The NS32016 was an advanced CISC microprocessor of the same generation as the Z8000 and 68000. It (and its successor the NS32032) had a reputation of being a very fast chip compared to the Zilog and Motorola efforts.
If Apple were to architect their own system controller chip for x86, as they already have done countless times for PowerPC, then they could lock out clones by simply not publishing the drivers for the ASIC in the Darwin source tree, or by refusing to license the ASIC to third party manufacturers.
That's solving the wrong problem, though.
Apple could relatively easily keep you from installing other operating systems on the Apple hardware that way, sure. But what would it do to keep you from installing Darwin components in an OS X install and bootstrapping OS X on a clone, the way XPostFacto does now?
Really, the one place I can see that Apple could effectively lock you out is in the video drivers. To get the video performance Mac OS X requires you need the proprietary Apple/ATI kext for the card. If that only works with Apple-ROMmed video cards, and Apple-ROMmed video cards only work on real apple hardware, you'd be stuck with a generic display driver and software OpenGL.
Give me a few and I'm pretty sure I could find others.
Please do. I'm honestly interested, I've been looking for one for a couple of years now.
I'm not an Apple apologist, nor would I ever say you're perfectly safe just because they haven't been exploited yet, or even because the system design is inherently safer than the Windows desktop. But the fact remains that is a better design (it could hardly be worse), and they haven't been subject to spyware or viruses... I think Apple has taken some unfortunate actions with Safari that have the potential of eventually changing the situation for the worse. But right now there are no cases of live viruses or spyware for OS X, and I can't see a credible avenue of attack that would be more effective than plain old social engineering.
Social engineering is a problem, but it's a managable one. You can learn not to be "phished"*.
Turn off "open safe files" in Safari (or, better, use Camino** instead) and don't be phishable: don't open email attachments unless you are expecting them, don't download and run random garbage, and don't trust URLs you get in email. You'll be as close to safe as you can be on any platform, and safer than you would be running Windows even if you run a firewall, a sniffer, antivirus, antispyware, and don't run ANY Microsoft-provided internet or Mail apps.
* I know that this term originated with people stealing passwords, credit card numbers, and other personal info... but the same kinds of social engineering attacks are used by viruses.
** I trust Camino more than Firefox. Mozilla's pluggable extension mechanism is better than Microsoft's ActiveX, but it's still an unnecessary risk. Once Flashblock came to Camino I quit using Firefox.
Sophos seems to think there's a few and I'm sure a quick google will find something more;)
These are typical examples of the knd of things I'm talking about.
Cowhand is not a virus or spyware, it's a rootkit component to be installed after you've already used an exploit (virus, direct attack, or social engineering) to get in. If this is an "OS X virus" then so is tinyproxy or socks. Strike one.
Amphimix is a demonstration exploit that can only be used through a social engineering attack. A social engineering attack can't be completely defended against by the OS, but Apple has taken steps to mitigate this one (less well than I'd like, but probably adequately) and there are zero examples of it in the wild. Strike two.
Renepo is a payload only, it's got no remote propogation mechanism... Sophus notes "that any attacker trying to plant this worm in your network would need to get root access on one of your boxes first, meaning that you would already be "owned". Strike Three.
Sorry, Casey.
Of course nothing is as bad as the numbers for Windows but to say that your completely safe is pure arrogance.
I didn't say that you're completely safe. I said that there are zero exploits currently out there for OS X. Of course "past performance is not proof of future gains", but the level of risk is so low that you're more likely to suffer data loss or system damage from running A/V software than not running it.
If you're running a webserver or some other "attractive nuisance", then there's a whole different level of risk and a whole different approach necessary. But for a desktop user? I'm not sure I'd even recommend turning on Apple's firewall unless you're on a shared LAN... and even then I'd want to check the settings before I treated it as anything but a placebo.
Depends. Are you asking "if Apple announces the switch" or "if Apple successfully makes the switch"?
