Mac OS X 10.6 to Drop PowerPC Support?

[shawnce]

I don't believe it. Because one has to ask: WHY? Delivering PPC code on OSX install DVD costs Apple nothing.

It costs them a lot in testing and maintenance of the PPC tool chain, etc.

With that said I don't see a strong reason why Apple would drop PowerPC support in 10.6 unless 10.6 brings some radical new things that make maintaining support PowerPC difficult.

 

[FoxyKaye]

It seems like a logical step to me. If Apple releases 10.6 in 2 years then the last Power PC computers will be 4 years old.

There's nothing logical about it whatsoever - it's just Apple getting in to the planned obsolescence game. A four year-old computer is perfectly functional depending on the use, but the real problem will come when developers stop using PPC compatible binaries because they're "no longer necessary."

The switch to Intel-based Macs is only two years old - what about all that BS about Universal Binaries and not leaving PPC Macs behind that Steve spouted when they went Intel?

 

[JesterJJZ]

I don't buy it. Apple will always develop for both platforms...just in case they ever decide to switch back. If they say they developed OSX for Intel since day one as well, they can keep going with OSX for PPC too.

 

[slffl]

I have a 17" Powerbook and I think this is GREAT news! Focus resources on something else.

 

[KindredMAC]

I agree with the few on here that have mentioned support for PPC Macs as a "Just in Case" clause...

If IBM happens to get over their hump of Speed vs Heat vs Power and comes up with a PPC chip that blows the doors off anything Intel or AMD have road mapped, do you really think Apple is going to let it sit there and not use it?

Why else have they been keeping OS X Universal for so many years?

IBM has some great prelim data on that POWER6 chip that they talked about a few months back. I would love to have a PowerMac with 2 of those babies inside running 16GB of RAM (or 32 if it could handle it).

If Apple holds tight to Universal Binaries and makes others hold fast to it, you could end up seeing computer lines that feature PPC in desktops and x86 in laptop/mobile devices.

The smart thing for Apple at this point is to keep the window open if the door ever gets closed.

Oh yeah and also..... don't know why I'm getting worked up over this while 10.4.11 and 10.5 haven't even hit market yet.

 

[mozmac]

This is one quality that separates Apple from Microsoft. Apple is willing to cut legacy weight and move forward. I get chills down my spine when I look at product specs on software that say, "Windows 98/ME/2000/XP/Vista" You shouldn't support an OS that was made before Google became a business, because a lot has changed since then. If Microsoft would have dropped a lot of that support for Vista, it would be a much better OS. They can't seem to let go of the past.

Apple ruffles lots of feathers with moves like this, but it allows them to continue to innovate. Where would OS X be if it still supported Classic?

 

[sosumi1981]

I wouldn't worry too much

Apple will have dropped support for the Mac entirely by 10.6. They'll just support iPhones and iPods

 

[age234]

Yep, no doubt. Supporting three processor platforms (ARM, PPC, and Intel) is a drain on resources. Time to shift those software engineers to other neat'o projects.

I doubt it's really that much of a drain on resources. They've been doing it for 7+ years, I'm sure they have things downpat.

Unless they do some really wild stuff, there is no performance-based reason to drop PPC support, at least for G5s. All it would do is satisfy the Thurston Howells of Intel Mac users.

 

[termite]

Contrary to what has been suggested, supporting the G3 is not significantly harder than supporting the G5. So either Apple will continue to support the G3, or it will drop all PowerPC support. That said, I think it would be extremely foolish to drop the PowerPC support. Here's the pros and cons from Apple's perspective:

Pro drop:

• Save a little bit of Apple labor on XCode and such.
• Save some testing time.
• Some people will buy new machines.

Con drop:

• Some real bugs show up better in one arch than the other so your code will be less buggy if you support multiple archs.
• Assuming Apple does not want to be tied to Intel permanently, they need to continue supporting other desktop archs in darwin.
• Apple supports the iPhone's processor, so it has to maintain portability anyway.
• Pissing off a first generation G5 user who has a good computer with quad-2GHz processor and 8GB of memory is silly -- that's still an excellent computer.

