That’s right, folks. In the words of a famous intergalactic travel guide: DON’T PANIC.
In a year from now, there will be Macs for sale with Intel inside. (Does that mean they will be forced to sport an ugly “Intel Inside” badge on their otherwise sleek and attractively designed cases? I doubt that Steve Jobs or Jonathan Ive would consider that for even a second. Have any recent Macs had any kind of IBM logo on their exterior?)
In two years from now, all new Macs will be running on Intel processors. (Does that mean you’ll be able to buy and install OS X on any Intel-based PC? Well, not officially. In fact, while it will no doubt be technically possible, I suspect it will remain very, very difficult for all but the geekiest geeks.)
During the next two years, anyone in the business of creating software for OS X will, at the very least, have to recompile their code to get it to run on the new Intel-based hardware. Even if they don’t, a translator called Rosetta will attempt to make PPC-designed code work transparently on the Intel chipset. Cross-platform ‘fat binaries’ will allow developers to create apps that run on either chip.
Anyone who thinks Apple is full of moles should think again, because there have been x86-based builds of every single OS X release for the last five years. They’ve been working on Rosetta and on the Intel transition kit for a long time. None of this stuff got leaked until the very last moment, when C|Net scooped us all last week. Even then, details were thin and speculation was allowed to go crazy over the weekend.
Apple’s been preparing for this for five years. That’s contingency, people. That’s planning ahead. That’s Steve Jobs all over.
Yes, there will be problems for sure. There’s bound to be fat binaries that won’t work on one, or both chipsets. There’s bound to be third-party hardware, software and hackery that goes haywire. There’s bound to be a lot of moaning and complaining, some of it justified.
Three thoughts lurking in my head this evening:
- Does this mean the current G4 laptop line gets no significant processing boost at all until next year?
- What does the “3.6 GHz Pentium 4 Mac” offered in the new developer kit look like? Is it just a Power Mac with different insides? Something completely different?
- No-one at Apple seems to have answered this question: “Why should I buy any Apple hardware between now and 2006?” Is a G4 or G5 worth the same sort of money now?
- I’m not worried. I’m not in any kind of panic. I’ve never felt any kind of loyalty to the chip, only to the computer as a whole. To my own surprise, I’m curious. I want see OS X for Intel at work. I can’t wait.
Go crazy, now’s as good a time as any
Apple loyalty
"I've never felt any kind of loyalty to the chip, only to the computer as a whole. To my own surprise, I'm curious. I want see OS X for Intel at work. I can't wait."
Too true. I saw way too many people after the keynote saying. "Well, there goes Apple." "I'll go get my Fedora CDs." To those people, I say "Good riddance." To tell the truth, how many of you have stopped what you were doing, sighed and said: "I love having a PPC processor!", before hugging your computer? None. The Macintosh experience is about the software, the user experience. Hardware just supports that. Switching from a PPC to an Intel processor won't change that. Steve himself demonstrated that everything still "just works".
As Giles said, don't panic. This'll be fun.
Issues in the transition...
From a developer persepective, here are a few major points worth considering:
A switch to Intel changes endianness of data. This really shouldn't be an issue for well-structured code. Darwin has been running on Intel since day 1, and open-source.
New instruction set will hit highly-optimised code using PPC assembly. This will be an issue.
It would be great if we could avoid having Intel's awful 32 bit emulator on the chip. We've switched to 64 bit, so why not 64-bit only Intel CPUs?
We've got "Rosetta" to run PPC-only code on Intel, what about the other direction? The "roadmap" Steve Jobs laid out involved phase-out of PPC support from 2008, which means buying a PPC mac would give it a 3-year lifespan, maximum.
A few bits
I'll admit to being barely coherent, yet -- still suffering the aftereffects of the RDF, and I'm not even at Moscone.
Some bits in your article, though, require a reply. "None of this stuff got leaked..." is not accurate. It's not *in*accurate, either. We've known of project Marklar (OS X on x86) for several years now. We've had rumors of X on x86 for several years now. Nobody with any sense put any credence in them. This year, though, well, things have changed, haven't they?
My point is that there are still plenty of leaks in Apple's organization. We all know that.
Your questions, though, show the proper direction of inquiry and analysis at this stage. Steve introduced iTunes 4.9 with podcasting, and then introduced the switch to Intel. That's it. No commentary on the current line-up, no new products, no new software (well, aside from XCode 2.1, iTunes 4.9 and X on x86, which is plenty if you stop to consider).
What's next?