It's a big risk, with the potential of crippling the company. That's more than enough reason to object now. If you had a newspaper from 2010 that showed how Apple navigated the waters, that'd be a different question... but I don't think that's the alternate universe we're in right now.
Secondly, what does "making the switch to x86" mean? If it just means running OS X on non-x86-based Apple-branded hardware, I honestly don't see the point in Apple doing this. I don't see any upside for Apple, and I see a huge downside for both Apple and the current Mac user base. If it means Apple's going head-to-head with Microsoft, so I can buy "OS X Intel" for a few hundred bucks retail and install it on non-Apple hardware? That's a whole different kettle of Dogcows.
So... If it meant I could run Mac OS X on a Thinkpad, even if I paid twice as much for OS X for the privilege of running it on a non-Apple computer, then I'd be (tentatively) for it. If it just means I still have to buy a Powerbook... just an x86-based one, then what's the upside? Your x86 Mac would be slower for a long time (even on a faster processor) until all the apps are ported and then re-optimised, it'd be less secure because buffer overrun attacks are easier with the x86 instruction set (purely by chance, there's fewer NUL bytes in x86 code than most RISC processors, and NUL bytes are a string terminator in C), and there's no particular advantage in the forseeable future.
I mean if you know the future is x86, why would you get a PPC mac?
If this rumor had been around a couple of months ago, I wouldn't have bought a mini... I'd have stuck with my old "Beige G4". If I knew for certain? I sure wouldn't buy a new PPC Mac. If they do this, they better have the new models ready across the product line on day one, or a guaranteed affordable upgrade (substantially complete buyback, or a motherboard or CPU+ROM replacement) for all current models.
Either that or some kind of solid and credible commitment to OS upgrades for the current platform over the projected lifetime of the machines currently in service. The oldest machines supported on Tiger date to January 1999, so that means OS upgrades for 2005 models through 2010.
Well, I do. There are zero viruses in the world for OS X.
I seem to remember there being one
You remember wrong. There was a port of a UNIX rootkit floating around, and a couple of theoretical exploits based on holes Apple patched (albeit clumsily) quickly. A rootkit is software that's used after one's already broken into the computer... it's useless without a way to get it into the computer in the first place.
What you remember is probably that Symantec claimed that there WOULD be viruses and spyware on OS X, with no actual evidence. They've made the same claims before for Pocket PC and Palm OS when they were trying to push their antivirus solutions for those. Nothing has EVER happened to people who ignored any of these attempts by Symantec to use FUD to push their product. But... people have had their systems damaged by antivirus software on all these platforms.
I don't think that the reason there is no (or very little) Mac spyware/adware/malware is really only due to installed base.
Of course it isn't. It's because of the way IE is integrated into the desktop the HTML control has to guess whether a document it's opening is trusted or not, so if you can trick it into thinking your spyware is a local file you can get local user access. Microsoft tries to pop up a warning dialog when you're in a gray area, but there's so much gray area that people get used to clicking "OK" automatically.
Of course Apple has started to open up the same opportunities, first by having Safari "open safe files after downloading", then by treating Widgets and software installers as "safe files", and finally by using a dialog instead of backing out the bad behaviour. It's nowhere near as bad as it is in Windows, but it's worrying.
Can someone explain the difference between hacking XP Home to turn it into XP Pro and downloading a warez copy?
The former only became illegal when the DMCA was passed. The DMCA makes a lot of previously legal and still necessary actions illegal, so the fact that it's illegal under the DMCA is by itself irrelevant to the morality of the act. So it comes down to the morality of boosting the performance of a factory-crippled product. You can buy products for doing that at any auto-parts store.
The latter? XP Pro includes software that isn't included in XP Home. Even if you bought a copy of XP Home you're not entitled to that. But if Microsoft sold you more than they claimed they did, and you can turn that extra software on?
Violating a EULA may be illegal, but I'm not going to call it unethical.
