Slashdot Mirror
Dvorak Thinks Apple Will Switch to Intel
SeanTobin was among several users who noted that Dvorak's latest column discusses the possibility of Apple going to Intel for future macs. Yeah, this rumor pops up pretty often, but I wonder how long before we'd get binary compatibility between other x86 unix OSs.
MacOS on Intel platform opens up lots of interesting "what ifs." Would you be able to order your Dell with XP or MacOS X? The real question then becomes, what would happen to MacOS support (i.e. MS Office for MacOS) from Microsoft once Apple and MS were competing on the same hardware platform.
Since OS X runs on a BSD base, would MS change its tune regarding Linux?
Could be an interesting time!
Dvorak needs to say these things every now and then to maintain his position as public enemy #2 (after Bill) to Mac users. As I recall, he had some interesting predictions on those newfangled "mouse" thingies as well.
people, this is the same idiot who claimed the mouse would never catch on... the same pin-headed moron who said Apple would never last and he predicted Apple's death for about a decade or so...
Apple will go with Intel when Osama bin Laden converts to Judaism.
This sounds like 1988... yes, 15 years ago. That's when the first "Apple will switch to Intel" story came out in MacWorld.
Of course it could happen. But is it likely? I don't think so. Apple's manrta is "being differnet". Using the standard PC processor out there isn't that different.
Of course, there could be a huge advantage to Intel if the Itanium CPU is picked up by Apple. Apple sells a lot of machines... although perhaps only 2% of the new machines are Apple, that's still a lot of hardware. ANd the fact that lots of folks look at Apple as a premium brand (unlike lots of the crap you can buy at Circuit City), Intel would have a winner.
Here is how the Inquirer reported the story on friday: April fools day comes early?
Whenever the offence inspires less horror than the punishment, the rigour of penal law is obliged to give way...
1) Jobs' ego. Jobs has said on the record that he'll resign before he builds an Apple box with an Intel chip. (I honestly don't remember where that rivalry originates.)
2) Developer opinion. Dvorak is primarly a PC man -- I think he missed much of the outcry that occurred when we switched from 68K to PPC. For that matter, there's still bits of Carbon that date back to 68K, such as setting and unsetting the A5 world register for callback routines. Also consider that the killer apps of the Mac world (Adobe products, Quark, etc) are just now becoming native to OS X. The outcry if we had to switch to a new OS would be massive. There's also the fact that the PPC ISA is backwards compatible with the 68K -- all existing apps for Apple would have to be emulated. Can you say "fuck no," children?
3) Architecture differences. True, you can recompile the Darwin microkernel for Intel. There's a lot of differences though in the hardware -- for example, Macs directly work with the INT# lines on the PCI bus, they don't have IRQs. It would be incredibly costly for Apple to eschew the current standards in PC motherboard design and make their own chipset.
4) IBM. The PowerPC architecture is not slow in and of itself -- it's just a spec for a RISC instruction set. The problem lies in Motorola, who no longer relies on Apple for business now that their wireless division supports the company, and who has been dragging their heels on their PPC line. IBM's new PowerPC 970 is a desktop version of their Power5 server processor (including its unusual pipeline design) planned to debut at 1.8GHz on a 0.13 micron process. Yum.
There's also the point that Dvorak is known as a rumor-spouting gasbag... and one who has a chip on his shoulder for Apple. The guy used to write for MacWorld until he had a falling out with Apple management, and has become notorious for his anti-Apple bias ever since.
Yea, you're right. Very misleading is the slashdot posting.
Anyways, Dvorak is silly. He knows he can write this kind of stuff and it doesn't matter if it comes true or not.
The Itanium, like you said, is just too expensive, and will remain so for a long time. Unless Apple wants to release a $19,000 workstation that nobody will buy and runs no old mac software (or runs old mac stuff slowly), I don't think this will happen anytime soon.
He says that they have been able to transition CPU architectures flawlessly in the past, which is true, but technology today it's a different story. The PPC was a LOT faster then the 68k processors, so emulation was pretty quick. Today, I'd like to see an Itanium emulate a modern PPC chip with any good speed. Yea, right.
The thing I found funny was when he writes "This new workstation will be optimized for Photoshop." Okay, since when does the mac user only use Photoshop?
I'm not a mac user, and I'm not even a mac fan. But even I can see the obvious flaws to his article.
