The 10.6 Report: Time for Apple to Drop PowerPC Support?
June 4th, 2008Before I get started with this commentary, let me assure you that I never take any of my predictions seriously. Yet it seems as if The Night Owl caught a wave with one particular article, and now other sites, including the standard rumor sources, are getting into the act.
So what am I talking about?
Well, on the heels of my admittedly casual prediction, recently, that we might learn preliminary details of Mac OS 10.6 at the WWDC next week, comes a claim that it will not support Macs with PowerPC chips.
Let’s, for the moment, consider the impact of this is true. You see, Intel-based Macs only went on sale in 2006, and the last conversions took place that summer, with the release of the Mac Pro. So if this particular rumor is to be taken seriously, and 10.6 appears some time next year, Apple would abandon support for many Macs that are barely more than three years old. That is certainly not in keeping with their previous approach on dealing with legacy hardware.
From Apple’s standpoint, this move could make a whole lot of sense, despite the pain they’ll be inflicting on many of you. Being able to strip PowerPC code from the system binaries means that upgrade files will be a whole lot smaller, and that would be a boon for those of you suffering lengthy downloads for the stuff that regularly shows up in your Software Update app. It would also mean simpler quality control testing, since only a single processor architecture need to evaluated for compatibility, resulting in far fewer complications and unexpected bugs.
Such a move would, of course, encourage third-party developers to get with the program and begin to offer more and more of their products as Intel-only, for the very same reasons. It would also make it easier for them to optimize cross-platform code bases, and thus such things as games would work more efficiently on Macs. Then again, a lot of the latter is because those games were originally designed for Microsoft’s DirectX graphics system, and they still have to be ported to OpenGL. Call it half a loaf.
From a practical business standpoint, it’s also true that sales of Intel-based Macs are off the charts, and the user base will be noticeably larger than PowerPC Macs at the end of that possible two-year gap between Leopard and its successor. This may indeed be a factor that’s weighing on the minds of Apple’s development and marketing teams, but you have to wonder whether this might just be far too draconian a measure.
Understand there is no confirmation about Apple’s plans for 10.6, or even if it will be showcased, however briefly, at the forthcoming WWDC. Just because a number of sites are raising that possibility, and the additional prospects that Apple will abandon support for the PowerPC, doesn’t mean it’s going to happen.
You see, I feel that action would be premature, even if 10.6 arrived in 2010. I tend to think that 10.7 would be a better operating system to mark the final nail in the coffin for PowerPC. That would take us to 2011 or 2012, and that would be more than enough time for this to happen.
If I were to venture a guess, I’d tell you I see little reason for 10.6 not to have the very same system requirements as Leopard. That means a G4 with an 867MHz processor or better. That, to me, is a perfectly sensible line of demarcation, because it would allow Macs from the early part of this century to survive to the end of the decade without having to be replaced.
This is not to say you shouldn’t get an Intel-based Mac at the earliest opportunity. Regardless of what Apple does, it’s a sure thing more and more application developers will abandon the Universal code base and embrace Intel-only. Adobe has done that with some of its creative applications already, and there are scattered apps from other companies that have the very same limitation.
Since Microsoft Office 2008 appears to suffer somewhat on a PowerPC, I dare say that Office 2010, or whatever it will be called, will make the full exclusive transition to Intel. That will also ease the process to deliver a restored Visual Basic for Applications, since Microsoft’s Mac Business Unit won’t have to work quite so hard when it only has a single processor platform to support.
Sure, the Universal feature in Apple’s Xcode was supposed to be able to simplify that dual-processor paradigm, but it still means more work optimizing an application’s code base. For a sprawling suite, such as the various Office components, that entaila an awful lot of work, and you can understand if the PowerPC wasn’t given full attention.
Regardless of how this all plays out, going to Intel is not quite as expensive as you might think. Believe it or not, the lowly Mac mini is actually more powerful than any Power Mac, other than the higher-end G5’s, and those with heavy-duty graphic cards. Should Apple accept the suggestion from me, Dan Frakes of Macworld, and Daniel Knight, of LowEndMac, to build a mid-priced “headless” iMac, the upgrade dilemma would be eased considerably.
I know there are still millions upon millions of PowerPC Macs in regular use in homes and businesses. I think Apple would be making a monumental mistake to abandon them too quickly. In fact, I think there would be lots and lots of protests if 10.6 did just that.
But I am still feeling more confident in my prediction about 10.6’s unveiling at WWDC 2008, and that’s at least one positive development — at least for me.
Print This Post
Putting out an Intel-only OS is fine by me, if there are good reasons for doing so. But since, as Gene noted, Power PC Macs were being sold until summer 2006, it would be reasonable to think that Apple would continue to support these Macs for their full life-cycle (i. e. out to maybe 2011 or 2012) in terms of such things as issuing security updates.
Next year, 1 TB drives will become common. Just because couple of overkill people “stripping” their applications on a OS are happier, Apple won’t make such a huge mistake as dropping PowerPC. Another thing: It is very, very wrong thing to do on Leopard. Nobody can predict what will Core OS do to tampered binaries and frameworks in 10.5.6 for example. Reject? Safari.app already gets problem when you strip it.