Hmmm. The ARM, as I recall, started out as a processor for the Acorn line of personal computers in England. It was pretty impressive as a PC, but workstation-class?
What do you think the average Windows user needs to know other than how to download, install, and use Firefox, AVG, Ad-aware, Spybot, Gaim, and firewall?
That's still a lot (why Firefox rather than Netscape, for example? I know the answer, do you?), and it's not enough. You forgot about a non-Microsoft mail program and a non-Microsoft media player. Oh, and some commercial game software ON CD has spyware in it.
They need to know how to configure a firewall, how to tell when a message from the firewall is a problem or a false alarm (YOUR DNS SERVER IS HAXORING ME), when NOT to trust Firefox, when a message from antivirus or antispyware is a false alarm (TEH REGISTRY IS SPYING ON ME)...
I've had to deal with failures in all of the above, and I'm not even supposed to be Windows support any more!
And that's ignoring DLL conflicts (DLL HELL is not dead), driver conflicts (not bad hardware, hardware that's perfectly good, it just doesn't work with other hardware), application conflicts, deliberately badly-behaved applications (to keep Acrobat Reader from installing itself as the handler for PDF files in the browser, you have to go in and remove a file from the Acrobat directory), antivirus software conflicting with Microsoft's system checkpoints, and half-assed troubleshooting from well-intentioned friends.
i don't seem all that much sucess running xbox code on other x86 boxes
Emulating an operating system is a much harder problem than patching an operating system to run on different hardware. An operating system is a much more complicated beast than any hardware.
In this case the hypothetical attacker would already have the operating system, as well as pretty complete documentation of everything that runs above it. All that would be needed is to replace the OS components with ones that don't do the same checks.
And on the Xbox there's the additional problem of emulating the GPU, because you're trying to run games that expect to have a custom GPU, whereas on the Mac everything's mediated by OpenGL.
you have the source to darwin but not to things like apples grapics system
I've already mentioned elsewhere that accelerated graphics is the most credible place for Apple to embed a lock, but you could probably still get unaccelerated graphics working: this has already been done for the earliest Powermacs and clones, as well as emulation under MOL (Mac on Linux) and PearPC.
Itanium would at least give Apple a shot at outperforming Windows PC's.
Itanium's got a lot of registers, but it's a bloody awful emulation host because the best possible emulation technique, recompilation, is going to be agonizingly difficult. The Itanium instruction set architecture is so complex that the Intel/HP supercompiler does cross-module optimization... it actually moves code from one object module to another... and analyzes trace logs from actual code runs to get the kind of performance that makes the Itanium worthwhile. It's the only compiler I've used in 20 years where you want to turn off optimizations just because the compiler itself is significantly slowed down when it's using them.
even if you use more or less standard pc parts its still pretty easy to change enough to break stuff look at the xbox for example
Yeh, it's a perfect example. All you need to do to unbreak the XBox is run some code... any code... on it. Once you can do that you're home free. The hard part about cracking the XBox is tricking it into running some code for you. A general-purpose operating system by definition does not have that protection, especially when you're trying to break it the other way: you're booting the OS into an environment you control.
and unless you have source to the os its bloody hard to deal with such breakages
I know that I would gladly fork out the $400 for OSX to run it on my laptop.
I don't think they could do it at $400, that's almost as much as a Mac mini and that includes the hardware as well. No, Apple would have to sell OS X for the current price ($129) or less to get the kind of market that would make it worthwhile.
But OS X on a Thinkpad? A laptop Mac without the horrible 'book hardware? I could deal with that.
Make a proprietary motherboard architecture just like they have now with PowerPC based boards, except with an x86 CPU in place of the Gx chip.
The only thing on an Apple Motherboard that isn't regular PC hardware is the PPC CPU. Other than the CPU there's almost nothing proprietary there. PCI, AGP, PC RAM, PC USB chips, PC firewire chips, PC everything.
Then just make it so that the OS requires an Apple mobo to run. How?