Some of your questions, though, are slightly misdirected. Will the laptop line-up get a significant processing boost? No, it will not. Once they moved to the G4, they haven't had *any* significant speed boost. Incremental, yes. Evolutionary, yes. Significant? Not since the jump from G3 to G4.
Why should someone buy apple hardware now, rather than wait for the new Intel-based Macs? Well, because you want or need to run OS X or an OS X-based application now. Because your Mac's value won't suddenly decline as soon as the Intel-based computers are available. If you buy today, your Mac will be useful for the next year. And the year after that. And the year after that. And, knowing Macs, very possibly the next two to three years after that. In one of the labs where I work, there's still a PMac 6100 running an instrument. It isn't used very often, but it's still there and it still works.
For what it's worth, I don't expect Apple to place a moratorium on all processor upgrades. As IBM (and FreeScale) manages to bring forth new classes of processor, I think we can expect to see new Macs based on them. At least, when and as appropriate for the line in question. So, since PowerBooks were recently updated, and iMacs more recently than that, and iBooks are semi-recent, as are PowerMacs... um, we should see new eMacs next week, right? :) More seriously, I wouldn't expect radical new introductions (e.g., dual core), but the occasional bump wouldn't surprise me.
Overall, though, I strongly share your curiosity. Or at least, I will once my case of WWDC-induced nerves settles down...
A couple more things
Discussion of AltiVec is conspicuously absent. Anyone doing hard core DSP in audio or video will likely have their work cut out for them.
I am interested to see what Apple does to keep sales up over the next year or two. I hope they don't fall victim to the "Osborne" effect.
There is a real market perception problem lurking here, I think. We've been able to say for a while that you can't directly compare the platforms, that it is not an "Apples to Apples" comparison. The Apple hardware will be so similar to Windows hardware, in many ways, that the comparison is inevitable. I think that Apple's margins are going to have to shrink because of this.
A few bits
Be careful! You're insulting me and at the same time *any* of the 50,000 NeXT/OpenStep users that always _knew_ (without the need to be an insider) that there always was a version of OpenStep/OSX running on a intel machine...
=;)
Can I panic a little?
I really don't care what CPU is in my computer. I used to back in the old days of Mac vs. Windows but since the rise of Java it should be obvious that the days of CPU dependency are coming to a close. I personally think that as soon as this transition to x86 is complete, Apple should stand up and say "We reserve the right to switch to any CPU architecture at any time. We will provide backward compatibility but don't expect us to be locked into any technology ever again."
Now for the (some) panic part. What happens in the next 18 months to Apple's sales? Apple is going to have to do some serious evangelizing about how the CPU doesn't matter or they will lose a huge percentage of their sales. This is going to look very bad on the bottom line for the next year or so. I haven't seen anything that says what Apple intends to do about it. That is very nervous making.
the mac experience includes the hardware...
...but not the chip!
so the only question which i have currently is: WILL X RUN ON DELL PCS? I hope not...
Can I panic a little?
"Apple is going to have to do some serious evangelizing about how the CPU doesn't matter or they will lose a huge percentage of their sales."
That's a bit of a no-win situation though. If one argues that the CPU doesn't matter, then why make the switch in the first place?
PANIC
There are a couple of reasons to flip out. None of them are related to will my stuff run... it will Apple is very good at that. We'll all be safe in that respect for prolly about 4 years from now...
PANIC #1: The poorest chip architecture on the market is now the the only viable architecture. x86 is a nasty inefficient beast. Intel has been able to keep it going simply by the shear volume of money it made selling unreliable, cheap buggy, chips. You only have the right t attribute half your Windows crashes to Microsoft the other half are all Intels fault.
PANIC #2: More Idealistic but very important. We call this one economics. Intel has essentially wipped out it's competition. Yes there is AMD who manufactures better, faster, and cooler chips, but in reallity AMD is kinda like Intels Flavor Flav. Intel no longer has a competing architecture on teh desktop. Expect x86 chipest prices to increase and innovation in chips to decrease in the future. Well at least if Adam Smith had even the remotest clue.
PANIC #3: IBM may be a slow moving bohemeth but let's be real, once it gets going, like a frieght train it goes solid and strong. There are two things to look at here IBM is replacing all of it's server chips with Power and PowerPC chips, one can expect things to get much more interesting in that front. And can does anybody remember the word CELL? Can you say 200 GFlops per chip by next summer? Wonder if Intel has anything like that up it's sleave, sooooo doubt it.
Bingo!
"No-one at Apple seems to have answered this question: 'Why should I buy any Apple hardware between now and 2006?' Is a G4 or G5 worth the same sort of money now?"