- It's not the Macs I hate. It's Digg users. -
I'm no mac user, but I can't see the benefits of using an Intel chip in a mac.
Personally, I'll take my high clocked hot as hell Athlon over a Mac anyday.
The x86 chips are fast now a days. The gap has gotten a lot smaller in the performance per Mhz between them and Apple's PPC chips. And, via brute-force (lots more Mhz) they outperform the PPC chips.
However, this doesn't seem like a forward step IMO.
Plus, Dvorak said "Intel" and in this case he meant "Itanium." Was my misunderstanding. Either way it's a load.
You know, I bet if Apple DID release MacOS for the x86 chips, it would do pretty well. I'd run it (for some stuff) and I bet it would get a lot of support. Unfortunately, it would probably also mean the demise of their hardware line, and it would be a lot of work. Of course, if it failed, Apple would probably go out of business.
- It's not the Macs I hate. It's Digg users. -
I think the chance of Itanium being used on the desktop is nil. Itanium is too expensive, and isn't that fast, nor does it have the really attractive property of other Intel CPU's: volume. No one's buying Itanium, thus Itanium isn't being manufactured in volume, thus Itanium isn't cheap. And Apple doesn't have the sales figures to bring Itanium's volume up.
Apple would be much better off going with AMD's Hammer, or IBM's upcoming PowerPC 970 chips, or even a P4/Athlon (not likely... I can see Jobs craving 64-bitness). I'd personnally choose Hammer because AMD is going to produce, and probably sell in volumes similar to the Athlon. Although, the 970 looks mighty tasty...
Lex orandi, lex credendi.
There are millions upon millions of gameboys worldwide that have used the z80 processor. Nintendo used them up until 1998 (with an 8MHz z80-esque processor made by sharp). There was still a market for them, possibly a larger market then the 8080 processors. The z80 passed test of time, while the 8080 just disappeared into oblivion.
heres a website with a lot of info on the z80
As a Linux switcher to OS X (and for the folks who are making their Linux apps compilable on OS X - thanks!), I can see the good and bad of this. Right now, my main desktop/laptop are Mac boxes, but I still have another 3 machines in the house still running Windows - my wife's work machine (soon to be retired after we move and she doesn't have to work), my Game box (because Raven Shield probably doesn't run under Wine), and my Linux server for my web server/mail server/etc.
I would love to switch my Linux box to a Mac OS X server box, mainly just to play around with another OS I haven't tried yet, and because I think it will be easier to maintain. For $1000 for an unlimited server licence, I could deal with that. (Yes, I could just go 10 for $500, but I'm evil that way.) The problem is that even finding an old Mac (like a G4 cube) is around $1000 even on Ebay.
The good part about OS X on Intel is that the machines would be damn cheap. I could probably take my current Windows P-800 machine and turn it into a decent - not great, but decent OS X server box. Cheap boxes with a great OS would be the true "Microsoft desktop killer" we've been waiting for. The operating system (OS X) is tried and true now, it's excellent, stable, and kicks Microsoft's ass all over the place.
But the problem in moving to the Intel platform is threefold:
1. Performance. Going from the PPC with all of its registers to Intel's platform will cost some performance - especially if some sort of PPC emulator is used to make all of the old apps run.
2. Drivers. Right now, Apple can ensure that every video card that's qualified to run on a Mac will run, and run without a problem. I've stuck all sorts of hardware into my Mac so far, and it all works flawlessly. Apple will lose that ability.
3. Cheap hardware. Yes, Apple's hardware costs more. And it depreciates a hell of a lot slower than just about any other PCs out there - look at my Cube situation again. These machines are like a rock - they run and run and run. A Compusa employee I once knew mentionted that he hated it when people came into his store and bought a Mac - because he never saw them again, while the Windows guys were in every few months because they "had" to upgrade. So you'll have hardware that won't last as long or as well.
Unless of course, Apple basically brands their own Intel based PCs and ensures that OS X only runs on "certified" machines. Remember - they make money from hardware, not software (though, if OS X became popular and runs on all Intel systems, they could become the next Microsoft in many ways, only with a decent desktop and server system).
I honestly don't believe that it's going to happen. Apple will most likely shift to IBM's new Power970 line - it's the most like the PPC, so no translators/emulators. IBM has a vested interest in making these chips fast and plentiful for their server systems, unlike Motorola which is making PPC chips for - well, pretty much Apple.
