Dedicated Mac users generally have little use for Windows emulation. Its main value would be in enticing switchers by lowering the "entry fee" to getting onto OS X.
I'm a dedicated Mac user, but I also have to earn aliving from people who are dedicated Windows users - I'm an IT consultant/project manager/cook/bottle washer - so I need to be able to run Project (FastTrack doesn't do everything I want) and a couple of other bits and pieces.
VPC or another emulator would not just be for making life easier for switchers - it has the effect of turning the Mac into the digital equivalent of the Swiss Army Knife or Leatherman tool.
I've felt the need to elaborate on my comments concerning the untenability of both the OS X x86 and the Mac OS X + Integrated Win32 Box ideas. Take a second to think about it with me.
Why OS X x86 Won't Come Out
As I said before, I believe OS X x86 is kept up-to-date with the PPC binary and that it is a last resort contingency option. Here's why Apple will not announce and market it:
1) Where are the apps? Here's an obvious issue. Customer puts OS X x86 on a PC and marvels at Aqua. He or she browses with Safari and writes something in TextEdit. Okay, Customer thinks, "OS X is a nice OS. Now what?" What applications will the customer run outside of those bundled with the OS? Where are the third party apps that so many rely on? Where's Photoshop or Flash MX? Where are the games? What can be done with this environment? The Mac applications on the market WILL NOT run on a Wintel box Of course, shareware people would pretty quickly recompile for Intel. And if there were incentive, some third parties would provide an Intel binary. But there wouldn't be any such incentive. The Mac's third party support is limited enough, so releasing OS X x86 with virtually no third party support would be a disaster.
2) It would kill Apple?s hardware If you wish to ignore the first reason, and you believe OS X x86 could be a success, you have to ask yourself why Apple would do it. It would obviously hurt its hardware division, as others have stated. If OS X x86 were a viable product, people would favor it because PC hardware will always be cheaper than Apple?s. Apple would have to price the OS far too high in order to make up lost revenue from hardware cannibalization. OS X x86 would kill Apple hardware, and I don?t think Apple?s prepared to do that.
3) It would panic the Mac community No matter what one says, a highly publicized, consumer version of OS X x86 would panic the Mac community. We would legitimately fear that Apple could plan to move to x86 fully. Especially if OS X x86 took off, that fear would ominously cloud the Mac?s future.
4) How to Support PCs A significant amount of Windows development costs goes to trying to support the myriad of PC variants out there. The PC world is a jumbled mess -- just look at how many different motherboards there are. Moreover, Apple has enough trouble supporting the comparatively few designs it produces. Now you can avoid this trap by suggesting only certain machines will be supported, but if you do you still have to defend your argument against all the other issues.
5) If successful, MS would go nuts If the product were successful, MS could quickly obliterate Apple. MS was able to evade the government yet again, so finally taking out Apple wouldn?t be too risky, especially if MS felt vulnerable.
Why Mac OS X + Integrated Win32 Box won?t come out
I have fewer objections to the scheme above, but I maintain it?s just as infeasible as OS X x86. It seems like a great idea ostensibly, though. What Mac user wouldn?t want to be able to run the occasional Windows application in a transparent, Classic-like environment? And with the 970, we?d have emulation speed that would be worth writing home about. Yet, the one major problem is the OS/2 example. IBM created an OS far superior to Windows that ran Windows software with nearly full speed and compatibility. So why did OS/2 die? It?s because it ran Windows software so well that no one bothered to code OS/2 applications!
Many developers barely have the desire to code for the Mac anyway, and now we?d be providing a really great incentive to halt development totally. Consider the following scenario:
Customer: ?Hi, I?m a Mac user who is really interested in running your software. When do you plan to make an OS X version??
Customer Service: ?Unfortunately, we have no such plans, sir.?
Customer: ?We?ll I?m really sorry to hear that. May I ask why not??
Customer Service: ?We decided to discontinue our Mac development a few months after Apple came out with Mac OS 10.3, which features great Windows emulation. We have certified our Windows product to run with full compatibility in Mac OS 10.3.?
Customer: ?Thanks so much for your time.?
If the Mac came out of the box with nearly perfect, adequately fast, transparent Windows support OS X development would die. If that were to happen, the Mac users who had not yet jumped ship would be forced to run emulated Windows applications much of the time. Mac sales would plummet through the ground, because no one would have any reason to buy a Mac at that point. (Are you going to purchase a premium-priced Mac to emulate Windows software on, or are you going to buy a cheap, much faster Dell?) I know that if I were a developer and my Windows binaries ran well emulated in OS X, I wouldn?t be able to justify OS X development. The only reason we have any third party software is because we?re barely relevant (and becoming increasingly irrelevant due to shrinking market share) and we?re a vocal group. If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
</strong><hr></blockquote>
... yeah, especially after the OS/2 lesson, it's virtually certain that the temptation to encapsulate Windows is a trap, a siren call to the rocks ...