Great question. I need to upgrade my Mac, but instead of doing so, I will be buying a cheapy Dell box and installing FreeBSD.
Bingo!
That said, I'd be giddy as a schoolgirl if this makes porting *games* to the platform easier, and facilitates simultaneous publishing by 2007.
Issues in the transition...
I didn't hear the announcement. But phasing out the PowerPC "*from* 2008 [emphasis added]" does not give PowerPC Macs "a 3-year lifespan, maximum". I think it would be a minimum.
PANIC
Well, while there is some truth to "Panic #1", there is also some falsehoods there as well. Linux runs very well and is very stable on the Intel architecture. In other words, it's all Microsoft's fault.
Also, the original x86 architecture was a mess compared to the very orthogonal 68000 architecture, my understanding is that current processors have a pretty clean underlying RISC architecture with an ugly emulation layer on top. Given that Apple has no need to support the ugliness of the past, wouldn't it be possible to make the OS and compiler use the cleaner later instruction set? Is there a true expert out there who can answer that?
Bingo!
It will only make porting games easier if Microsoft decides to port DirectX (unlikely.)
PANIC
Okay, you had a point at the beginning of #1, but this:
Intel has been able to keep it going simply by the shear volume of money it made selling unreliable, cheap buggy, chips. You only have the right t attribute half your Windows crashes to Microsoft the other half are all Intels fault.
well, this is just hooey.
First of all, OS stability is the same for Windows and the various Unices on Intel chips as on AMD. Secondly, 85% or more of catastrophic system crashes in Windows are due to third party hardware drivers. The Windows kernel itself is quite stable at this point and is mostly unaffected by application crashes as well. This is no different from Linux: when I load the nVidia binary kernel module and run X11 with their driver, my machine instantly goes from rock solid to moderaterly crashy.
Yes, Intel has put an enormous amount of engineering into keeping x86 alive, and that is indeed because that design sucks. Not that it is mattering that much these days, because all the architectures are closing in on one another anyway; large on-chip memory (and I’m not talking just about caches) has brought them closer together in efficiency. But calling Intel chips cheap, or even unreliable, or in fact buggy, well, that’s just laughable. A primitive combustion engine design that achieves power through enormous cubic capacity might rightly be called inefficient, but that doesn’t make it unreliable or cheap.
You were spouting nonsense in the recent Java licensing discussion too. Please note that this is not supposed to be Slashdot.
PANIC
No, that’s not possible, unless you go for something with an explicit emulation layer of some sort, like Intel’s Itanium architecture or the Transmeta chips. Neither has enjoyed a lot of market success.
Will There Be an Apple?
In a year from now, there will be Macs for sale with Intel inside.
In a year from now, after trying to peddle a Macintosh hardware platform they've already abandoned, will there be an Apple?
Microsoft will ship Longhorn before Apple is shipping a full-line of non-abandoned hardware!
Personally, I think Apple just dropped out of the computer business.
Will There Be an Apple?
Bah!
I think I have all the Macs I need for a while, so I'm probably not the right focus group for this, but...
At this point, I'm having a hard time seeing a substantial difference between buying a G3 iBook when the G4 was announced, and buying a G5 PowerMac when the Px has been announced. If you want a computer now, you evaluate pricing and availability now, and you buy the best fit for your current and anticipated future needs. Why wouldn't a current Mac be that machine? It's not going to suddenly be worthless next year. It won't stop functioning. G3 machines run Tiger even now.
Software shouldn't be a problem, either, as long as developers follow a Universal Binaries model (Xcode or otherwise). If you buy software now and never upgrade, the change won't matter. If you upgrade software on a regular basis anyway, future versions will automatically become compatible with the new systems (maintaining your flexibility to switch hardware in the future). If you buy software three years from now, there's little doubt that developers will still be selling Universal Binaries. Many developers are still selling OS 9 compatible software, long after the supported life of OS 9 compatible hardware expired. It's supposed to be even *easier* for developers to keep PPC support in their code going forward.
The place where Apple may really take a hit (at least from *rational* purchasers) is in the periods leading up to actual availability. That may mean a weak quarter, not a weak year. With the staggered rollout process, each product line will take this hit independently. Besides, again, what's different between this and the anticipated upgrade for any other massive speed boost? When people thought 3 GHz G5s were coming out in the next PowerMac upgrade, they held off on purchasing until the new line was announced.
Finally, with the iPod product line going strong (perceptible stagnation notwithstanding) and the we-really-hope-its-real media hub or some other non-Mac product in the pipeline, Apple will have plenty of other sources of income to span over quarters with diminished Mac sales.
So let's not get carried away with the end-of-Apple speculation. The future is just not that grim.