Anyway, there's my $0.02. And of course, I could be wrong.
52 Weeks, 52 Religions with John Hummel
- ALl they would hvae to do is port darwin, and recompile the rest. That's not the momentous task you make it out to be... in fact, it's one of the reasons for using a microkernel in the first place.
- Developers would NOT be writing for a "new OS". They would be compiling for a different target architecture.... not the same thing at all. Look at linux, and all the apps that work in ppc, linux, alpha, etctera.
- The PPC was not 68k compatable.. they had to emulate the 68k completely.
For one thing, they have already started making the motherboards.
-You may license this sig for only $6.99.
The concept of an itanium notebook, considering their current power usage (current power... geddit?) is almost laughable.
Powerbooks are a huge part of Apple's market now, and even if their desktops fared the equal of a PC in pure grunt, would still be a major source of revenue.
Pick an option - Itanium all through the line including powerbooks, or PPC Powerbooks and Itanium desktops. It doesn't seem likely to me.
Then again, Itanium XServes doesn't sound quite as far-fetched.
Dvorak didn't say Apple would switch to Pentium.
First Dvorak says Apple wouldn't have to rationalize lower closkspeed (MHz myth) anymore if it switched to Intel. Then he suggests Apple will choose Itanium? Last time I checked Itanium 2 was no faster than a G4 in clock speed. Dvorak should inform himself about Intel's products if he wants to be credible.
Apple-OS (Any flavor) rocks because apple knows what they will be working with down to the last transistor - Other-OS is generally made to work on a wide range of machines - This is why windows 95 was so awful - bad device support
Apple on i386 = not so good
Over time, users would face the choice between going back to Windows, or buying a Mac, if they wanted to continue to get support. The Intel OS X itself would become steadily outdated and users reliant on third party hacks to support more modern hardware, so just users just sticking to it wouldn't necessarily be an option.
It would be a massive jump for Apple and they'd have to consider it a long term project. Perhaps they'd even, to make the thing profitable, have to sell each copy at a margin approaching their low end hardware margins, or alternatively ship the OS as a "Demo version" which would expire after a period of time.
I wonder if they could get the concept to work?
You are not alone. This is not normal. None of this is normal.
So don't be surprised if there's a follow-up to this saying as such.
This is my digital signature. 10011011001
Itanium expensive yes. Itanium not that fast, no. I've benchmarked them and am in the process of buying 3 of em. And when the 1.5GHz CPUs come out later this year they will be the fastest procs in their class, period.
I doubt Apple will go to any x86 varient because that will turn them into a software company and kill their business, unless they think that they will be able to make up the loss of profits off of hardware by increased OS sales. Apple's software is basically thier OS, and even M$ with their stronghold on the PC (and desktop in general) market makes mucho dinero off of apps like Office. I don't think that ppl want another PC OS, remember BeOS?
On a side note, its pretty much common knowledge that Dvorak is a moron, and articles here based on something that he says is usually flamebait from the getgo.
OK, when you say "I'd run it" was that "I'd run it legally by going to buy a box on a shelf at $129.95" or "I'd run it if I could get somebody to give me a copy, or if I could download one from Kazaa."?
Regardless of the way you'd answer, a lot of people would answer the second way. So what's the benefit for Apple?
Why yes, I AM a rocket scientist!
This couldn't work...
The reason mac stuff is so nice is that mac knows the hardware that goes into it. Apple controls the computers that run OS X so it always runs flawlessly. Intel hardware is too unpredictable for this type of thing.
Also, most people consider an OS a "house", they use it and live in it for years until they have a reason to move. it's an immobile part of their computer. People like you and me switch OS's like socks, but most ppl use the same OS, same INSTALLATION of the os, for years.
my parents have been using the same install of win98 on their pII 450 since...well..98.
in short, nobody will go for a demo OS unless they're a hacker, and nobody will like MacOS if it runs on intel stuff.
Typed from a G3 800 ibook.
You like your new Mac more than you like me, don't you, Dave? Dave? I asked...She said Yes.
Apple will undoubtedly choose to go with whichever processor will provide their users with the best balance of price/performance. The fact that they have stuck with PPC for so long is merely a coincidence.
Much like SGI using MIPS processors (although they did end up buying the company), and their official stance being whatever provides the best performance (be it Intel, MIPS, Sparc, PPC) is what they will in turn use.