They can't do it. Darwin is already out, it's open source, it's the Mac OS X kernel, and it runs on x86 hardware.
And in any case, the performance gap between PPC and x86 is tiny, sometimes PPC even gets ahead of x86 despite having a fraction of the resources spent on it. So there's no significant performance win for Apple going to x86, and a huge performance downside because they will have to do emulation for some apps, and that will be agonizingly slow.
So the only possibile reason for going to x86 would be to leave the hardware business and go head-to-head with Microsoft on x86 clones. And they MIGHT be able to pull it off, but it'd be a big gamble and a big loss in the short term.
Only on a new Apple handheld running Newton OS. The ARM is a good low-power embedded chip, but it's nowhere near fast enough even for an entry-level Mac.
How about if they put an Intel chip in a low end computer, like the Mini, recompile the Apple-supplied applications so they run reasonably fast, and then emulate PPC for the 3rd party applications only.
Why? It's not the low end computers that need new chips. They're not even using the fastest available chip in the mini, and they're underclocking it at that.
And it would produce a system that's slower, more power-hungry, and more expensive than the mini. Running any 3rd party applications on it would be like running them on a Powermac 7600. The Beige G3, the oldest system OS X is supported on, would be faster. So to get acceptable speed they'd need to build it with a faster CPU, faster RAM, and more RAM... and it would still be slower.
And where do you get this "most buyers of the Mini wouldn't run 3rd party applications anyway" from, anyway?
but if the G4 and G5 can't hold up against the Pentium M and sequelae in the laptop game Apple will have to do something
That's up to Freescale. They're supposed to be sampling later this year. Dual-core G4 powerbooks early 2006?
I also agree with you that the x86 isn't an ideal PPC emulation platform, but where else will Apple turn?
IBM seems to have broken the 3 GHz barrier with the 970 core, so there's nothing stopping Apple from using those chips in the G5 to keep up with the numbers game. There's also dual-core G5s coming.
And unless there's something badly wrong with both the Freescale low-power chips and the IBM high-end chips, this will all come out sooner than they could possibly wield an x86 Mac with any kind of credible application base.
Will Intel risk Microsoft's wrath by luring Apple with a well-designed ISA?
I assume you're talking about Intel licensing the PPC instruction set from IBM, which is certainly a possibility... but that would render this whole discussion moot.
what's to stop Microsoft from porting Windows to that generation of Mac hardware?
Microsoft's already ported Windows to THIS generation of Mac hardware. That's what they use for their XBox 360 emulation. But they've got no reason to sell it, why would someone buy a Mac to run Windows on it unless they're an XBox developer? Changing the CPU over to an x86 wouldn't change that at all... if someone looking to buy a Mac wants to buy a Windows box they can buy one a LOT cheaper just buy buying a Windows box.
Microsoft could even entice a few Taiwanese whitebox makers to churn out Mac clones at a much smaller fraction of the cost Apple would want to sell them for.
They could do that now. You can buy a PPC motherboard right now, if you want, but the only market is the embedded systems business. And without the PPC a "Mac Compatible x86" is just another clone: you can already run the core of OS X on one today
Since Apple's margin on the iPod is more like 40% they're still making a nice profit, and a certain amount of the sales will be to people who hadn't planned on buying a new iPod but are reminded of the old 1st gen with the cracked screen in their drawer.
And at least some of the returns will be good enough to refurbish and feed back into their support and warranty system.
Slashdot Mirror
The NS32016 was an advanced CISC microprocessor of the same generation as the Z8000 and 68000. It (and its successor the NS32032) had a reputation of being a very fast chip compared to the Zilog and Motorola efforts.
If Apple were to architect their own system controller chip for x86, as they already have done countless times for PowerPC, then they could lock out clones by simply not publishing the drivers for the ASIC in the Darwin source tree, or by refusing to license the ASIC to third party manufacturers.
That's solving the wrong problem, though.
