Does OPTIONS mean this....
Apple first launched the server version of OSX (called Rhapsody or something like that) running on G3 machines. The operating system was then modified and released only to run on G4 machines (after Altivec acceleration). The core of OSX was tested for atleast 2 years in a server environment before it turned up with Aqua for the General Desktop user. This gave OSX the stability that very few os's have when they are first launched.
Assuming that Apple has known for more than 12 months that Motorola would not be able to continue development of new processors, in all likelihood they would have approached IBM and asked them to develop a new processor (unless this was done during the 500Mhz fiasco).
I have a theoretical situation here and require some answers from people who are programmers on whether the following is possible or not.
Ever since the launch of OSX and more so in the past 6-8 months, Steve Jobs has mentioned many a time that Apple needs to shift away from OS9 as soon as possible. Pooling developer resources into one OS only is the most logical arguement.
However consider the following scenario....
OSX is based on the Nextstep. It was basically coded for the x86 platform. If, Apple had decided to have a Plan C and go with yet another CPU maker. I would assume that they would have decided AMD rather than Intel (this is based on heresay and the enemies enemy is a friend syndrome).
I am further assuming that Apple decided to skip the current generation processor family and go with the Hammer instead). Apple can proceed by first porting the base OSX to run on the Hammer CPU. With this, they would have all the Iapps and the other extras with OSX. These Apple developed machines (3U rack server is my guess as the first machine), would blow away most other machines by having a brilliant OS and very fast and modern harware. Since these machine are essentialy servers, Apple would only have to ensure that there is a minimum set of third part software that need run on these machines from day one. Of this probable Oracle could be the single most important application. Won't be too tough to convince them to put in the resources would it?
Apple could then wait and see how these machines perform against whatever PPC processor they have at that time and then tailor make machines for different markets. And stay on the curve with regards to Hardware. They could then play one supplier against the other and ensure that they always have the fastest chips.
For this people would have to get away from OS9 apps as these are never going to run on the Hammer.
All this is based on the assumption that a recompile on the part of the 3rd party software developers would be required to generate a PPC version vs a Hammer version of their applications for OSX.
My question is, is this possible? (From what I know, there were PPC and Intel versions of NT and of Linux and Unix). If it is possible is too troublesome that developers would crib?
There are other advantages:-
1. Maybe a Classic type software from Apple that can run on these hammer boxes and run Windows apps
Apple can then take on Microsoft head on. No version of Windows can offer the tight intergration of hardware and software that OSX has. Also, Apple has proven with Safari that they can code apps like no one else.
I'm also assuming that sometime like 2 years back the also decided to make Appleworks 7 better than MS Office (entirely in cocoa) and take MS heads on.
If the hurdles are not too great, then I think that Apple has a very very bright future
Please let me have your opinions...
thecube1359
Assuming that Apple has known for more than 12 months that Motorola would not be able to continue development of new processors, in all likelihood they would have approached IBM and asked them to develop a new processor (unless this was done during the 500Mhz fiasco).
I have a theoretical situation here and require some answers from people who are programmers on whether the following is possible or not.
Ever since the launch of OSX and more so in the past 6-8 months, Steve Jobs has mentioned many a time that Apple needs to shift away from OS9 as soon as possible. Pooling developer resources into one OS only is the most logical arguement.
However consider the following scenario....
OSX is based on the Nextstep. It was basically coded for the x86 platform. If, Apple had decided to have a Plan C and go with yet another CPU maker. I would assume that they would have decided AMD rather than Intel (this is based on heresay and the enemies enemy is a friend syndrome).
I am further assuming that Apple decided to skip the current generation processor family and go with the Hammer instead). Apple can proceed by first porting the base OSX to run on the Hammer CPU. With this, they would have all the Iapps and the other extras with OSX. These Apple developed machines (3U rack server is my guess as the first machine), would blow away most other machines by having a brilliant OS and very fast and modern harware. Since these machine are essentialy servers, Apple would only have to ensure that there is a minimum set of third part software that need run on these machines from day one. Of this probable Oracle could be the single most important application. Won't be too tough to convince them to put in the resources would it?