But there's one difference between the IBM OS/2 situation and the current Apple one, I have no idea if it's significant, so I'll let others decide.
Apple constrols the hardware. Perhaps, by integrating so much better "Digital Lifestyle" type ports (FireWire) and software, that for people to bother trying to do the same thing on windows becomes just too much of a hassle, and it's just cheaper and less of a headache to just buy the Mac.
Once again, this is a very loose argument, it depends on that case actually being true, and Apple being able to effectively market that truth past MS FUD ...
Unless Apple has some sort of Killer value proposition well beyond the half fleshed out escape from the OS/2 trap as mentioned above, I think, if Apple tried the OS/2 road ... they'd probably be making a big mistake.
What if the 970's 64bit abilities put it so far past the x86 architecture, that if Apple decided to include a very powerful OS/2 type emulator for 32bit Windows in OSX ... it would still force any developer who wanted to write killer 64bit version of their program to write it to OSX, rather than to the 32bit Windows emulator?
Thus, if you buy a 970.
1 - all your old 32bit Windows software will still work.
2 - but any new killer 64bit software, can't be written for the emulator, and must be written to OSX native, and will kick ass.
Thus OSX + the 970 becomes a nice migration path for Windows users to go 64bit, and switch at the same time.
As to whether this will benefit Apple more, than it will hurt Apple by the effects of the, as earlier mentioned "OS/2 Trap" is something better speculated, than tested in the real world at this point.
If the Mac came out of the box with nearly perfect, adequately fast, transparent Windows support OS X development would die. If that were to happen, the Mac users who had not yet jumped ship would be forced to run emulated Windows applications much of the time. Mac sales would plummet through the ground, because no one would have any reason to buy a Mac at that point. (Are you going to purchase a premium-priced Mac to emulate Windows software on, or are you going to buy a cheap, much faster Dell?) I know that if I were a developer and my Windows binaries ran well emulated in OS X, I wouldn?t be able to justify OS X development. The only reason we have any third party software is because we?re barely relevant (and becoming increasingly irrelevant due to shrinking market share) and we?re a vocal group. If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
</strong><hr></blockquote>
The key here to me is the term "adequately fast". OS/2's problem was that there was no real speed difference between its native OS/2 and emulated Windows environments, since they both ran natively on x86 hardware - so indeed there was no real advantage (at least superficially) for using native OS/2 software.
Emulating the x86 environment on the PPC will of necessity incur a performance penalty. If the 970 offers high enough performance that emulated x86 is as fast as current x86 hardware, the native PPC code will be substantially faster. I have a hard time believing that people would be content running emulated Windows programs, when they know an OS X native version would run significantly (20-50%) faster. I may put up with it for the current version of PhotoShop that I own, but for darn sure the next version would be the OS X version.
If half of what David M says comes true, we'll all be in Apple heaven sometime by 2004. :-)
My only add to David's comments would be for the iBriq to be silent as well (a la Cube). I do not *need* the cube shape, but I do want a no-compromise headless machine that is ___silent___ (for media room).
Other than that, I 110% agree with David M's suggestions! :-)
The thing about the cube was that it wasted so much space. 3 inches of the Cube's hight was wasted by a clear base. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> It would be nice to have a modular cube, like the Linux Briq, that could be used for processor intensive tasks that don't require the 4 hard drive support and 66-mhz, 64-bit PCI slots that the xServe offers.
As I said before, I believe OS X x86 is kept up-to-date with the PPC binary and that it is a last resort contingency option. Here's why Apple will not announce and market it:
</strong>
Definitely, OS X is in ready shape. With Motorola as partner, it would be irresponsible not to have a contingency plan. However, it's also a wise engineering decision to find subtle bugs (which may be missed on PPC) and keeping options open (which Steve would appreciate).
<strong>
1) Where are the apps? [...] The Mac applications on the market WILL NOT run on a Wintel box
</strong>
Wrong. I would bet Apple has taken pains to ensure their API's are platform (CPU) independent. NeXT was doing that over a decade ago, and running on Intel would require it (see your point above). Otherwise, even the Apple apps would be prone to breakage. Granted, a few API oversights probably exist, but overall Apple *needs* the API's to be CPU architecture independent for the contingency strategy you outline above.