 

[fourthtunz]

There are some very fast (and expensive) Power PCs being made, just not by Apple.
Apple went to Intel for performance per dollar and the ability to run Windows inside a VM.

How many people actually upgrade the OS when a new version comes out. I'd bet it's at most half. I think a lot of user run with whatever OS version was on their system when they bought it. A lot of people still use Panther. So, when 10.6 is released in 2009 or 2010 all those G3 and G4 users, many of them will still be using 10.3 or 10.4

Not being able to upgrade the OSdoes not mean you have to stop using the computer. It will still work just as well as before.

You are right but how many of us want the latest OS anyway? I do
I guess it depends on your perspective too, I have a G5 and a couple of G4's. I wouldn't expect the G4's to be supported too much longer but the G5's, well, we'll have to see what happens when it happens
daniel

 

[gmanrique]

Couldn't agree more...

This article should hardly make it to Page 2

Let's get 10.5 on store shelves before we even start pretending rumors about 10.6 are relevant to our lives.

 

[mozmac]

I agree with the few on here that have mentioned support for PPC Macs as a "Just in Case" clause...

If IBM happens to get over their hump of Speed vs Heat vs Power and comes up with a PPC chip that blows the doors off anything Intel or AMD have road mapped, do you really think Apple is going to let it sit there and not use it?

Why else have they been keeping OS X Universal for so many years?

IBM has some great prelim data on that POWER6 chip that they talked about a few months back. I would love to have a PowerMac with 2 of those babies inside running 16GB of RAM (or 32 if it could handle it).

If Apple holds tight to Universal Binaries and makes others hold fast to it, you could end up seeing computer lines that feature PPC in desktops and x86 in laptop/mobile devices.

The smart thing for Apple at this point is to keep the window open if the door ever gets closed.

Oh yeah and also..... don't know why I'm getting worked up over this while 10.4.11 and 10.5 haven't even hit market yet.

You make a great point. Apple is great about keeping their options open, so I can see them officially canceling PPC support, but secretly keeping development going in case the PPC chips develop the cure for cancer.

 

[visor]

What a stupid post. I mean - we haven't even got our hands on 10.5 and we're speculating about 10.6. Complete nonsense.
Nevertheless, we won't see 10.6 anytime before 2010 in the open, by then, i will have bought at about two new Macs in a row. I'm fairly convinced that 10.6 will run on both of them.

So, who cares anyway? Besides, an OS will at it's core always need to be multi architecture. There is just not much reason not to be. They use gcc as compiler, it compiles on any architecture anyway ...

 

[mklos]

I can't believe how many people are saying that it would "make sense".

It would not make sense, and Apple is not going to do it. Think about it, kids: the bulk of the Apple Insider article is about a mere 67Mhz increase in the minimum requirements for Leopard...and why? Because they discovered that it simply wouldn't be performant on lesser machines -- and that's a reasonable expectation, since threading things (like the network stack) does have some overhead associated with it and the older, slower machines are going to be the ones bit by that tradeoff.

But WTF...you think a quad, or dual, or even single G5 at 2.0 Ghz is not going to run 10.6 adequately for some reason? As if. The trend in OSX from release to release is that, on most machines with adequate memory, it gets faster -- not slower. The machines that get slower are the edge cases.

It costs Apple almost nothing to maintain the PPC port of this software (which, you should recall, ran on 040, SPARC, and HP PA-RISC, and now runs on PPC, x86, and ARM and who knows what else in the lab) and they're not going to anger anybody with even a dual G4 tower for a long, long time.

I'd put money on it.