Just imagine how happy will Dell/Microsoft guys will be. “Apple on enterprise? They abandoned their 3 years old machines!”. They are _already_ doing it using the comical 64bit Intel only Java 6 while Java 6 for P3 500 machines running windows 2000 around everywhere.
While speaking about PPC, everyone makes same mistake as focusing only to Mac Mini and iMacs or old Powerbooks.
The G5 line just managed to get a pure 64bit possible OS with Leopard. For example, this machine has 4 G5 64bit processors with 16GB of Max memory. I actually know/used a 16GB maxed machine which reached full potential after Leopard.
A Quad G5 or even a dual G5 with Fiber/SCSI/High End/Pro video card can easily reach $12.000 just for hardware. Consider the high end program licenses. It is not like we throw out a poor mini and buy the intel one.
The NeXT which is the root of OS X is way more than 2 different processors. OS X has support for 10 different architectures in a single executable. If you look at the original NeXT idea (which still exists!), see latest Safari on Windows install the Core Image framework like OSX gems and imagine a market which Mac hits 30% share, you can predict very interesting things for next year.
If Apple got anyone suggesting they drop PowerPC support, they should start a internal investigation for that guy. The respect in enterprise (which is sadly low) they got for all these years will completely go back to negative values.
It is not like they spare months of programmer time to optimise for Altivec eh? G4/G5 users actually get surprised when they approach anything optimised for their CPU. I personally have a single .app which uses multiple architecture/family support of universal binary. It is “X2.App”, a game. It has G3/G4/G5 and Intel binaries in single executable. Size of binary on such a complex game? Comical. About size of a single mp3.
Interesting to speculate, but it ain’t going to happen.
I read a couple stats. that suggest the PPC vs Intel user base is around 50/50. You can’t possibly consider pulling the plug on half of your customers.
I’m in favor if bifurcating Mac OS X for now. I think Apple has pretty much given people the hint and like always people refuse to believe Apple right up to the point where Steve takes the stage and drops the “BOOM”. Hints that are out in plain sight.
1. 64-Bit Intel only Java 6
2. There’s a paucity of WWDC sessions pertaining to Universal Binary or Carbon development
3. Rumors state that 10.6 will not be feature laden. Consider this a concession.
4. The iPhone SDK doesn’t support PPC
5. Large developers like Adobe and Microsoft are eschewing PPC support (Soundbooth, Silverlight)
The writing is clearly on the wall. PPC support is going to be gone soon as in “we not waiting until 2010”. Yes this is going to ruffle some feathers but other than some tweaks to Leopard API and probably the addition of Cocoa Touch to 10.6 to support future devices PPC owners should be able to get what they need done Leopard. Most of the bellyaching will be done as a matter of principal.
Why is Apple doing this? Because even if you can easily compile a UB application you still have twice the QA work. If you decide to target the iPhone with a product now you have 3 platforms to QA. Are we having fun yet? 64-bit apps are Cocoa only as Apple killed Carbon 64.
So what I believe the Two Golden Gate Briges signify is a splitting of OS X. Leopard will be the last PPC supporting OS and 10.6 will be more in sync with the OS X platform as a whole from Xcode to the applications. 10.6 will be heavily optimized for Intel which includes SSE4 and perhaps VPRO management
PPC wasn’t an issue when Apple simply had two platforms to code to. But now we have Mac Intel, Apple TV Intel, Mac PPC. iPhone ARM. Ask yourself this “Do I want to continue to see Apple release buggy OS that require frequent stability downloads?” Folks there’s no free lunch here. OS X engineer bandwidth is finite and asking these guys and gals to produce PPC and Intel versions of OS X with Carbon and Cocoa framworks was a bit much. How many simultaneous transitions are we going to put them through.
10.6 needs to boil things down to what works across the OSX platforms. Intel is the defacto standard and ARM is the mobility standard. Three’s a crowd and PPC has to go for the sake of stabilit and security.
I can think of a LOT worse things than dealing with leopard for another year as a PPC owner. By 10.7 the legions of PPC owners will be refreshing their Macs and enjoying the sane decision made by Apple to pull the reigns back and regroup while they can. The time is perfect. The time is now.
I think Murchison has it about right. This makes good sense from a programmer’s point of view, and the writer of yesterday’s Ars Technica article on this rumor suggested that a better-optimisized OS might have a payoff in longer battery life for present and future portable devices. But the marketing of a new OSX consisting primarily if not exclusively of “under the hood” stuff is another thing entirely. At least since OSX.1, every new feline has come equipped with some wizzy new features and interface improvements (or in some cases “improvements”) which have been promoted to make users willing to pay for the upgrade (Time Machine in Leopard and so forth). Without similar enticements in OSX.6, Apple might experience difficulty in convincing Macintosh users to shell out money for an upgrade intended to make their programmers’ lives easier and perhaps to help develop and market Apple’s line of non-Mac portable devices they sell, which might seem to be of more visible benefit to Apple than to the consumers themselves (no matter how incomplete or inaccurate such a customer perception might be). So, unless “Snow Leopard” is going to be distributed for free, as I seem to remember OSX.1 was, I have trouble accepting the part of the rumor that says no significant new consumer-oriented “hooks” will be forthcoming.