Apple could relatively easily keep you from installing other operating systems on the Apple hardware that way, sure. But what would it do to keep you from installing Darwin components in an OS X install and bootstrapping OS X on a clone, the way XPostFacto does now?
Really, the one place I can see that Apple could effectively lock you out is in the video drivers. To get the video performance Mac OS X requires you need the proprietary Apple/ATI kext for the card. If that only works with Apple-ROMmed video cards, and Apple-ROMmed video cards only work on real apple hardware, you'd be stuck with a generic display driver and software OpenGL.
Give me a few and I'm pretty sure I could find others.
Please do. I'm honestly interested, I've been looking for one for a couple of years now.
I'm not an Apple apologist, nor would I ever say you're perfectly safe just because they haven't been exploited yet, or even because the system design is inherently safer than the Windows desktop. But the fact remains that is a better design (it could hardly be worse), and they haven't been subject to spyware or viruses... I think Apple has taken some unfortunate actions with Safari that have the potential of eventually changing the situation for the worse. But right now there are no cases of live viruses or spyware for OS X, and I can't see a credible avenue of attack that would be more effective than plain old social engineering.
Social engineering is a problem, but it's a managable one. You can learn not to be "phished"*.
Turn off "open safe files" in Safari (or, better, use Camino** instead) and don't be phishable: don't open email attachments unless you are expecting them, don't download and run random garbage, and don't trust URLs you get in email. You'll be as close to safe as you can be on any platform, and safer than you would be running Windows even if you run a firewall, a sniffer, antivirus, antispyware, and don't run ANY Microsoft-provided internet or Mail apps.
* I know that this term originated with people stealing passwords, credit card numbers, and other personal info... but the same kinds of social engineering attacks are used by viruses.
** I trust Camino more than Firefox. Mozilla's pluggable extension mechanism is better than Microsoft's ActiveX, but it's still an unnecessary risk. Once Flashblock came to Camino I quit using Firefox.
Sophos seems to think there's a few and I'm sure a quick google will find something more ;)
These are typical examples of the knd of things I'm talking about.
Cowhand is not a virus or spyware, it's a rootkit component to be installed after you've already used an exploit (virus, direct attack, or social engineering) to get in. If this is an "OS X virus" then so is tinyproxy or socks. Strike one.
Amphimix is a demonstration exploit that can only be used through a social engineering attack. A social engineering attack can't be completely defended against by the OS, but Apple has taken steps to mitigate this one (less well than I'd like, but probably adequately) and there are zero examples of it in the wild. Strike two.
Renepo is a payload only, it's got no remote propogation mechanism... Sophus notes "that any attacker trying to plant this worm in your network would need to get root access on one of your boxes first, meaning that you would already be "owned". Strike Three.
Sorry, Casey.
Of course nothing is as bad as the numbers for Windows but to say that your completely safe is pure arrogance.
I didn't say that you're completely safe. I said that there are zero exploits currently out there for OS X. Of course "past performance is not proof of future gains", but the level of risk is so low that you're more likely to suffer data loss or system damage from running A/V software than not running it.
If you're running a webserver or some other "attractive nuisance", then there's a whole different level of risk and a whole different approach necessary. But for a desktop user? I'm not sure I'd even recommend turning on Apple's firewall unless you're on a shared LAN... and even then I'd want to check the settings before I treated it as anything but a placebo.
Depends. Are you asking "if Apple announces the switch" or "if Apple successfully makes the switch"?
It's a big risk, with the potential of crippling the company. That's more than enough reason to object now. If you had a newspaper from 2010 that showed how Apple navigated the waters, that'd be a different question... but I don't think that's the alternate universe we're in right now.
Secondly, what does "making the switch to x86" mean? If it just means running OS X on non-x86-based Apple-branded hardware, I honestly don't see the point in Apple doing this. I don't see any upside for Apple, and I see a huge downside for both Apple and the current Mac user base. If it means Apple's going head-to-head with Microsoft, so I can buy "OS X Intel" for a few hundred bucks retail and install it on non-Apple hardware? That's a whole different kettle of Dogcows.
