There is not much that can be said now that everyone's head has exploded. I just have one thing to add that i think people are missing. Even if there was acomplete swithc to x86, it would insane trying to cover all the bases of mothernoards, and hardware, at least in my opinion. One of the thing i like about OSX is it either works or doesnt. For me there is no tweaking and trying to install drivers that don't work or then break your system ala windows. Or lack of any support at all like linux in some cases. But that is just my personal opinion. I would acttually like to see a dual core pentium-m powerbook next year with an altivec unit. But hey I can dream.
Would be a custom chip-set for Apple only. It would be easy to drop in hardware based security. and not have to worry about OSX runing on Dells. As far as other equipemtn goes I would expect that code examples in x86 are easier to find than PPC. I know that Intel has said that they have wanted Apple for a customer for about 20 years. I think that the big win for Apple would be first cost of development, and availability. If you go to Intel.com they talk about having PentiumM Dual Core chips for late '05 or early '06, Centrino as well, workstation class will go dual core with HT. Dual cores 3+GHz, able to support 4 threads per core. Numbers are hard to pin down but they look fine and if you look almost every different processor type is going dual core. They have 15 different projects for dual core. This could be a huge win for Apple and I almost forgot Intel is building WiFi and all things wireless into their MBs no add-ons, and they have the Intel Multimedia support built into their boards as well. Intel is kicking or should be again very soon. If the rumor is correct I would suspect that iBooks and Powerbooks were a real big thing for Apple and the PPC not going past 3.0GHz a ful, year after they had been promised that they would ship 3GHz chips was the other big hitter.
I found this article via Google news. Its take is that the move to Intel is pushed by Job's desire for the Pentium D chips which in turn are being pushed by Hollywood in order to implement DRM on video. The article takes the stance that Apple is not willing to cede the video entertainment possibilities.
I found this article via Google news. Its take is that the move to Intel is pushed by Job's desire for the Pentium D chips which in turn are being pushed by Hollywood in order to implement DRM on video. The article takes the stance that Apple is not willing to cede the video entertainment possibilities.
That pretty much makes sense.........hollywood will be one step closer to ruling the world.
It must be true. The NYTimes is reporting the switch:
SAN FRANCISCO, June 5 - Steven P. Jobs is preparing to take an unprecedented gamble by abandoning Apple Computer's 14-year commitment to chips developed by I.B.M. and Motorola in favor of Intel processors for his Macintosh computers, industry executives informed of the decision said Sunday.
It seems like the Times simply read the rumor sites and turned all the info into an article based on unnamed "industry executives."
It must be true. The NYTimes is reporting the switch:
SAN FRANCISCO, June 5 - Steven P. Jobs is preparing to take an unprecedented gamble by abandoning Apple Computer's 14-year commitment to chips developed by I.B.M. and Motorola in favor of Intel processors for his Macintosh computers, industry executives informed of the decision said Sunday.
It seems like the Times simply read the rumor sites and turned all the info into an article based on unnamed "industry executives."
Normally, I'd be laughing with you, but the reporter in question is John Markoff.
Markoff is one of the few reporters Uncle Steve cannot afford to screw with. For long, anyway.
I was a bit polite. Let me re-iterate. I know that Big endianess is not a problem.
I call BS.
I recently did an OS X port of some open source C code which had originally been written to run under x86 Linux. There were endian problems galore. Whoever wrote a few of the code modules couldn't decide between handling a few big tables of data as arrays of bytes or arrays of 32-bit words, and kept doing both all over the place, gleefully casting back and forth between uint8_t* and uint32_t*.
To get the code to run on PPC required adding some byte-order swapping in some places, but not in others, and determining where to swap and where not to wasn't at all straight forward. The port took measurable time and effort -- no way this was a case of "Oh, just recompile it!" I could have tried to rewrite the code to make it endian agnostic, but that would have taken even more time and effort -- a lot of time and effort since I didn't really understand how the code worked.
Would it have been better if the code had been written endian agnostic right from the start? Of course! But I also think it would be better if were paid 20 times my currently salary and only had to show up at work a couple of hours a week. If wishes were horses...
The world is fully of less-than-optimal, less-than-ideally-engineered code. In fact, the world depends on mountains of crappy code. Anyone who thinks the endian issue, and plenty of other issues, wouldn't make an x86 transition extremely painful hasn't a clue about the problems involved in that kind of porting.