IOW, the API's will largely support the current software with a recompile. Various apps may have problems introduced with the transition, but those should be vendor specific code --- which would need fixing before shipping app fat for both PPC and Intel OS X boxes. NeXT already did fat binaries as well, so that's a proven technology which Apple already has (they probably use it regularly to run apps on Marklar for testing, verification).
<strong>
2) It would kill Apple?s hardware
</strong>
I suspect this is the issue at the forefront of Apple's current strategy. Yes, this is a key issue. I wonder what happens if Apple's hardware is so superior that this issue is moot, though. Then, what does Apple do?
With the prospect of PPC970, this question becomes interesting. Would Apple then feel comfortable releasing OS X Intel? If so, prospective customers looking to switch could do so without the huge barrier presented by (new Apple) hardware costs. Instead, they could deploy OS X on their existing hardware, and phase in new Apple hardware on their own schedule as desired.
Apple gains potentially huge numbers of customers, serious revenue from OS X Intel sales in large numbers, and (in long run) gains a large base of upgrade customers to Apple hardware.
OS X Intel could transform Apple almost overnight.
<strong>
If you wish to ignore the first reason, and you believe OS X x86 could be a success, you have to ask yourself why Apple would do it.
</strong>
Whoops, already answered that. See above, i.e., to gain easy OS X revenue, convert large numbers of corporate users, and gain future hardware/software customers (not to mention grow their customer base to attract 3P software development).
<strong>
3) It would panic the Mac community
</strong>
Can't follow that line of thought. I bet a huge number of prospective corporate customers would be relieved that Apple has a growth strategy. They would feel much more comfortable using existing hardware rather than betting the farm on all new hardware (at great expense). So, I see comfort going up, though it would be fun to read all the misperceptions in media until the reality set in (:-).
<strong>
4) How to Support PCs
</strong>
Good point. I assume two strategies. 1) Apple would support hardware selectively. In GPU (graphics), OS X already requires latest generation hardware, so that removes many generations of ugly hardware (except in VGA mode or something). So OS X itself requires certain calibre hardware. 2) OS X Intel numbers would be quite large ($$$). Marklar doesn't really cost Apple anything more than it currently does. Apple has a good device driver architecture. Apple would need to use that architecture and spend some of it's Intel revenue to support hardware (that big customers ask it to support). Also, there are now very few major hardware vendors out there. The Intel market has been consolidating. Apple could focus on supporting HP, Dell, and IBM hardware offerings, and other stuff could be supported 3P or with additional effort. The really needy customers would find a way to get the hardware, and Apple would be supporting most corporate customers already by supporting the big 3.
This would cost $, but probably in small proportion to the revenue Apple gained from OS X Intel sales. It would be an ongoing cost though, until h/w vendors helped out as OS X gained market share (and h/w vendor customers demanded such support direct).
<strong>
5) If successful, MS would go nuts If the product were successful, MS could quickly obliterate Apple.
</strong>
I would hope not, but this is a real point. Even if MS chose not to obliterate Apple, they could control their destiny in some ways. If the MS control (in general marketplace) continues to erode, however, even MS would be unable to stop the Linux/IBM/Apple juggernaut.
I guess the key here is timing. Apple needs to be fully committed to this strategy before employing it. There won't be much room for backpedaling once Apple unleashes OS X Intel. Apple will need to be serious and committed, since it's such a high risk/high return strategy.
<strong>
Why Mac OS X + Integrated Win32 Box won?t come out
</strong>
You've given me a cool idea. What if Apple were to release a custom, headless Intel box which talked directly to Mac/OS X through firewire 800 (i.e., networking, for all graphics, h/d access, and keyboard/mouse) to provide a PC compatibility? The box could be cheap (cheaper even than Xbox since it would contain no hard drive), and could be speedy silent and compact. Apple could sell it as optional equipment. It could run Windows (for backwards compatibility). It wouldn't be great for games, but business could love it. Apple could tie it as tightly as they want to Mac OS X using Windows drivers, etc.. The Apple box would give it it's display, networking, keyboard/mouse, power (?), etc.. The Windows box would show up as a window (or variety of windows a la VirtualPC). That could be a nice optional piece of equipment that would solve the VPC void created by MS's recent purchase. The box would be relatively cheap and easy to upgrade, and Apple could equip it with all the Apple stuff Apple desires (e.g., FW800, Rendezvous, etc.). Hmm...
<strong>
If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
</strong>
Moot? <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> This issue was interesting even before the latest IBM 2.5GHz announcement. It's been interesting for years. It just continues to get more interesting as MS's monopoly grip weakens, MS licensing and security issues alienate well-heeled corporations and (!) governments, Apple's hardware begins standing on its own merits, and Apple's OS X matures into a genuine value add for corporate America and other large organizations (with Linux's complementary help). I just wonder when the time will come.