So... If it meant I could run Mac OS X on a Thinkpad, even if I paid twice as much for OS X for the privilege of running it on a non-Apple computer, then I'd be (tentatively) for it. If it just means I still have to buy a Powerbook... just an x86-based one, then what's the upside? Your x86 Mac would be slower for a long time (even on a faster processor) until all the apps are ported and then re-optimised, it'd be less secure because buffer overrun attacks are easier with the x86 instruction set (purely by chance, there's fewer NUL bytes in x86 code than most RISC processors, and NUL bytes are a string terminator in C), and there's no particular advantage in the forseeable future.
I mean if you know the future is x86, why would you get a PPC mac?
If this rumor had been around a couple of months ago, I wouldn't have bought a mini... I'd have stuck with my old "Beige G4". If I knew for certain? I sure wouldn't buy a new PPC Mac. If they do this, they better have the new models ready across the product line on day one, or a guaranteed affordable upgrade (substantially complete buyback, or a motherboard or CPU+ROM replacement) for all current models.
Either that or some kind of solid and credible commitment to OS upgrades for the current platform over the projected lifetime of the machines currently in service. The oldest machines supported on Tiger date to January 1999, so that means OS upgrades for 2005 models through 2010.
Don't know about the newest OS X
Well, I do. There are zero viruses in the world for OS X.
I seem to remember there being one
You remember wrong. There was a port of a UNIX rootkit floating around, and a couple of theoretical exploits based on holes Apple patched (albeit clumsily) quickly. A rootkit is software that's used after one's already broken into the computer... it's useless without a way to get it into the computer in the first place.
What you remember is probably that Symantec claimed that there WOULD be viruses and spyware on OS X, with no actual evidence. They've made the same claims before for Pocket PC and Palm OS when they were trying to push their antivirus solutions for those. Nothing has EVER happened to people who ignored any of these attempts by Symantec to use FUD to push their product. But... people have had their systems damaged by antivirus software on all these platforms.
I don't think that the reason there is no (or very little) Mac spyware/adware/malware is really only due to installed base.
Of course it isn't. It's because of the way IE is integrated into the desktop the HTML control has to guess whether a document it's opening is trusted or not, so if you can trick it into thinking your spyware is a local file you can get local user access. Microsoft tries to pop up a warning dialog when you're in a gray area, but there's so much gray area that people get used to clicking "OK" automatically.
Of course Apple has started to open up the same opportunities, first by having Safari "open safe files after downloading", then by treating Widgets and software installers as "safe files", and finally by using a dialog instead of backing out the bad behaviour. It's nowhere near as bad as it is in Windows, but it's worrying.
The PowerMac 8100 ran emulated 68k code faster than the fastest 68k mac ever made (the 840av).
Cite?
you quote 97 for archimedies which seems way way wrong to me and indeed wikipedia claims a more reasonable date of 1987
Sorry, that was a typo. Of course it should have been 1987 and 1989.
As for the magazine articles... sure you're not thinking of the Cambridge workstation (6502+NS32016)?
Can someone explain the difference between hacking XP Home to turn it into XP Pro and downloading a warez copy?
The former only became illegal when the DMCA was passed. The DMCA makes a lot of previously legal and still necessary actions illegal, so the fact that it's illegal under the DMCA is by itself irrelevant to the morality of the act. So it comes down to the morality of boosting the performance of a factory-crippled product. You can buy products for doing that at any auto-parts store.
The latter? XP Pro includes software that isn't included in XP Home. Even if you bought a copy of XP Home you're not entitled to that. But if Microsoft sold you more than they claimed they did, and you can turn that extra software on?
Violating a EULA may be illegal, but I'm not going to call it unethical.
Archimedes (mid '97) came over a year before the Acorn RISC Workstation (early '99).
Hmmm. The ARM, as I recall, started out as a processor for the Acorn line of personal computers in England. It was pretty impressive as a PC, but workstation-class?