Even if "Marklar" is ready to go and tuned to perfection on a few x86 Mac boxes deep in Apple's secret labs, that would represent only a tiny fraction of the development time and effort needed to get a user-satisfying number of x86/OS X apps, both Apple and third-party, up and running and reliable.
With separate binaries for x86 and PPC, only half the code is read, anyway, so the other stuff just sits there unused AND unloaded. I never heard anyone complain about Fat apps being bloated when the PPC came out, because people understood that some code would be ignored, and other run, depending on the machine. (Now people complained the files were big, and went through some effort to delete the PPC part if they didn't need it, because disk space was much more of a premium, and much more expensive per megabyte, then).
Also, nothing says an app has to be both a PPC and Intel app. Some 'fat' apps (that ran both on the PPC and 680x0) came in pairs, one for PPC and one for 680x0. But the only advantage here was disk space. Adobe's installer, for example, could be set up to install either binary, rather than both. However, the theory is that they'd need just one set of code to be compiled for each platform (note: no comment here on length of making it compatible for either).
And how long did it take developers to stop producing 68k versions of their applications? The answer was not long at all. And how did developers charge for PPC upgrades of their apps? Full price. Look, I don't want to be accused of spreading FUD here, let's assume the story is true: If we're going to rely on the generosity and competence of third parties to provide PPC/86k binaries for Apple's customers, we should start making those Linux contingency plans now. Apple is not in the position it was in the pre-Windows 95 days, and it's going to be hard pressed to convince most every developer that this alleged drastic, abrupt transition will pay off. Unless Transitive is the ABI holy grail, an Apple transition to 86k is dead in the water.
I recently did an OS X port of some open source C code which had originally been written to run under x86 Linux. There were endian problems galore. Whoever wrote a few of the code modules couldn't decide between handling a few big tables of data as arrays of bytes or arrays of 32-bit words, and kept doing both all over the place, gleefully casting back and forth between uint8_t* and uint32_t*.
Correct if you're porting C code that does all the nasty things C can do.
Well written C++ has fewer problems, and fewer still if it is written using the Carbon APIs and uses the standard objects because they handle byte-swapping themselves.
More so with Cocao.
This task is exactly what I had to do in porting 500K lines of Objective-C for NEXTSTEP (68k) to NEXTSTEP/Intel (i386).
90% of it was a simple recompile, all the GUI, all the disk I/O was fine except where I had stepped outside the APIs and wrote my own stuff that did not account for endianness of the code.
My expectation is that people will find the same things in graphics intensive software where you monkey around but bits and bytes yourself. If you're using APIs to do it (e.g. CoreImage, quicktime) you won't have any problems.
That means that 90% of software will be trivial to port. My guess is that apps like iWork, Pages, Safari, iPhoto, even FCP and DVDSP will be pretty simple because all the frameworks are already multi-architecture. Apps that do more complex things (e.g. Photoshop) will be more problematic, but it's still doable.
I've been through this before and yes, it's a pain, but nowhere near the pain of having to completely rewrite your code - unless you've been a bad programmer in which case you deserve to be spanked, preferably by someone in leather - and you're NOT allowed to enjoy it.
"Transitive's QuickTransit allows any software to run on any hardware with no performance hit, or so the company claims. The techology automatically kicks in when necessary, and supports high-end 3D graphics. It was developed by Alasdair Rawsthorne.
When I wrote about the software for Wired News last fall, the company had PowerBooks and Windows laptops running Linux software, including Quake III, with no performance lag whatsoever.
If Apple has licensed QuickTransit for an Intel-powered Mac, all current applications should just work, no user or developer intervention required.
Programmers could port their software to the new platform slowly and steadily, and the shift would be as relatively painless as the recent move from OS 9 to OS X, which, of course, relied on emulation in the Classic environment.
But why would Apple do this? Because Apple wants Intel's new Pentium D chips."
INFO for you new timers, Apple has used Transitives software a couple of times in the past.
"According to News.com, the Intel transition will occur first in the summer with the Mac mini, which I'll bet will become a mini-Tivo-cum-home-server.
Hooked to the internet, it will allow movies to be ordered and stored, and if this News.com piece is correct, loaded onto the video iPod that's in the works."
"Presumably, Jobs used his Pixar moxie to persuade Hollywood to get onboard, and they did so because the Mac platform is seen as small and isolated -- just as it was when the record labels first licensed music to iTunes. The new Mac/Intel platform will be a relatively isolated test bed for the digital distribution of movies and video.
Will current Mac users like this new locked-down platform? I doubt it, which I guess is why it's going into consumer devices first.