The fact that you say it costs Apple nothing to maintain a PPC version is just wrong. Its not like they can just make an update and then port it over. This is a lot of hard work. Originally, OS X has over 86 million lines of code. With Intel its a lot more. So when you start making a major release (Mac OS 10.6) then you have to put a TON of engineering into making your OS work for TWO entirely different processor architectures. If you don't think that doesn't cost Apple a lot of extra time and money you're surely mistaken... Plus, every time you make a small update (10.6.1) you have to make sure it works for both processor architectures. It no longer makes any sense to continue to do 2 processor architectures.

If I remember correctly, Apple did say they would only support the PPC processor for another 3, maybe 4 years. If so, Leopard will be the last PPC release. This gives users 4 years to make the switch.

As for servers, well if it ain't broke, don't fix it. Just because there's an update available, doesn't mean you have to install it. Yes, its very expensive to update servers alone. So if they don't need to be updated then don't. I know a few places that still run Xserves with Mac OS 10.2 and 10.3 Server. They run just fine with what they have.

 

[shawnce]

Onwards and upwards.
Universal binaries add so much bloat. And run slower than optimised binaries.
Don't believe me? Download xSlimmer and tell it to look at your /applications folder. You'll save Gb's
the program strips out the unecessary code (for your machine) and give you a intel ot PPC only app. iTunes for example went from 90Mb to 30Mb, and launches much faster.

Most of that savings in storage space comes from stripping out the localizations and not the universal binary aspect. Also stripping out PowerPC from the universal binary (or vice versa) does not change how fast the program will execute.

At most stripping universal aspects from iTunes will save you around 12 MiB and again it wont affect how fast it executes.

[0:557] > lipo -detailed_info /Applications/iTunes.app/Contents/MacOS/iTunes
Fat header in: /Applications/iTunes.app/Contents/MacOS/iTunes
fat_magic 0xcafebabe
nfat_arch 2
architecture ppc
cputype CPU_TYPE_POWERPC
cpusubtype CPU_SUBTYPE_POWERPC_ALL
offset 4096
size 13650368 << ~ 12 MiB
align 2^12 (4096)
architecture i386
cputype CPU_TYPE_I386
cpusubtype CPU_SUBTYPE_I386_ALL
offset 13656064
size 13373008 << ~ 12 MiB
align 2^12 (4096)

[0:558] > ll /Applications/iTunes.app/Contents/MacOS/iTunes
-rwxrwxr-x 1 root admin 25M Sep 14 09:34 /Applications/iTunes.app/Contents/MacOS/iTunes*

 

[happylittlemac]

PPC systems are already starting to show their age and by 9/10 will be very much out of date, still even 10.5 will still be pretty cutting edge for a few years anyhow, 10,2 and 3 are even still in use in many PPC Mac's.

 

[Antares]

Well, what kind of new features do you think would possibly be included in 10.6? Does forcing Universal support for both PowerPC and Intel chips actually preclude Apple from including any features in the OS? Or, is it simply just an extra development expense which will see little or no return for Apple?

Another 2 or 3 years of support for PowerPC Macs should be fine. It's not like they would suddenly stop working once 10.6 is released. As other people stated, the majority of people who keep older systems aren't the type that upgrade software with every new release. And, couldn't software developers write their applications so that it supports previous version of OSX?

 

[mklos]

I don't buy it. Apple will always develop for both platforms...just in case they ever decide to switch back. If they say they developed OSX for Intel since day one as well, they can keep going with OSX for PPC too.

Apple stated when they announced the switch to Intel that they will drop PPC support in 4 years. That was in what, 2004 or 2005? So by 2009, no PPC Macs will be supported.

 

[Much Ado]

Does this news really bother people, considering we are still waiting for 10.5?

 

[LizKat]

That sucks. There's a lot of life left in the PPC Macs.

Well there's life left in Tiger, too. Panther's not dead either. Depending on what you do with your PPC machines, they could be fine for years not even moving up to Leopard.

But, when you let your setups drift back like that, it's not a free ride. You may have a lot of extra work to do when you finally decide to get up to speed. Your apps may use proprietary formats not supported in later versions on later OS. Your external media or their sharing modes may become obsolete. That can make bringing your data forward pretty painful.

