loose ends of OS X so that the last OS for PPCs is as fast and as best as can be, meaning that option 2 would be the next OS.

2: The foundation for a new OS (like an early OS 11) which would have no PPC support. This would be logical since as the base for the future, it wouldn't need to lay the future foundations for a PPC OS.

3: Like option 1, but due to it not being a big leap (unless you're a developer), for the consumer market to upgrade, it will be very cheap ($20 please)

Prices:

It could be cheap, but they may tout the speed and security advantages and charge $129. Maybe if it is a sum up, it'll be like $69.

The name Snow Leopard makes me think it wont be option 2, but option 1 or a mix of 1 and 2. The name is too similar to be a new base.

 

[matticus008]

It's too damn early to drop PPC support. Probably half of all Mac users have PowerPC machines and as soon as Apple drops PPC support developers are going to drop Universal binary creation

Why? If 10.6 preserves API compatibility with 10.5, most software is pretty much a tickbox in compiling to put out a UB, particularly if it's already Universal. There'd be no incentive to stop doing so except when there are no longer enough PPC users to warrant the option (or if the software is resource-intensive, in which case a 4 year old machine isn't the target hardware to begin with).

And yet they did it for the first 4 versions of OSX with no software sales or revenue generated by the Intel version.

As a skunk project not for public consumption, and certainly without optimization. "Works" and "works well" are not the same thing.

The PPC market may look dead now, but who knows what the future will bring.

There's nothing stopping them from retaining basic compatibility in a back lab. That's not the same as preparing a product for release.

PPC code cannot execute on Intel, Intel code cannot execute on PPC the machine languages are not compatible.

Clearly, but this is not relevant.

Optimizations can be made to the internal functons and compilers

That only carries you so far. At the end of the day, it's not the compiler that needs to be trimmed down.

If they're using any software or machine language to mimic Altivec on Intel I'd be amazed.

That's not in slightest what I said. I said that if they're working on optimizations dealing with SSE and MMX family technologies, those don't automagically bounce back to Altivec, and such changes may not be worth the effort or even possible with VMX.

And just because they said they were optimizing it for speed and efficiency does not mean they are dropping support for a large percentage of the Mac install base.

Correlation is not causation. The optimization goal does not imply PPC is being dropped; the system requirements do.

Until a developer or Apple shows the official announcement of PPC being dropped it's safe to assume that PPC is still there.

On the contrary, as long as the developer builds are Intel-only, it's safe to assume that PPC is on the chopping block. Until and unless a PowerPC version is seeded, it seems the rumors are probably true.

Since no one has produced any evidence other than the requirements of the Developer Preview I'd put money on PPC support being in the final release, just not quite ready for developers to play with.

How does that make sense? Why would they strip out PPC rather than just incorporate the existing code with whatever modifications they'd completed? There's no reason to do this unless they've done absolutely nothing with the PPC side, which further supports they idea that they're not going to, because certain kinds of changes would be made in tandem, e.g. implementing OpenCL, Exchange integration (which wouldn't require anything special off the bat and would immediately expand the testing base).

 

[atszyman]

2: The foundation for a new OS (like an early OS 11) which would have no PPC support. This would be logical since as the base for the future, it wouldn't need to lay the future foundations for a PPC OS.

And when OSX started there was no compelling reason to lay the foundations for an Intel version of the OS, yet they did it for years before the Intel switch.

As long as Apple pushes for developers to create universal binaries, they'll have internal development running PPC just like they did for Intel. It's not like they are going to lay off all of their PPC experts or stop internal PPC development the minute the OS no longer supports PPC in the public release.

It probably won't take as many resources as maintaining two full customer releases but they've already done one fairly smooth switch, why not hedge their bets for the future keep the best of both worlds.

Maybe IBM will make a huge breakthrough in the POWER CPU line and it will make sense to have PPC machines, possibly even a mix of PPC and Intel depending on their different performance metrics.

 

[Cabbit]

Our PPC systems are running Leopard as fine as an Intel mac does, why should we not be able to get the latest and greatest on a processor that works absolutely fine?

What latest and greatest, nothing new nothing lost. This is a foundation release for 10.7 and all other future OS releases that means no PPC as there are no future PPC macs on the horizon.
Why build a foundation for two houses when your only going to build one of them?

 

[atszyman]

That only carries you so far. At the end of the day, it's not the compiler that needs to be trimmed down.

That's not what I meant. They are optimizing OS functions. There are two paths for this. One is to recode the functions to get the compiler to recognize that the function can be handled more efficiently by the target processor. The other option is to fix the compiler so that code written will be compiled to take better advantage of the target architecture. By doing compiler optimization it is very likely that you can take code that was written to take advantage of one processor feature (like Altivec) and get the compiler to recognize that it can be done efficiently using another processor's feature (like MMX, or SSE).