OS X being direct descendant of OpenStep/NextStep, is a very portable OS (If you will recall it ran on x86, PPC, and Sparc Hardware). The only reason that I can argue that Apple has not gone to x86 compatible hardware up to this point is Microsoft's influence. Microsoft will not continue with Office for the Mac if it is going to lead to people choosing the Mac OS over Windows.
There is already some talk about Microsoft dropping Office for the Mac because of "low sales figures", which is fine by me as I tend to use OppenOffice.org anyway. Frankly as soon as Sun realizes the market for OpenOffice.org on MacOS, they will start marketing it under the StarOffice name and provide support, all at a price that Microsoft can not even begin to compete with.
If Apple does choose to go to an "x86" processor, it would be more than likely an offering from AMD (in the form of a Hammer) than Intel. Any thought of the Itanium processor is merely wishful thinking on the part of Intel (remember these things cost nearly $3000 per processor, and Apple has joined the SMP revolution).
If you were to see OS X on "Intel" hardware, I would expect that you would see it in a strange combination of technologies. For example, you would likely see is special PCI card which would be the boot media (Kernel in Flash) with special system identifiers in ROM, to insure that is is a Apple authorized installation. This would be the configuration of the "Clones". The "real" Apple hardware would have these components integrated into the mainboard. The real Apple hardware would not support booting Windows (much the same as you can not boot AIX on a Mac (except for the ANS) even though it is a compatible platform on which to do so. Obvious omissions from the firmware are noted.).
Like most people so far in this thread, I agree that Dvorak's prediction is a bit absurd...but does anyone have an explanation for the seeming development of a relationship between Apple (or at least Steve Jobs) and Intel hinted at by the three things Dvorak mentioned (Intel sales conference keynote, Pixar switching to Intel, Intel executives at Macworld)?
It makes me wonder, and I haven't read any alternate theories.
You have to admit, Dvorak has a pretty sweet gig. Somehow he's figured out a way to get paid to be an uninformed, foaming nutwad. Even years on, when his predictions have turned out to be no better than (if not worse than) random guessing, he's still making it work for him.
Personally I just want to see apple apply their design skills to faster processors. I'll run linux on it anyway, so they can do whatever they want for os. Of course it's a lot more convenient to run linux on x86 hardware.
It's still beyond me why no x86 laptop manufacturer has managed to come up with a design on a par with PowerBooks.
I think we can make a few assumptions...
-Apple will NOT switch to a chip that's not 64-bit. That's simply not an option. The costs of switching to a new platform will not be justifiable if they have to switch again in a few years.
-Apple will not abandon PowerPC until IBM's PowerPC 970 sinks or swims. It's a very mutually beneficial relationship, and while it may not keep up with x86's power, it won't be that far behind, and it *will* fit in the form factors that Apple needs. The really fast x86's put out way too much heat.
It'll be 5 or more years before they switch processor architechtures, maybe even longer, maybe never. x86 does not offer sufficient advantages to put up with the heat of the fast x86 processors. Apple is very strong with laptops, and they're only going to get stronger there. Even their desktop offerings are compact. Small is important, quiet is important, batteries are important, and x86 can not beat PPC in with these.
When someone might yell at me, it has to be OpenBSD.
Also, he mis-understands Marklar. Apparently, this is a complete x86 Intel port of OS X. It acheives very little in targeting Itanium as a processor, as x86 is as much another slow emulation like PPC.
"Flyin' in just a sweet place,
Never been known to fail..."
Since when has apple cared if their stuff is expensive?
He is truly an idiot. The move towards Intel is the move towards commodity pricing. No future there for Apple....or Sun for that matter. Apple should merge with Sun and Sparc. Sun would rise to the occasion and produce a cheap Sparc chip to power Mac's. It's the only hope for Sun and maybe the only hope for Apple. Despite Apple's client side glitz their server story is weak in the extreme. But the two companies' synergies are enormous. Apple would add immense value to Sun since they understand the client side so well and Sun has repeatedly demonstrated it is clueless in this regard. Sun would add immense value to Apple since Sun's whole business is the server. Neither Sun nor Apple have a future without proprietary hardware, because you can't download hardware. but you certainly can download software...
Dvorak has time and time again shot massive holes through his credibility when it comes to the topic of Macintosh and Apple. I'm surprised he's not so thoroughly embarrassed by this point to avoid the subject completely. It's likely he only put this column up to kick up hits to his column which is precisely why I'm not going to it.