<strong>The thing about the cube was that it wasted so much space. 3 inches of the Cube's hight was wasted by a clear base. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> It would be nice to have a modular cube, like the Linux Briq, that could be used for processor intensive tasks that don't require the 4 hard drive support and 66-mhz, 64-bit PCI slots that the xServe offers.</strong><hr></blockquote>
So true. The only thing it needs is support for a full size AGP slot/grafix card.
And im not talking nVidia FX which needs an extra PCI place for a hairblower!
The exact same quote appeared in their article of last year as I pointed out on the MacRumors comment system. Don't worry - it's cut and paste journalism, not a change of plans (as far as we know).
Does Apple use those non-fan, thermoelectric cooling modules in their machines? It would be nice if they did. They are silent, power-efficient, and small. Would be nice in the PB's, or the iBooks for that matter.
The exact same quote appeared in their article of last year as I pointed out on the MacRumors comment system. Don't worry - it's cut and paste journalism, not a change of plans (as far as we know).</strong>
<hr></blockquote>
Nice catch. I was hoping it was something of the sort but didn't know where to look.
Your regularly scheduled 970 anticipation can now return to normal.
The article also states the 1.8 - 2.5GHz range for the 970 AFTER it migrates to a 0.09µm process.
Can't seem to access the IBM press release any more, but I thought it mentioned the higher range still using a 0.13µm process.</strong><hr></blockquote>
<strong>Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.
Well Intel is doing quite nicely making all those processors for the PC market. I thnk IBM may be prepared to invest in the hope that a good chip will lift Apples share of the market.
[quote]Originally posted by GardenOfEarthlyDelights:
<strong>Beats me what they do with Pentiums. It's already designed-- they just have to pack sand, er, so to speak.</strong><hr></blockquote>
I'm not talking about designing processors, I'm talking about the actual time it takes to manufacture a processor.
[quote]Originally posted byMacJedai:
<strong>And "Rickag", I tend to agree with "Rhumgod", because the 970 production process testing was half way completed in early November 2002 on 200mm and 300mm wafers.</strong><hr></blockquote>
I'm not talking about designing processors, I'm talking about the actual time it takes to manufacture a processor.
Obviously my communication skills are very lacking. Let me start over.
#1. Transcendental Octothorpe stated,"PPC970 production starting in MARCH."
#2 Based on Transcendental Octothorpe's statement, I said that since it takes about 60 days to manufacture a cpu:
-if IBM starts making 970's on March 1st the first finished cpu's will come off the production floor around June.
-if IBM starts making 970's on March 31st the first finished cpu's will come off the production floor around July.
#3. Then an inventory must be built up before shipment/production by Apple. That would indicate what a July to August introduction?
Because of subsequent posts, I'm now confused. How long does it take to manufacture a cpu. I'm not talking about designing the cpu. And published literature states unequivocally that it took 90 DAYS, around the clock, 24 hours a day to manufacture a pentium processor. THEY were not talking about designing, ramping up, tweeking the design they were talking about manufacturing the cpu.
I guess my point is, that even if Transcendental Octothorpe is correct, then it will be months before Apple will even announce an Apple computer that contains a 970, and shipping products 5-6 weeks after the announcement which is now a hallmark of Apple computer.
Big Mac : Great post. I agree with a lot of it. I used to think that really great IA-32 emulation and integration built in to OS X could work, but your points made good sense and I don't think I'd support that idea anymore.
[quote] Originally posted by Over Toasty:
What if the 970's 64bit abilities put it so far past the x86 architecture, that if Apple decided to include a very powerful OS/2 type emulator for 32bit Windows in OSX ... it would still force any developer who wanted to write killer 64bit version of their program to write it to OSX, rather than to the 32bit Windows emulator? <hr></blockquote>
I disagree. I have large doubts that software developers would adopt 64-bits for such a small platform if that platform can emulate their optimized 32-bit x86 version of the software. It won't cost them a penny more and they can save an assload by halting OS X development, but still making it available for OS X, running under emulation. Then they can then concentrate all of their effort on IA-32 and cover all bases. When they already have a good product for 95% of the market that runs fine in 32-bits, why put money into making it that much better for 3% of the market with 64-bit code. You don't have to believe me, but think for a bit about what _developers_ would do (and what would be in the developors' best interests) and you might come to the same conclusion. Maybe not, but I think my point holds SOME water.