That's not in slightest what I said. I said that if they're working on optimizations dealing with SSE and MMX family technologies, those don't automagically bounce back to Altivec, and such changes may not be worth the effort or even possible with VMX.

No, but if the Altivec (or other processor features) is already there and enhanced, optimizing for SSE/MMX doesn't have to break it either, you can write code that will use differently optimized functions depending on your target architecture.

Correlation is not causation. The optimization goal does not imply PPC is being dropped; the system requirements do.

It's a developer preview, system requirements for alpha software don't necessarily correspond with the final release. I'll believe PPC support is dropped when someone produces an official announcement.

On the contrary, as long as the developer builds are Intel-only, it's safe to assume that PPC is on the chopping block. Until and unless a PowerPC version is seeded, it seems the rumors are probably true.

And without an official announcement all they are is rumors. If it's such a pain for Apple to maintain both PPC and Intel optimizations why would they not let their developers know that PPC support was being dropped so they don't have to go through the time and pain of maintaining or testing PPC apps? And don't tell me it's just a check box in Xcode to develop a Universal binary. If that were true we'd have had universal binaries for MS Office and Photoshop the day after the Intel switch was announced.

How does that make sense? Why would they strip out PPC rather than just incorporate the existing code with whatever modifications they'd completed? There's no reason to do this unless they've done absolutely nothing with the PPC side, which further supports they idea that they're not going to, because certain kinds of changes would be made in tandem, e.g. implementing OpenCL, Exchange integration (which wouldn't require anything special off the bat and would immediately expand the testing base).

Who says they stripped out PPC? The only thing we can infer from these requirements is that this preview version doesn't support it. The disk space and memory requirements are the same as they were for 10.5. Since the PPC compiled code will not run on Intel there's no reason to delay developer release if the Intel compiled kernel is stable enough for developers to play with and the requirement of an Intel processor does not mean that the PPC code has been stripped. If the PPC code is such a waste of space and speed why isn't the OS footprint significantly smaller?

Do you support removing Rosetta as well?

Until there's an official announcement we're both speculating. You think Apple is going to drop it because it costs them money to maintain and that it's somehow holding them back. I think that they don't want to risk angering developers so if it were being dropped it would be a big announcement.

I understand where everyone is coming from, I just can't believe that Apple would drop PPC support in an OS without a big announcement to avoid angering any developers, and would sneak the announcement in by hiding it in the system requirements for a developer preview release.

 

[matticus008]

And when OSX started there was no compelling reason to lay the foundations for an Intel version of the OS, yet they did it for years before the Intel switch.

It started in the early 1990s, in fact, long before OS X, and it was specifically because they weren't sure which platform they'd need to use, coming off the 68000 line. In those days, many people thought x86 wouldn't last, because the future looked bright for RISC CPUs. It didn't pan out, and I don't think anyone expects PowerPC to make a consumer-space comeback. What replaces the Core architecture and the x86 family will be something different.

For OS X, NextStep came with Intel compatibility. BSD was Intel-compatible. "No compelling reason" is a misstatement. The foundations had already been laid by preceding projects, and by the time Apple started real development on the Intel version, they'd grown worried about the PowerPC roadmap.

 

[atszyman]

It started in the early 1990s, in fact, long before OS X, and it was specifically because they weren't sure which platform they'd need to use, coming off the 68000 line. In those days, many people thought x86 wouldn't last, because the future looked bright for RISC CPUs. It didn't pan out, and I don't think anyone expects PowerPC to make a consumer-space comeback. What replaces the Core architecture and the x86 family will be something different.

And back in those days x86 was thought to be a dead end architecture. No one really knows what the future of CPUs is and with the video game consoles pushing processing power and sales in a myriad of different directions and two of the big three consoles on PPC derivatives, we could see another turnaround. Why not hedge their bets for the time being.

For OS X, NextStep came with Intel compatibility. BSD was Intel-compatible. "No compelling reason" is a misstatement. The foundations had already been laid by preceding projects, and by the time Apple started real development on the Intel version, they'd grown worried about the PowerPC roadmap.

And the foundation is already there for PPC in OSX. I'm not saying they need to be worried about the Intel roadmap but it never hurts to be cautious. I realize that it doesn't mean they have to release PPC or other processor builds to the public which let's them remain in a rougher state and cheaper to produce maintain, I just don't think Apple is ready to completely halt PPC design and I don't think this is going to be the OS version that drops PPC support.