Is it possible to have a “hot rod” version of 10.6 for us Intel owners, and a “stretch limo” version that covers both?
Just because Apple may not support PPC in 10.6 doesn’t necessarily mean that they are “dropping” support altogether. I would imagine that they will issue bug fixes for 10.5 for quite a while. And just because a new OS comes out doesn’t mean that your older one will suddenly stop working…
See prior comments vis a vis the QA process. I do QA for a living. 2 versions of 10.6 makes much less sense than one version supporting multiple platforms. If 10.6 is Intel only, then you already have your “stretch limo” – 10.5.
As for dropping PPC support, I agree completely with Ranger Jay. There are plenty of PCs in use that are supported but can’t run Vista (even though some of them are “Vista capable” or whatever they called it). At some point this will have to happen. Whether or not this is the —well, ask me on Tuesday!
😉
Cheers!
HORRIBLE idea for 10.6…Drop G4 support, maybe.
Yes, but that leaves all those pre-MacBook notebooks unsupported. I don’t think that’s a terribly good idea either. I think it would be all or nothing, and I think just killing off a few additional models would be added torture.
Peace,
Gene
There are no new “I gotta have it” features. I don’t see what people’s hangups are given the dearth of information that we have. Most G4 Macs couldn’t accept enough RAM to make 64-bit worthwhile. I think people are stuck with the thought that this is somehow going to be another OS update with 300 new features and marketing blitz. Nope. Expect this to be the new OS X 10.1.
I don’t see why we can’t get a release to fix what’s been going off the rails lately that supports the same hardware as 10.5. Should Apple do that they’ll gain immense respect and good will from the IT world. Cutting off 3 year old hardware will do the opposite. A quick look at what’s happened on the Redmond side of the fence proves that. Vista cut off all but the most recent hardware and it’s being actively avoided. Yes I know there are other reasons for that, but indulge me. Fortunately Windows users can still choose XP which is supported by all software developers and runs fine on virtually any PC made since 2000.
The IT community is on a 3-5 yr refresh cycle. They understand support issues far more than the typical consumer. Which each successive release IT members in larger companies must continually rebuild and test their images to make sure things work. If 10.6 represents merely a security and stability upgrade there is nothing that prevents Apple from taking the same approach with 10.5.x. What I believe 10.6 will bring is a defined roadmap for Apple and developers. Carbon is out and PPC could be out. The iPhone, .Mac/MobileMe and upgrade development tools will all congeal into a platform that is easier for Apple to manage and easier for developers to build supporting apps. Fretting over PPC support is intriquing because eventually those PPC Macs are going to be replaced and that replacement will be with Intel based Macs. You may as well let Apple optimize the hell out of Macintel and reap the rewards after the next refresh.
Apple stated when Intel came out that they would support the G5’s for many years, I do not remember the exact number, but we are no where near it.
I agree as a part time programmer that cutting out PPC is best for apple in the long run. I also agree that this would be a significant short term mistake due to the risk of alienating those that bought the last round of G5’s, as well as those like myself that have G4’s and G5’s and Intel Macs. I am currently running OSX on most and still run OS9 on a few due to legacy software that is unlikely to change until i is rewritten in a few years.
I feel that the next OS should remain binary but stipulate that the following one would be Intel only so as to avoid bad feelings,
And though the IT world may have a 3-5 year cycle, home Mac’s have had a 5-10 year usage cycle and PPC’s were still being sold up to 2-4 years ago.
If they force PPC users by alienating them from paid OS upgrade even before Leopard still not completely de-rushed, they may really move to another thing rather than Intel based Mac.
It doesn’t have to be a Mac you know… If a Vendor plays such games to force people to do things, see what happened to Microsoft. How couple of mistakes made them lose entire user trust and sympathy.
I actually saw a Microsoft representative (reseller) use Java 6 being 64bit Intel only as an example about how Apple can’t be trusted for business things.
I better repeat: If Apple makes PowerPC Mac users mad, they may switch to other solutions. Also Linux is out of question. It doesn’t have ppc drivers (still in progress), no flash support etc. Putting customers to that situation wouldn’t be great for Apple.
I think hmurchison has this right.
I expect that the security and stability improvements will actually appear in both 10.6 and a free upgrade to 10.5.x (available for both PPC and Intel). I think the selling point (the reason that someone will pay) for 10.6 will be speed improvements based exclusively on Intel architecture – something that cannot be offered to existing PPC owners.
Expect the 10.6 upgrade to be less expensive – somewhere between $50 and $70 – and then Intel Mac owners just need to decide if they want more speed. This is a near-perfect way for Apple to signal the end of PPC without actually ending support while also laying the base for all future OS X upgrades (as opposed to security releases).
Everything went the way you said it would. What to do if you bought a powerpc in the summer of 2006? I kinda feel left in the dark here.