[quote] As to whether this will benefit Apple more, than it will hurt Apple by the effects of the, as earlier mentioned "OS/2 Trap" is something better speculated, than tested in the real world at this point. <hr></blockquote>
That pretty much holds true for all of this speculation, but by examining similar events in the past and working out _logical_ scenarios, I think prediction of what would happen in the real world becomes a bit more clear.
Comments
Sources are saying Apple is a customer.
But...not until next(?!) year?
Tell me...say it aint so...
Lemon Bon Bon
<hr></blockquote>
Boo-hoo?
Lemon Bon Bon
<strong>
Dedicated Mac users generally have little use for Windows emulation. Its main value would be in enticing switchers by lowering the "entry fee" to getting onto OS X.
[ 02-28-2003: Message edited by: TJM ]</strong><hr></blockquote>
I'm a dedicated Mac user, but I also have to earn aliving from people who are dedicated Windows users - I'm an IT consultant/project manager/cook/bottle washer - so I need to be able to run Project (FastTrack doesn't do everything I want) and a couple of other bits and pieces.
VPC or another emulator would not just be for making life easier for switchers - it has the effect of turning the Mac into the digital equivalent of the Swiss Army Knife or Leatherman tool.
Why OS X x86 Won't Come Out
As I said before, I believe OS X x86 is kept up-to-date with the PPC binary and that it is a last resort contingency option. Here's why Apple will not announce and market it:
1) Where are the apps? Here's an obvious issue. Customer puts OS X x86 on a PC and marvels at Aqua. He or she browses with Safari and writes something in TextEdit. Okay, Customer thinks, "OS X is a nice OS. Now what?" What applications will the customer run outside of those bundled with the OS? Where are the third party apps that so many rely on? Where's Photoshop or Flash MX? Where are the games? What can be done with this environment? The Mac applications on the market WILL NOT run on a Wintel box Of course, shareware people would pretty quickly recompile for Intel. And if there were incentive, some third parties would provide an Intel binary. But there wouldn't be any such incentive. The Mac's third party support is limited enough, so releasing OS X x86 with virtually no third party support would be a disaster.
2) It would kill Apple?s hardware If you wish to ignore the first reason, and you believe OS X x86 could be a success, you have to ask yourself why Apple would do it. It would obviously hurt its hardware division, as others have stated. If OS X x86 were a viable product, people would favor it because PC hardware will always be cheaper than Apple?s. Apple would have to price the OS far too high in order to make up lost revenue from hardware cannibalization. OS X x86 would kill Apple hardware, and I don?t think Apple?s prepared to do that.
3) It would panic the Mac community No matter what one says, a highly publicized, consumer version of OS X x86 would panic the Mac community. We would legitimately fear that Apple could plan to move to x86 fully. Especially if OS X x86 took off, that fear would ominously cloud the Mac?s future.
4) How to Support PCs A significant amount of Windows development costs goes to trying to support the myriad of PC variants out there. The PC world is a jumbled mess -- just look at how many different motherboards there are. Moreover, Apple has enough trouble supporting the comparatively few designs it produces. Now you can avoid this trap by suggesting only certain machines will be supported, but if you do you still have to defend your argument against all the other issues.
5) If successful, MS would go nuts If the product were successful, MS could quickly obliterate Apple. MS was able to evade the government yet again, so finally taking out Apple wouldn?t be too risky, especially if MS felt vulnerable.
Why Mac OS X + Integrated Win32 Box won?t come out
I have fewer objections to the scheme above, but I maintain it?s just as infeasible as OS X x86. It seems like a great idea ostensibly, though. What Mac user wouldn?t want to be able to run the occasional Windows application in a transparent, Classic-like environment? And with the 970, we?d have emulation speed that would be worth writing home about. Yet, the one major problem is the OS/2 example. IBM created an OS far superior to Windows that ran Windows software with nearly full speed and compatibility. So why did OS/2 die? It?s because it ran Windows software so well that no one bothered to code OS/2 applications!
Many developers barely have the desire to code for the Mac anyway, and now we?d be providing a really great incentive to halt development totally. Consider the following scenario:
Customer: ?Hi, I?m a Mac user who is really interested in running your software. When do you plan to make an OS X version??
Customer Service: ?Unfortunately, we have no such plans, sir.?
Customer: ?We?ll I?m really sorry to hear that. May I ask why not??
Customer Service: ?We decided to discontinue our Mac development a few months after Apple came out with Mac OS 10.3, which features great Windows emulation. We have certified our Windows product to run with full compatibility in Mac OS 10.3.?
Customer: ?Thanks so much for your time.?