In the PC industry, Apple lost the productivity/office era to Microsoft, but it's trying to get the jump on the next big thing: the entertainment/creativity era, and it's going to drag it users, even if they're kicking and screaming, with it."
Look, I don't want to be accused of spreading FUD here, let's assume the story is true: If we're going to rely on the generosity and competence of third parties to provide PPC/86k binaries for Apple's customers, we should start making those Linux contingency plans now.
I'm afraid you may be right. Too much change may mark the end of Apple. How many rivers must we cross to stick with an admittedly great platform? Jobs may end up trying everyone's patience, from his customers to his developers to his own amazing employees. (And the husbands and wives at Cupertino were complaining about too much overtime a few months ago ... Just think what the mood at Apple will be like if there's a full-court press for x86.)
... here they are on the cusp of a huge rise in Apple Macintosh marketshare ....
The huge rise you allude to will turn to vapor unless Steve and company have plans for an amazingly seamless transition to Intel. Like everyone, I'd love to be wowed by news tomorrow of a painless segue, but I'll believe it when I see it.
Why? You don't want a computer that's always at least as fast hardware wise as your windows "friends"? You'd rather it get slower? Apple has always had problems since the beginning of time. They always come out with something super fast and amazing only to not be able to update it for 2 years... The history with motorola, ibm has been bad. At the very least with Intel you'll get a faster computer now.
I love my dual 2.5ghz g5, fully loaded, that cost me $3,900.... However my $1500 windows machine at work is much faster, and cheepier to maintain.
Imagine more affordable computers with OS X, that are even faster than what you have now.
Quote:
Originally posted by dstranathan
amen, brotha. this will be "character building" for many of us.
Comments
Originally posted by myahmac
There is not much that can be said now that everyone's head has exploded. I just have one thing to add that i think people are missing. Even if there was acomplete swithc to x86, it would insane trying to cover all the bases of mothernoards, and hardware, at least in my opinion. One of the thing i like about OSX is it either works or doesnt. For me there is no tweaking and trying to install drivers that don't work or then break your system ala windows. Or lack of any support at all like linux in some cases. But that is just my personal opinion. I would acttually like to see a dual core pentium-m powerbook next year with an altivec unit. But hey I can dream.
Would be a custom chip-set for Apple only. It would be easy to drop in hardware based security. and not have to worry about OSX runing on Dells. As far as other equipemtn goes I would expect that code examples in x86 are easier to find than PPC. I know that Intel has said that they have wanted Apple for a customer for about 20 years. I think that the big win for Apple would be first cost of development, and availability. If you go to Intel.com they talk about having PentiumM Dual Core chips for late '05 or early '06, Centrino as well, workstation class will go dual core with HT. Dual cores 3+GHz, able to support 4 threads per core. Numbers are hard to pin down but they look fine and if you look almost every different processor type is going dual core. They have 15 different projects for dual core. This could be a huge win for Apple and I almost forgot Intel is building WiFi and all things wireless into their MBs no add-ons, and they have the Intel Multimedia support built into their boards as well. Intel is kicking or should be again very soon. If the rumor is correct I would suspect that iBooks and Powerbooks were a real big thing for Apple and the PPC not going past 3.0GHz a ful, year after they had been promised that they would ship 3GHz chips was the other big hitter.
Originally posted by brent1a
The illustrated dude overlayed on the pic looks like a NAMBLA reject.
That is somebodies rendition of SJ !?!?!?
Originally posted by Kishan
I found this article via Google news. Its take is that the move to Intel is pushed by Job's desire for the Pentium D chips which in turn are being pushed by Hollywood in order to implement DRM on video. The article takes the stance that Apple is not willing to cede the video entertainment possibilities.
That pretty much makes sense.........hollywood will be one step closer to ruling the world.
SAN FRANCISCO, June 5 - Steven P. Jobs is preparing to take an unprecedented gamble by abandoning Apple Computer's 14-year commitment to chips developed by I.B.M. and Motorola in favor of Intel processors for his Macintosh computers, industry executives informed of the decision said Sunday.
It seems like the Times simply read the rumor sites and turned all the info into an article based on unnamed "industry executives."
Originally posted by jante99
It must be true. The NYTimes is reporting the switch:
SAN FRANCISCO, June 5 - Steven P. Jobs is preparing to take an unprecedented gamble by abandoning Apple Computer's 14-year commitment to chips developed by I.B.M. and Motorola in favor of Intel processors for his Macintosh computers, industry executives informed of the decision said Sunday.