Apple could then wait and see how these machines perform against whatever PPC processor they have at that time and then tailor make machines for different markets. And stay on the curve with regards to Hardware. They could then play one supplier against the other and ensure that they always have the fastest chips.
For this people would have to get away from OS9 apps as these are never going to run on the Hammer.
All this is based on the assumption that a recompile on the part of the 3rd party software developers would be required to generate a PPC version vs a Hammer version of their applications for OSX.
My question is, is this possible? (From what I know, there were PPC and Intel versions of NT and of Linux and Unix). If it is possible is too troublesome that developers would crib?
There are other advantages:-
1. Maybe a Classic type software from Apple that can run on these hammer boxes and run Windows apps
Apple can then take on Microsoft head on. No version of Windows can offer the tight intergration of hardware and software that OSX has. Also, Apple has proven with Safari that they can code apps like no one else.
I'm also assuming that sometime like 2 years back the also decided to make Appleworks 7 better than MS Office (entirely in cocoa) and take MS heads on.
If the hurdles are not too great, then I think that Apple has a very very bright future
Please let me have your opinions...
thecube1359
Comments
<strong>OSX is based on the Nextstep. It was basically coded for the x86 platform.</strong><hr></blockquote>
NeXTStep was _ported_ to the x86 platform. It was basically coded for NeXT.
The "moving to x86" was discussed at lenght here and the sum up would be:
Not gonna happen. Not financially viable, would put Apple head to head with Dell et.al. and Microsoft. Who would code for OSX on Hammer if their software is already running fine on Windows/Hammer? The PowerPC is more suitable for servers than x86 CPUs are (heat/power consumption being one reason).
Moving software from x86 to PPC _can_ be a simple recompile, but in most cases you need different optimizations and a whole set of new hardware to support (not even starting about Cocoa). Once people get a Hammer OSX they will also demand drivers for their Wildcat or Kyro 3D cards. Or SiS or VIA. They will want their Sound Blaster WhateverModel supported and also, say, their 2nd ethernet controller. Either Apple starts writing tons of drivers (or tries to force the manufacturers to do so) or people will have even less reasons to switch to a Hammer OSX machine.
Besides the Wintel CPU speeds are basically due to insane clocking of the chips (a 3GHz Pentium 4 uses 100 Watts) and to most people but some POWER users there is no difference between a dual 1.25 GHz G4 machine and a 3.0 GHz Pentium4. Most people are even happy with a 1 GHz G4 or a 1.8 GHz Pentium4 machine.
<strong>I am talking about Apple Made hardware that carry the hammer chips. Apple can then only support the display cards that they want to, similar to what they are doing now.</strong><hr></blockquote>
Yeah, but still sooner or later someone might come up with a "clone" that mimics the Apple hardware and runs OSX which would hurt Apple sales. And still, _if_ Apple gets the PowerPC 970 from IBM that would mean one more new CPU to support and developers are not keen on supporting all new CPUs from the start, especially if it's uncertain how good it will sell.
It's not that it wouldn't be interesting to at least compare performance - quite the opposite. It's just that with a lower marketshare Apple can not afford experimenting too much, since it might wound the company severely or even kill it (if they bank all their money on the wrong horse).
If Apple made servers only it wouldn't be that bad, but they constantly are at risk of eating into their own sales and if a Hammer Mac is the only thing they'd be selling well that pretty much nullifies all their investment in PowerPC.
It's simply too risky, imho. But then, we don't even know how the 970 will truly compare to the Hammer - maybe it will be twice as fast/effective or maybe twice as slow. Certainly the 970 would mean less work/cost to port all the OSX stuff to, though.
Also, if I am not mistaken the 970 is backward compatible and also altivec compatible.
gong
[ 01-16-2003: Message edited by: Krassy ]</p>
[quote]Originally posted by thecube1359:
<strong>Apple first launched the server version of OSX (called Rhapsody or something like that) running on G3 machines. The operating system was then modified and released only to run on G4 machines (after Altivec acceleration).</strong><hr></blockquote>
OS X runs on G3 processors very well; see the iBooks and CRT iMacs, for instance. Much of it is optimized for altivec, however, and will perform much better on a G4.