If the Mac came out of the box with nearly perfect, adequately fast, transparent Windows support OS X development would die. If that were to happen, the Mac users who had not yet jumped ship would be forced to run emulated Windows applications much of the time. Mac sales would plummet through the ground, because no one would have any reason to buy a Mac at that point. (Are you going to purchase a premium-priced Mac to emulate Windows software on, or are you going to buy a cheap, much faster Dell?) I know that if I were a developer and my Windows binaries ran well emulated in OS X, I wouldn?t be able to justify OS X development. The only reason we have any third party software is because we?re barely relevant (and becoming increasingly irrelevant due to shrinking market share) and we?re a vocal group. If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
[ 03-01-2003: Message edited by: Big Mac ]</p>
<strong>
If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
</strong><hr></blockquote>
... yeah, especially after the OS/2 lesson, it's virtually certain that the temptation to encapsulate Windows is a trap, a siren call to the rocks ...
But there's one difference between the IBM OS/2 situation and the current Apple one, I have no idea if it's significant, so I'll let others decide.
Apple constrols the hardware. Perhaps, by integrating so much better "Digital Lifestyle" type ports (FireWire) and software, that for people to bother trying to do the same thing on windows becomes just too much of a hassle, and it's just cheaper and less of a headache to just buy the Mac.
Once again, this is a very loose argument, it depends on that case actually being true, and Apple being able to effectively market that truth past MS FUD ...
Unless Apple has some sort of Killer value proposition well beyond the half fleshed out escape from the OS/2 trap as mentioned above, I think, if Apple tried the OS/2 road ... they'd probably be making a big mistake.
What if the 970's 64bit abilities put it so far past the x86 architecture, that if Apple decided to include a very powerful OS/2 type emulator for 32bit Windows in OSX ... it would still force any developer who wanted to write killer 64bit version of their program to write it to OSX, rather than to the 32bit Windows emulator?
Thus, if you buy a 970.
1 - all your old 32bit Windows software will still work.
2 - but any new killer 64bit software, can't be written for the emulator, and must be written to OSX native, and will kick ass.
Thus OSX + the 970 becomes a nice migration path for Windows users to go 64bit, and switch at the same time.
As to whether this will benefit Apple more, than it will hurt Apple by the effects of the, as earlier mentioned "OS/2 Trap" is something better speculated, than tested in the real world at this point.
</mooting>
<strong>
If the Mac came out of the box with nearly perfect, adequately fast, transparent Windows support OS X development would die. If that were to happen, the Mac users who had not yet jumped ship would be forced to run emulated Windows applications much of the time. Mac sales would plummet through the ground, because no one would have any reason to buy a Mac at that point. (Are you going to purchase a premium-priced Mac to emulate Windows software on, or are you going to buy a cheap, much faster Dell?) I know that if I were a developer and my Windows binaries ran well emulated in OS X, I wouldn?t be able to justify OS X development. The only reason we have any third party software is because we?re barely relevant (and becoming increasingly irrelevant due to shrinking market share) and we?re a vocal group. If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
</strong><hr></blockquote>
The key here to me is the term "adequately fast". OS/2's problem was that there was no real speed difference between its native OS/2 and emulated Windows environments, since they both ran natively on x86 hardware - so indeed there was no real advantage (at least superficially) for using native OS/2 software.
Emulating the x86 environment on the PPC will of necessity incur a performance penalty. If the 970 offers high enough performance that emulated x86 is as fast as current x86 hardware, the native PPC code will be substantially faster. I have a hard time believing that people would be content running emulated Windows programs, when they know an OS X native version would run significantly (20-50%) faster. I may put up with it for the current version of PhotoShop that I own, but for darn sure the next version would be the OS X version.
My only add to David's comments would be for the iBriq to be silent as well (a la Cube). I do not *need* the cube shape, but I do want a no-compromise headless machine that is ___silent___ (for media room).
Other than that, I 110% agree with David M's suggestions! :-)
<strong>
Why OS X x86 Won't Come Out
</strong>
Good topic!
<strong>
As I said before, I believe OS X x86 is kept up-to-date with the PPC binary and that it is a last resort contingency option. Here's why Apple will not announce and market it:
</strong>
Definitely, OS X is in ready shape. With Motorola as partner, it would be irresponsible not to have a contingency plan. However, it's also a wise engineering decision to find subtle bugs (which may be missed on PPC) and keeping options open (which Steve would appreciate).