What do you think the average Windows user needs to know other than how to download, install, and use Firefox, AVG, Ad-aware, Spybot, Gaim, and firewall?
That's still a lot (why Firefox rather than Netscape, for example? I know the answer, do you?), and it's not enough. You forgot about a non-Microsoft mail program and a non-Microsoft media player. Oh, and some commercial game software ON CD has spyware in it.
They need to know how to configure a firewall, how to tell when a message from the firewall is a problem or a false alarm (YOUR DNS SERVER IS HAXORING ME), when NOT to trust Firefox, when a message from antivirus or antispyware is a false alarm (TEH REGISTRY IS SPYING ON ME)...
I've had to deal with failures in all of the above, and I'm not even supposed to be Windows support any more!
And that's ignoring DLL conflicts (DLL HELL is not dead), driver conflicts (not bad hardware, hardware that's perfectly good, it just doesn't work with other hardware), application conflicts, deliberately badly-behaved applications (to keep Acrobat Reader from installing itself as the handler for PDF files in the browser, you have to go in and remove a file from the Acrobat directory), antivirus software conflicting with Microsoft's system checkpoints, and half-assed troubleshooting from well-intentioned friends.
i don't seem all that much sucess running xbox code on other x86 boxes
Emulating an operating system is a much harder problem than patching an operating system to run on different hardware. An operating system is a much more complicated beast than any hardware.
In this case the hypothetical attacker would already have the operating system, as well as pretty complete documentation of everything that runs above it. All that would be needed is to replace the OS components with ones that don't do the same checks.
And on the Xbox there's the additional problem of emulating the GPU, because you're trying to run games that expect to have a custom GPU, whereas on the Mac everything's mediated by OpenGL.
you have the source to darwin but not to things like apples grapics system
I've already mentioned elsewhere that accelerated graphics is the most credible place for Apple to embed a lock, but you could probably still get unaccelerated graphics working: this has already been done for the earliest Powermacs and clones, as well as emulation under MOL (Mac on Linux) and PearPC.
Itanium would at least give Apple a shot at outperforming Windows PC's.
Itanium's got a lot of registers, but it's a bloody awful emulation host because the best possible emulation technique, recompilation, is going to be agonizingly difficult. The Itanium instruction set architecture is so complex that the Intel/HP supercompiler does cross-module optimization... it actually moves code from one object module to another... and analyzes trace logs from actual code runs to get the kind of performance that makes the Itanium worthwhile. It's the only compiler I've used in 20 years where you want to turn off optimizations just because the compiler itself is significantly slowed down when it's using them.
even if you use more or less standard pc parts its still pretty easy to change enough to break stuff look at the xbox for example
Yeh, it's a perfect example. All you need to do to unbreak the XBox is run some code... any code... on it. Once you can do that you're home free. The hard part about cracking the XBox is tricking it into running some code for you. A general-purpose operating system by definition does not have that protection, especially when you're trying to break it the other way: you're booting the OS into an environment you control.
and unless you have source to the os its bloody hard to deal with such breakages
That's why it's a ludicrous idea: we do have the source.
Now we are waiting for IE to support the ACID2 test.
IE's target is still the Electric Kool-Aid Acid Test.
The PowerPC was a good fit to replace the 68k range. It was significantly faster, and could be used with the same endianisms[...]
How about a big-endian x86? That'd put the kibosh on the clone problem.
(no, I'm not serious, it'd still be stupid, because it'd effectively be a new architecture... and a really really horrid one...)
I know that I would gladly fork out the $400 for OSX to run it on my laptop.
I don't think they could do it at $400, that's almost as much as a Mac mini and that includes the hardware as well. No, Apple would have to sell OS X for the current price ($129) or less to get the kind of market that would make it worthwhile.
But OS X on a Thinkpad? A laptop Mac without the horrible 'book hardware? I could deal with that.