[quote] OSX is based on the Nextstep. It was basically coded for the x86 platform.<hr></blockquote>
Actually, NeXT's OS was originally coded for Motorola processors (the 68000 series) and was later ported to x86 as OpenStep.
Blark
<strong>
Yeah, but still sooner or later someone might come up with a "clone" that mimics the Apple hardware and runs OSX which would hurt Apple sales.</strong><hr></blockquote>
Well, PPC chips can be sold to virtually every computer maker around the world! Cloning a x86-based Mac wouldn't be easier than cloning a PPC-based Mac... if Apple ever released a x86-based Mac, they wouldn't open their platform for cloning, for sure! Don't you remember how they killed clone-makers and the CHRP platform back in 97?
However, that's right, moving to x86 would be a pain in the ass for Apple and its developers. And for the Mac users too. After all, i moved to the PPC in 93, and that was painful. I moved to Mac OS X two years ago, and that was painful too (and it's still painful sometimes...). I don't want to move anymore... well, at least until 2006-2007; let me breath!!!!
The Nocturnal
<strong>Well, PPC chips can be sold to virtually every computer maker around the world! Cloning a x86-based Mac wouldn't be easier than cloning a PPC-based Mac...</strong><hr></blockquote>
Yes, but who wants to start building PPC machines at this point? Surely not many would like a 1.25GHz top of the line model. Plus there might be companies already having experience with x86 chipsets whom may find it easier.
Altivec is just a name for a set of vector processing routines, no? (excuse my terminology?) I mean, why couldn't AMD lincense Altivec and build an X86 chip with it? Or better yet, why couldn't AMD build a PPC compatible chip for Apple? Are we missing something here by focussing only on the X86 possibility? I mean, the AMD is already a RISC chip at it's core...
I think that if iit were true that the G4 could perform as well as a pentium 4, i doubt many Mac users would be complaining. But as it currently stands, the Mac is much more expensive and comparitively underpowered, although they more or less run the same software apps.
IF Apple is able to close the PERFORMANCE GAP (forget MHz), Why would they even need to port the OS to X86 at all, assuming that the only reason they'd do it is for the performance edge?
Personally, I get the feeling that Apple is listening, that they're focused on fixing the performance gap and that we should be doing a polite little queen-wave to the G4/moto era soon enough.
<strong>
Yes, but who wants to start building PPC machines at this point? Surely not many would like a 1.25GHz top of the line model. </strong><hr></blockquote>
Well, maybe server makers would like to use POWER4 chips, or maybe low-end desktop makers would use G3 class CPUs (some say that IBM produces 1.6GHz G3 CPUs), who knows?
[quote]<strong>Plus there might be companies already having experience with x86 chipsets whom may find it easier.</strong><hr></blockquote>
That wouldn't change anything: the Mac platform is proprietary. Even if cloning a Mac is technically possible, that remains illegal...
<strong>
That wouldn't change anything: the Mac platform is proprietary. Even if cloning a Mac is technically possible, that remains illegal...</strong><hr></blockquote>
Yeah, but so is MP3 sharing, DVD DECSSing, and a host of other things people do. If someone made it easy enough to hack a mac, people would do it and Apple could do little about it.
<strong>
Yeah, but so is MP3 sharing, DVD DECSSing, and a host of other things people do. If someone made it easy enough to hack a mac, people would do it and Apple could do little about it.</strong><hr></blockquote>
Well, MP3 sharing and DVD DECSSing don't need much R&D investment! However, creating a clone from the ground up
costs money: and do you think that one can raise funds so easily for an illegal project, or do you think that big companies would launch illegal research? I don't think so! If it were so, Umax, Power Computing and even Mot would have gone on producing clones: after all, their businesses were not going so bad when Apple forced them to cease their activity!
The Nocturnal
<strong>or do you think that big companies would launch illegal research? I don't think so!</strong><hr></blockquote>
BWAHAHAHAHAHAHAHAHAHAHAHAHAHAHA!
*sigh*
Sorry, but this is just too silly.
Companies perform illegal research all the time... they just try not to get caught doing it.
<strong>
BWAHAHAHAHAHAHAHAHAHAHAHAHAHAHA!