<strong>
1) Where are the apps? [...] The Mac applications on the market WILL NOT run on a Wintel box
</strong>
Wrong. I would bet Apple has taken pains to ensure their API's are platform (CPU) independent. NeXT was doing that over a decade ago, and running on Intel would require it (see your point above). Otherwise, even the Apple apps would be prone to breakage. Granted, a few API oversights probably exist, but overall Apple *needs* the API's to be CPU architecture independent for the contingency strategy you outline above.
IOW, the API's will largely support the current software with a recompile. Various apps may have problems introduced with the transition, but those should be vendor specific code --- which would need fixing before shipping app fat for both PPC and Intel OS X boxes. NeXT already did fat binaries as well, so that's a proven technology which Apple already has (they probably use it regularly to run apps on Marklar for testing, verification).
<strong>
2) It would kill Apple?s hardware
</strong>
I suspect this is the issue at the forefront of Apple's current strategy. Yes, this is a key issue. I wonder what happens if Apple's hardware is so superior that this issue is moot, though. Then, what does Apple do?
With the prospect of PPC970, this question becomes interesting. Would Apple then feel comfortable releasing OS X Intel? If so, prospective customers looking to switch could do so without the huge barrier presented by (new Apple) hardware costs. Instead, they could deploy OS X on their existing hardware, and phase in new Apple hardware on their own schedule as desired.
Apple gains potentially huge numbers of customers, serious revenue from OS X Intel sales in large numbers, and (in long run) gains a large base of upgrade customers to Apple hardware.
OS X Intel could transform Apple almost overnight.
<strong>
If you wish to ignore the first reason, and you believe OS X x86 could be a success, you have to ask yourself why Apple would do it.
</strong>
Whoops, already answered that. See above, i.e., to gain easy OS X revenue, convert large numbers of corporate users, and gain future hardware/software customers (not to mention grow their customer base to attract 3P software development).
<strong>
3) It would panic the Mac community
</strong>
Can't follow that line of thought. I bet a huge number of prospective corporate customers would be relieved that Apple has a growth strategy. They would feel much more comfortable using existing hardware rather than betting the farm on all new hardware (at great expense). So, I see comfort going up, though it would be fun to read all the misperceptions in media until the reality set in (:-).
<strong>
4) How to Support PCs
</strong>
Good point. I assume two strategies. 1) Apple would support hardware selectively. In GPU (graphics), OS X already requires latest generation hardware, so that removes many generations of ugly hardware (except in VGA mode or something). So OS X itself requires certain calibre hardware. 2) OS X Intel numbers would be quite large ($$$). Marklar doesn't really cost Apple anything more than it currently does. Apple has a good device driver architecture. Apple would need to use that architecture and spend some of it's Intel revenue to support hardware (that big customers ask it to support). Also, there are now very few major hardware vendors out there. The Intel market has been consolidating. Apple could focus on supporting HP, Dell, and IBM hardware offerings, and other stuff could be supported 3P or with additional effort. The really needy customers would find a way to get the hardware, and Apple would be supporting most corporate customers already by supporting the big 3.
This would cost $, but probably in small proportion to the revenue Apple gained from OS X Intel sales. It would be an ongoing cost though, until h/w vendors helped out as OS X gained market share (and h/w vendor customers demanded such support direct).
<strong>
5) If successful, MS would go nuts If the product were successful, MS could quickly obliterate Apple.
</strong>
I would hope not, but this is a real point. Even if MS chose not to obliterate Apple, they could control their destiny in some ways. If the MS control (in general marketplace) continues to erode, however, even MS would be unable to stop the Linux/IBM/Apple juggernaut.
I guess the key here is timing. Apple needs to be fully committed to this strategy before employing it. There won't be much room for backpedaling once Apple unleashes OS X Intel. Apple will need to be serious and committed, since it's such a high risk/high return strategy.
<strong>
Why Mac OS X + Integrated Win32 Box won?t come out
</strong>
You've given me a cool idea. What if Apple were to release a custom, headless Intel box which talked directly to Mac/OS X through firewire 800 (i.e., networking, for all graphics, h/d access, and keyboard/mouse) to provide a PC compatibility? The box could be cheap (cheaper even than Xbox since it would contain no hard drive), and could be speedy silent and compact. Apple could sell it as optional equipment. It could run Windows (for backwards compatibility). It wouldn't be great for games, but business could love it. Apple could tie it as tightly as they want to Mac OS X using Windows drivers, etc.. The Apple box would give it it's display, networking, keyboard/mouse, power (?), etc.. The Windows box would show up as a window (or variety of windows a la VirtualPC). That could be a nice optional piece of equipment that would solve the VPC void created by MS's recent purchase. The box would be relatively cheap and easy to upgrade, and Apple could equip it with all the Apple stuff Apple desires (e.g., FW800, Rendezvous, etc.). Hmm...