So, when the official "evolve or die" flare goes up from Cupertino about the last OS a PPC machine can run, it's time to round up at least a low-end machine that will run whatever OS is then current or about to roll out.

 

[Cloudane]

Lots of moaning and gurning as usual

Seems sensible to me. 10.6 is a long way off, and the PowerPC models are feeling aged already. I don't see the problem - the types who like to cling on to a computer until it falls over and dies (and yes, I do know a few) can always use 10.5 or earlier still so it's not like they'll lose anything? In most cases such people are too tight to buy an OS upgrade anyway (never mind a better computer!). Those who aren't tight are simply resistant to change - so again, I don't see them upgrading the OS. I still know of someone who is determined to stick with Windows 98.

Perhaps they learned something from Windows, which is *still* clinging on to the same old BIOS from the early 80s and suffered from clinging on to 16 bit for decades more than was necessary.

 

[invalidname]

Called it

Dang, I predicted this, two years ago:

So, I suspect that Leopard is the end of the line for PowerPC, and that 10.6 will be Intel-only. That means you are buying into a four-year dead-end on PowerPC.

This really shouldn't surprise anyone. Apple sees hardware as obsolete after about four years, and they haven't sold a PPC system since mid-'06.

 

[macFanDave]

And let's not pretend that everyone. . .

switches to the new OS version the day it is released!

Suppose the next system, 10.6, is named for the biggest of the big cats: Garfield. Garfield will probably be released around October 2009. It will take several months for a large majority of the base to install it. Also, as history shows, the advent of apps that require Garfield at a minimum will be around two years later (i.e. late 2011).

I'd also wager that people who hold on to old hardware year after year are not those who are early adopters of new OS's. (I'd be an exception because the Mac's durability gives me great pride, but I like the new OS's and my work on the Mac is usually not time-critical, so a delay caused by reverting to an older system is no big deal.)

I'd also bet that having to support two platforms, Intel and PPC, is one of the reasons that Leopard development has been delayed. Making Garfield only for Intel will probably ease the development path, and that would outweigh the cost of throwing PPC users to the wind.

For the record, my family is using three PPC machines right now, but replacing them in 2009/2010 is a reasonable plan.

See you at Leopard night next month, and at Garfield night in '09 or '10!

 

[yetanotherdave]

Most of that savings in space comes from stripping out the localizations and not the universal binary aspect. Also stripping out PowerPC from the universal binary (or vice versa) does not change how fast the program will execute.

At most stripping universal aspects from iTunes will save you around 12 MiB and again it wont affect how fast it executes.

[0:557] > lipo -detailed_info /Applications/iTunes.app/Contents/MacOS/iTunes
Fat header in: /Applications/iTunes.app/Contents/MacOS/iTunes
fat_magic 0xcafebabe
nfat_arch 2
architecture ppc
cputype CPU_TYPE_POWERPC
cpusubtype CPU_SUBTYPE_POWERPC_ALL
offset 4096
size 13650368 << ~ 12 MiB
align 2^12 (4096)
architecture i386
cputype CPU_TYPE_I386
cpusubtype CPU_SUBTYPE_I386_ALL
offset 13656064
size 13373008 << ~ 12 MiB
align 2^12 (4096)

[0:558] > ll /Applications/iTunes.app/Contents/MacOS/iTunes
-rwxrwxr-x 1 root admin 25M Sep 14 09:34 /Applications/iTunes.app/Contents/MacOS/iTunes*

Just reporting what I experienced, I saw a 60Mb drop in footprint in that one application alone (tell me that doesn't do good things for RAM performance) and faster load times for applications.

 

[flopticalcube]

By the time 10.6 comes out there will be very few PPC Macs around, given the average lifespan of 4 years. With Apple selling 10M Intel PCs per year, this is what resources should be focused on. Furthermore, if G4 support being dropped is a given (since G3 was dropped for Leopard), that leaves the G5 iMac and the PowerMacs which is probably not a big enough base worth supporting, especially by 2009/10.