Make a proprietary motherboard architecture just like they have now with PowerPC based boards, except with an x86 CPU in place of the Gx chip.
The only thing on an Apple Motherboard that isn't regular PC hardware is the PPC CPU. Other than the CPU there's almost nothing proprietary there. PCI, AGP, PC RAM, PC USB chips, PC firewire chips, PC everything.
Then just make it so that the OS requires an Apple mobo to run. How?
They can't do it. Darwin is already out, it's open source, it's the Mac OS X kernel, and it runs on x86 hardware.
And in any case, the performance gap between PPC and x86 is tiny, sometimes PPC even gets ahead of x86 despite having a fraction of the resources spent on it. So there's no significant performance win for Apple going to x86, and a huge performance downside because they will have to do emulation for some apps, and that will be agonizingly slow.
So the only possibile reason for going to x86 would be to leave the hardware business and go head-to-head with Microsoft on x86 clones. And they MIGHT be able to pull it off, but it'd be a big gamble and a big loss in the short term.
Intel's going to be making PPCs.
That's what increasingly seems likely to me.
Apple's switching to ARM.
Only on a new Apple handheld running Newton OS. The ARM is a good low-power embedded chip, but it's nowhere near fast enough even for an entry-level Mac.
How about if they put an Intel chip in a low end computer, like the Mini, recompile the Apple-supplied applications so they run reasonably fast, and then emulate PPC for the 3rd party applications only.
Why? It's not the low end computers that need new chips. They're not even using the fastest available chip in the mini, and they're underclocking it at that.
And it would produce a system that's slower, more power-hungry, and more expensive than the mini. Running any 3rd party applications on it would be like running them on a Powermac 7600. The Beige G3, the oldest system OS X is supported on, would be faster. So to get acceptable speed they'd need to build it with a faster CPU, faster RAM, and more RAM... and it would still be slower.
And where do you get this "most buyers of the Mini wouldn't run 3rd party applications anyway" from, anyway?
but if the G4 and G5 can't hold up against the Pentium M and sequelae in the laptop game Apple will have to do something
That's up to Freescale. They're supposed to be sampling later this year. Dual-core G4 powerbooks early 2006?
I also agree with you that the x86 isn't an ideal PPC emulation platform, but where else will Apple turn?
IBM seems to have broken the 3 GHz barrier with the 970 core, so there's nothing stopping Apple from using those chips in the G5 to keep up with the numbers game. There's also dual-core G5s coming.
And unless there's something badly wrong with both the Freescale low-power chips and the IBM high-end chips, this will all come out sooner than they could possibly wield an x86 Mac with any kind of credible application base.
Will Intel risk Microsoft's wrath by luring Apple with a well-designed ISA?
I assume you're talking about Intel licensing the PPC instruction set from IBM, which is certainly a possibility... but that would render this whole discussion moot.
what's to stop Microsoft from porting Windows to that generation of Mac hardware?
Microsoft's already ported Windows to THIS generation of Mac hardware. That's what they use for their XBox 360 emulation. But they've got no reason to sell it, why would someone buy a Mac to run Windows on it unless they're an XBox developer? Changing the CPU over to an x86 wouldn't change that at all... if someone looking to buy a Mac wants to buy a Windows box they can buy one a LOT cheaper just buy buying a Windows box.
Microsoft could even entice a few Taiwanese whitebox makers to churn out Mac clones at a much smaller fraction of the cost Apple would want to sell them for.
They could do that now. You can buy a PPC motherboard right now, if you want, but the only market is the embedded systems business. And without the PPC a "Mac Compatible x86" is just another clone: you can already run the core of OS X on one today
Since Apple's margin on the iPod is more like 40% they're still making a nice profit, and a certain amount of the sales will be to people who hadn't planned on buying a new iPod but are reminded of the old 1st gen with the cracked screen in their drawer.
And at least some of the returns will be good enough to refurbish and feed back into their support and warranty system.
I don't see this as a cost for Apple at all.