*sigh*
Sorry, but this is just too silly.
Companies perform illegal research all the time... they just try not to get caught doing it.</strong><hr></blockquote>
All right! That was silly! But do you think that such companies would release the fruits of their illegal research?? Maybe some firms have Mac clones in their labs, but they won't ever release them unless Apple gives them the right to! Which won't happen... unless Steve Jobs quits (or else: maybe he could get assassined by some mac clone fanatic or even worse, by some M.Dell maniac!)
[ 01-16-2003: Message edited by: The Nocturnal ]</p>
<strong>Questions about porting and altivec:
Altivec is just a name for a set of vector processing routines, no? (excuse my terminology?) I mean, why couldn't AMD lincense Altivec and build an X86 chip with it? Or better yet, why couldn't AMD build a PPC compatible chip for Apple? Are we missing something here by focussing only on the X86 possibility? I mean, the AMD is already a RISC chip at it's core...
I think that if iit were true that the G4 could perform as well as a pentium 4, i doubt many Mac users would be complaining. But as it currently stands, the Mac is much more expensive and comparitively underpowered, although they more or less run the same software apps.
IF Apple is able to close the PERFORMANCE GAP (forget MHz), Why would they even need to port the OS to X86 at all, assuming that the only reason they'd do it is for the performance edge?
Personally, I get the feeling that Apple is listening, that they're focused on fixing the performance gap and that we should be doing a polite little queen-wave to the G4/moto era soon enough.</strong><hr></blockquote>
If AMD were to graft on AltiVec into their X86 processors it would no longer be compatiable with MMX or whatever the X68 is using these days. This would mean that the processors would take a hit in performance on the windows when compared to Intel processors. Unless AMD were to get Microsoft to optamize for AltiVec, this would be a VERY BAD IDEA for AMD.
<strong>My question is, is this possible? (From what I know, there were PPC and Intel versions of NT and of Linux and Unix). If it is possible is too troublesome that developers would crib?</strong><hr></blockquote>
Yes, this is possible. There would be some work making sure that their i/o code was endian-safe, but other than that, it should be a simple recompile, both the Carbon and Cocoa APIs are endian-safe already.
In fact, in Project Builder right now, you can select x86 as a build platform (it's always been this way). It'll fail on linking for lack of x86 frameworks/libs, but hey...
<strong>
That wouldn't change anything: the Mac platform is proprietary. Even if cloning a Mac is technically possible, that remains illegal...</strong><hr></blockquote>
Not true. It is no more illegal to clone Mac hardware than it is to clone IBM hardware. However, it is illegal to build a computer that uses Apple's intellectual property without the company's permission. Back when Apple used the Toolbox ROM, it was simply a lot harder to reverse engineer the Mac than it was to reverse engineer the IBM PC. However, if you could have proven in court that your Mac clone used no Apple intellectual property, then it would have been legal for you to manufacture and sell it. That was never done.
<strong>
Not true. It is no more illegal to clone Mac hardware than it is to clone IBM hardware. However, it is illegal to build a computer that uses Apple's intellectual property without the company's permission. Back when Apple used the Toolbox ROM, it was simply a lot harder to reverse engineer the Mac than it was to reverse engineer the IBM PC. However, if you could have proven in court that your Mac clone used no Apple intellectual property, then it would have been legal for you to manufacture and sell it. That was never done.</strong><hr></blockquote>
You are 100% right!
Anyway, it would not be that hard for Apple to modify Mac OS X to prevent such clones from working with Mac OS... Plus that's always the same problem than the emulation problem: pretty hard to know where copy stops and where innovation starts... Connectix have been suited several times by Sony in similar cases (although i dunno if they won or not...)
We can speculate all we like, but we can say that due to limited resources, Apple will take only one of those two roads: they will either go Hammer or 970. Motorola may be delivering 7457s this year, but that's probably the end of the G4 line; 1.6GHz will be as fast as they ever go.
If I were SJ, I'd be inclined toward the 970. But the decision which looms is not one which can be made easily, nor can it be unmade easily. It's got to be the best, most informed decision possible. Therefore, I'd make sure I had SEVERAL machines of BOTH types running in the lab. I'd set up two teams to compete.