Will There Be an Apple?
Those who bought G3s after G4s became available did so with some confidence that any software they bought for the machine would work even better when they later moved up to a newer machine. I have no such confidence this time. Based on reports I've read, you can forget about running anything that uses Altivec or requires a G4 or G5 on the new Pentium-class machines. If it's tuned for performance on a Macintosh, it won't run at all on a Macintel.
Usually, the delay between announcement of a new product and availability is a few months, at most. Even at that, Apple often suffers when they have trouble getting new products into the retail pipeline. This is the first time I know of when the delay between announcement and delivery is going to be up to 2-1/2 years.
I've been wrong many times before. This doesn't feel like one of those times.
Apple loyalty
Actually, I did.
...well, ok, not quite the hugging part, but knowing that the Macs were running on a good and interesting processor architecture just added that final spark of sexyness to the machines, lack of speed be damned.
Did you notice?
The speed of the presented applications on the iMac "Intel-inside", during the WWDC keynote speech of Steve Jobs, was impressive.
Issues in the transition...
- there are people STILL running OS 9!!
- there are people STILL running OS X 10.1!!!
- there are people STILL running OLD MACS
- the Quadra 8/900 are extremely sturdy machines!
so, depends on what you define as "life span" and what you want to use your computer for - a powerpc could be good for a longer time.
plus, you don't know what's happening on the linux-ppc side either.
Apple loyalty
I'm the opposite: I think, hardware does play a big role. And I want Apple to be FAST and AFFORDABLE. Sure, the user experience is what we pay the extra$$ for - but the hardware should be good.
So, if Apple manages to do that using PowerPC, fine; for some reason though, they don't, and one can't blame IBM for dropping the somewhat hysterical home PC market.
So now Apple switches to Intel? Fine with me. I most definitely want one of those sleek fast new Macintosh 64-bit Intel multiple-core multiple-processor machines the moment they're out and reliable :-) :-) I don't care how we gonna cool them as long as there's ThermalTake and Papst out there - so, get going, dogs!
At the Crossroads
You've often read messages that begin, "I was the owner of the first Mac, or Apple, or whatever, been a loyal Apple fan since 1983, etc., etc." OK, here's another.
I'm just a computer user, and stayed studiously away from 'though peripherally aware of the expert argument about the internal architecture. Here is what I can tell you from up on topside of the deck. Macs are slow. Apple's lost it, and from what I can tell, are going to further lose. I have been waiting for a new PowerMac, even if unaffordable, that would put me at the head of the class again, but now I see that that is not to come for a couple of years, and then there will be a big shake-down cruise while all the developers reconfigure for Intel and re-sell me "my" programs.
My company just gave me the lightest little Dell laptop and it beats anything I have played with in The Apple Store, even at the very top end, for opening and working applications in the business productivity arena where I have to wrestle with increasingly complex multi-format documents and avalanches of e-mail every day.
Yes, the Windows interfaces are clunky and full of gimcracks I don't need, and, as someone else said recently of XT, "DOS is still in there somewhere; I KNOW it!" But, face it, it is getting better, and Mac OS is not. I am digging around in Tiger for something more than Spotlight and Widgets, but only finding that the code being dragged around seems heavier and heavier.
Sorry, the tradeoff is just too close now, and this Mac owner, who was a bellwhether in the early days-- among the first customers at New York's first Apple dealer-- is right at the crossroads now.
Let me tell you one more thing: This vaunted creativity of Steve Jobs. I am sorry, but the iPod had to saturate its market sooner or later, and whatever new gadget he comes up with will do the same. And I do not detect a major shift to Macs as a result. Steve lost his edge long ago, IMHO, soon after he properly and insightfully "appropriated" the desktop idiom and GUI from Xerox Park or Parc or whatever it was, but he has failed to bring that idiom to the next step, which I hereby reveal to both the Apple and Windows worlds: the personal secretary. Being a few years too young and growing up in the Bay area, where they didn't have personal secretaries, it was probably fated. My boss back East did, and I can tell you that it was one hell of a productivity boost for him to be able to say to his secretary, "Write a letter to so-and-so and tell him that we're still awaiting that contract," and have it back on his desk, typed perfectly on letterhead, even after three drafts on a Selectric oerhaps, ready for his signature. Beats e-mail, because he didn't type or even think much of it himself. That is the way the personal computer should have and still could go-- toward a real robotic, personal secretary interface, perfectly responsive to audio commands, interactive and subservient. Man, what I would give to have that capability. Apple missed a big boat.
That's all. Tear me a new one if you like.