It seems like the Times simply read the rumor sites and turned all the info into an article based on unnamed "industry executives."
Normally, I'd be laughing with you, but the reporter in question is John Markoff.
Markoff is one of the few reporters Uncle Steve cannot afford to screw with. For long, anyway.
Originally posted by jante99
It seems like the Times simply read the rumor sites and turned all the info into an article based on unnamed "industry executives."
Nope, note that it says that is contains comments from industry execs on Sunday. That means they've dug deeper.
Real news organizations don't just quote internet stories, they research it and if they have confirmation they publish them.
The NYT isn't a blog...
Originally posted by asdasd
I was a bit polite. Let me re-iterate. I know that Big endianess is not a problem.
I call BS.
I recently did an OS X port of some open source C code which had originally been written to run under x86 Linux. There were endian problems galore. Whoever wrote a few of the code modules couldn't decide between handling a few big tables of data as arrays of bytes or arrays of 32-bit words, and kept doing both all over the place, gleefully casting back and forth between uint8_t* and uint32_t*.
To get the code to run on PPC required adding some byte-order swapping in some places, but not in others, and determining where to swap and where not to wasn't at all straight forward. The port took measurable time and effort -- no way this was a case of "Oh, just recompile it!" I could have tried to rewrite the code to make it endian agnostic, but that would have taken even more time and effort -- a lot of time and effort since I didn't really understand how the code worked.
Would it have been better if the code had been written endian agnostic right from the start? Of course! But I also think it would be better if were paid 20 times my currently salary and only had to show up at work a couple of hours a week. If wishes were horses...
The world is fully of less-than-optimal, less-than-ideally-engineered code. In fact, the world depends on mountains of crappy code. Anyone who thinks the endian issue, and plenty of other issues, wouldn't make an x86 transition extremely painful hasn't a clue about the problems involved in that kind of porting.
Even if "Marklar" is ready to go and tuned to perfection on a few x86 Mac boxes deep in Apple's secret labs, that would represent only a tiny fraction of the development time and effort needed to get a user-satisfying number of x86/OS X apps, both Apple and third-party, up and running and reliable.
Originally posted by Louzer
With separate binaries for x86 and PPC, only half the code is read, anyway, so the other stuff just sits there unused AND unloaded. I never heard anyone complain about Fat apps being bloated when the PPC came out, because people understood that some code would be ignored, and other run, depending on the machine. (Now people complained the files were big, and went through some effort to delete the PPC part if they didn't need it, because disk space was much more of a premium, and much more expensive per megabyte, then).
Also, nothing says an app has to be both a PPC and Intel app. Some 'fat' apps (that ran both on the PPC and 680x0) came in pairs, one for PPC and one for 680x0. But the only advantage here was disk space. Adobe's installer, for example, could be set up to install either binary, rather than both. However, the theory is that they'd need just one set of code to be compiled for each platform (note: no comment here on length of making it compatible for either).
And how long did it take developers to stop producing 68k versions of their applications? The answer was not long at all. And how did developers charge for PPC upgrades of their apps? Full price. Look, I don't want to be accused of spreading FUD here, let's assume the story is true: If we're going to rely on the generosity and competence of third parties to provide PPC/86k binaries for Apple's customers, we should start making those Linux contingency plans now. Apple is not in the position it was in the pre-Windows 95 days, and it's going to be hard pressed to convince most every developer that this alleged drastic, abrupt transition will pay off. Unless Transitive is the ABI holy grail, an Apple transition to 86k is dead in the water.
Originally posted by brent1a
That pretty much makes sense.........hollywood will be one step closer to ruling the world.
I told somebody...
Pixar "MIND CONTROL MOVIE"
Originally posted by shetline
I call BS.
I recently did an OS X port of some open source C code which had originally been written to run under x86 Linux. There were endian problems galore. Whoever wrote a few of the code modules couldn't decide between handling a few big tables of data as arrays of bytes or arrays of 32-bit words, and kept doing both all over the place, gleefully casting back and forth between uint8_t* and uint32_t*.
Correct if you're porting C code that does all the nasty things C can do.
Well written C++ has fewer problems, and fewer still if it is written using the Carbon APIs and uses the standard objects because they handle byte-swapping themselves.
More so with Cocao.
This task is exactly what I had to do in porting 500K lines of Objective-C for NEXTSTEP (68k) to NEXTSTEP/Intel (i386).