--Rick "If it isn't broken, take it apart and find out why."
Itanium's EPIC architecture fits neatly into Apple's image of always using superior technology.
Claus
Well, I don't know about his track record for predictions, as I don't read him that often, but the idea doesn't seem so far-fetched to me. Reading through some of the objections to it here makes his article seem well thought out by comparison.
BSD is a micro-kernel operating system. In theory even easier to port to other platforms than Linux. What advantage is there in remaining tied to a single hardware architecture? Don't get me wrong, I'm a big fan of Power-PC over Intel, but to the end user it really shouldn't make any difference. Very few people (including audiophiles) evaluate new stereo components based on the details of the integrated circuits used internally. The commodity PC will be that way too. Users will expect a certain number of basic features and the rest will be a matter of price, performance, and case color. Apple understands this a lot better than anyone else right now apparently. Their biggest weakness is price, not performance.
Don't forget they are selling servers now too. I doubt their goal is to lose gobs of money at this. They need to be flexible in terms of running on towers, blades, low power Transmetas and whatever other variations come out next year. To be perpetually tied to the fortunes of the Power-PC processor would be insane for them.
While I think OS X is far from being perfect, Apple is the the cat-bird seat with respect to Microsoft, who is just stuck selling software. With a fairly clean (by comparison with Windows) operating system that can be tailored to a number of uses, and an established reputation as a designer of hardware they not only compete with Microsoft, but with Sun and IBM as well. If, unlike these two, they start out with their server side software offerings as multiplatform compatible they'll have a big lead over everyone (except Linux) in terms of an OS that runs as comfortably on a desktop as it does in the computer room.
While there may be some technical challenges as pointed out here in Slashdot, I doubt that any of them are insurmountable. The more important question is: Is this a good idea from the Apple business perspective. I think the answer is a resounding YES!
Agreed. Although if he had said Opteron instead of Itantium, I might have half belived it.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
All Windows versions since Windows 2000 have run on Itanium.
Back in 1983 I was working for Convergent Technologies, a company that originally specialized in Intel-based workstations. (I think they may even have coined the word.) Dvorak reported a rumor that we we're working on a Motorola 68000-based portable computer. He discounted the rumor, since everybody knew that we only did Intel boxes.
Dvorak was wrong in two different ways. First he or his source combined two different rumors. There was a portable computer, but it was based on a Hitachi 6303. There was a 68000-based computer, but that was a completely separate project.
Which I was hired to help document. The MegaFrame actually used both 68000 and 80186 processors in its Unix config. (It could also be configured as a workstation server using only 80186 processors.) So in fact we were not only not committed to 80x86 architecutre, we were into two other architectures.
(The 6303 was also a Motorola architecture, being based on the Motorola 6800. But that's completely different from the 68000, because Motorola decided to make a clean break when then went from 8-bit to 16-bit processors. Unlike Intel, which made the 8086 vaguely backward compatible with the 8080. Which is part of the reason Intel's chips are standard and Motorola's are dead. But I digress.)
Dvorak's other error seemd particularly stupid: the assumption that all programmers targeted specific CPUs. Which might have actually been true in the homebrew micro culture he came from, but was never true of programming in general.
Actually, Dvorak might be a very smart guy, behind all the stupid stuff he keeps saying. A lot of computer pundits are people who have some Big Insight that's either completely bogus or only valid in a certain context. They hold onto these ideas for years, against all logic. I guess they'd lose too much face by admitting they're wrong.
One example is Vernor Vinge, who used to be one of my favorite SF writers. But now he considers himself a computer expert, based on a lot of second hand knowledge, and some practical experience with things like client-server computing. The way his pseudo-knowledge dominates his stories completely destroys my ability to enjoy his work. Which is a shame -- in many ways he's grown a lot as a writer.
Another example is Neal Stephenson, who's still one of my favorites, despite all the non-sequiturs in books like Snowcrash. (Come on people, do you really think that you can design seriou VR in machine language???!!!) The Big Idea that really drives me crazy is Stephenson's belief that a Turing Machine is something you can actually build. (Neither Radio Shack nor CDW stock infinite-length tapes. I'll apologize if anybody can point me to a source.) So far, his work is original and creative enough to make me overlook crap like this. But give him time!