90% of it was a simple recompile, all the GUI, all the disk I/O was fine except where I had stepped outside the APIs and wrote my own stuff that did not account for endianness of the code.
My expectation is that people will find the same things in graphics intensive software where you monkey around but bits and bytes yourself. If you're using APIs to do it (e.g. CoreImage, quicktime) you won't have any problems.
That means that 90% of software will be trivial to port. My guess is that apps like iWork, Pages, Safari, iPhoto, even FCP and DVDSP will be pretty simple because all the frameworks are already multi-architecture. Apps that do more complex things (e.g. Photoshop) will be more problematic, but it's still doable.
I've been through this before and yes, it's a pain, but nowhere near the pain of having to completely rewrite your code - unless you've been a bad programmer in which case you deserve to be spanked, preferably by someone in leather - and you're NOT allowed to enjoy it.
http://www.wired.com/news/mac/0,2125...w=wn_tophead_1
"Transitive's QuickTransit allows any software to run on any hardware with no performance hit, or so the company claims. The techology automatically kicks in when necessary, and supports high-end 3D graphics. It was developed by Alasdair Rawsthorne.
When I wrote about the software for Wired News last fall, the company had PowerBooks and Windows laptops running Linux software, including Quake III, with no performance lag whatsoever.
If Apple has licensed QuickTransit for an Intel-powered Mac, all current applications should just work, no user or developer intervention required.
Programmers could port their software to the new platform slowly and steadily, and the shift would be as relatively painless as the recent move from OS 9 to OS X, which, of course, relied on emulation in the Classic environment.
But why would Apple do this? Because Apple wants Intel's new Pentium D chips."
INFO for you new timers, Apple has used Transitives software a couple of times in the past.
"According to News.com, the Intel transition will occur first in the summer with the Mac mini, which I'll bet will become a mini-Tivo-cum-home-server.
Hooked to the internet, it will allow movies to be ordered and stored, and if this News.com piece is correct, loaded onto the video iPod that's in the works."
"Presumably, Jobs used his Pixar moxie to persuade Hollywood to get onboard, and they did so because the Mac platform is seen as small and isolated -- just as it was when the record labels first licensed music to iTunes. The new Mac/Intel platform will be a relatively isolated test bed for the digital distribution of movies and video.
Will current Mac users like this new locked-down platform? I doubt it, which I guess is why it's going into consumer devices first.
In the PC industry, Apple lost the productivity/office era to Microsoft, but it's trying to get the jump on the next big thing: the entertainment/creativity era, and it's going to drag it users, even if they're kicking and screaming, with it."
WIRED always has the BEST news !!!
originally posted by Big Mac:
Look, I don't want to be accused of spreading FUD here, let's assume the story is true: If we're going to rely on the generosity and competence of third parties to provide PPC/86k binaries for Apple's customers, we should start making those Linux contingency plans now.
I'm afraid you may be right. Too much change may mark the end of Apple. How many rivers must we cross to stick with an admittedly great platform? Jobs may end up trying everyone's patience, from his customers to his developers to his own amazing employees. (And the husbands and wives at Cupertino were complaining about too much overtime a few months ago ... Just think what the mood at Apple will be like if there's a full-court press for x86.)
I would die before buying a powerbook with an Intel Inside sticker.
god damn.
Originally posted by soulcrusher
we've come to hate Intel so much. I don't think many will be able to accept the switch. seriously.
I would die before buying a powerbook with an Intel Inside sticker.
god damn.
I don't mind Intel at all, in a lot of ways IBM is worse then Intel.
Now Microsoft, without question they IS BEAST !!!
Originally posted by soulcrusher
we've come to hate Intel so much. I don't think many will be able to accept the switch. seriously.
I would die before buying a powerbook with an Intel Inside sticker.
god damn.
amen, brotha. this will be "character building" for many of us.
originally posted by 9secondko:
... here they are on the cusp of a huge rise in Apple Macintosh marketshare ....
The huge rise you allude to will turn to vapor unless Steve and company have plans for an amazingly seamless transition to Intel. Like everyone, I'd love to be wowed by news tomorrow of a painless segue, but I'll believe it when I see it.
I love my dual 2.5ghz g5, fully loaded, that cost me $3,900.... However my $1500 windows machine at work is much faster, and cheepier to maintain.
Imagine more affordable computers with OS X, that are even faster than what you have now.
Originally posted by dstranathan
amen, brotha. this will be "character building" for many of us.