<strong>
If anyone has any problems with the logic contained herein, please feel free to challenge it. Otherwise, let?s not waste time on a moot issue.
</strong>
Moot? <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> This issue was interesting even before the latest IBM 2.5GHz announcement. It's been interesting for years. It just continues to get more interesting as MS's monopoly grip weakens, MS licensing and security issues alienate well-heeled corporations and (!) governments, Apple's hardware begins standing on its own merits, and Apple's OS X matures into a genuine value add for corporate America and other large organizations (with Linux's complementary help). I just wonder when the time will come.
[ 03-01-2003: Message edited by: occam ]
[ 03-01-2003: Message edited by: occam ]</p>
<strong>The thing about the cube was that it wasted so much space. 3 inches of the Cube's hight was wasted by a clear base. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> It would be nice to have a modular cube, like the Linux Briq, that could be used for processor intensive tasks that don't require the 4 hard drive support and 66-mhz, 64-bit PCI slots that the xServe offers.</strong><hr></blockquote>
So true. The only thing it needs is support for a full size AGP slot/grafix card.
And im not talking nVidia FX which needs an extra PCI place for a hairblower!
<strong>
Boo-hoo?
Lemon Bon Bon
The exact same quote appeared in their article of last year as I pointed out on the MacRumors comment system. Don't worry - it's cut and paste journalism, not a change of plans (as far as we know).
<strong>
The exact same quote appeared in their article of last year as I pointed out on the MacRumors comment system. Don't worry - it's cut and paste journalism, not a change of plans (as far as we know).</strong>
<hr></blockquote>
Nice catch. I was hoping it was something of the sort but didn't know where to look.
Your regularly scheduled 970 anticipation can now return to normal.
<strong>
Boo-hoo?
Lemon Bon Bon
The article also states the 1.8 - 2.5GHz range for the 970 AFTER it migrates to a 0.09µm process.
Can't seem to access the IBM press release any more, but I thought it mentioned the higher range still using a 0.13µm process.
<strong>
The article also states the 1.8 - 2.5GHz range for the 970 AFTER it migrates to a 0.09µm process.
Can't seem to access the IBM press release any more, but I thought it mentioned the higher range still using a 0.13µm process.</strong><hr></blockquote>
It DID say that...
<strong>Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.
[ 02-28-2003: Message edited by: Anonymous Karma ]</strong><hr></blockquote>
Well Intel is doing quite nicely making all those processors for the PC market. I thnk IBM may be prepared to invest in the hope that a good chip will lift Apples share of the market.
<strong>Beats me what they do with Pentiums. It's already designed-- they just have to pack sand, er, so to speak.</strong><hr></blockquote>
I'm not talking about designing processors, I'm talking about the actual time it takes to manufacture a processor.
[quote]Originally posted byMacJedai:
<strong>And "Rickag", I tend to agree with "Rhumgod", because the 970 production process testing was half way completed in early November 2002 on 200mm and 300mm wafers.</strong><hr></blockquote>
I'm not talking about designing processors, I'm talking about the actual time it takes to manufacture a processor.
Obviously my communication skills are very lacking. Let me start over.
#1. Transcendental Octothorpe stated,"PPC970 production starting in MARCH."
#2 Based on Transcendental Octothorpe's statement, I said that since it takes about 60 days to manufacture a cpu:
-if IBM starts making 970's on March 1st the first finished cpu's will come off the production floor around June.
-if IBM starts making 970's on March 31st the first finished cpu's will come off the production floor around July.
#3. Then an inventory must be built up before shipment/production by Apple. That would indicate what a July to August introduction?
Because of subsequent posts, I'm now confused. How long does it take to manufacture a cpu. I'm not talking about designing the cpu. And published literature states unequivocally that it took 90 DAYS, around the clock, 24 hours a day to manufacture a pentium processor. THEY were not talking about designing, ramping up, tweeking the design they were talking about manufacturing the cpu.
I guess my point is, that even if Transcendental Octothorpe is correct, then it will be months before Apple will even announce an Apple computer that contains a 970, and shipping products 5-6 weeks after the announcement
[ 03-01-2003: Message edited by: rickag ]</p>
<strong>
It DID say that...
Just seconding this.
IBM's own release said '1.8-2.5 GHz' and '0.13um.' YMMV.
I'll say anywhere between 275,000 to 325,000 (there's a lot of built up "im waiting for the G5" purchases)
[quote] Originally posted by Over Toasty:
What if the 970's 64bit abilities put it so far past the x86 architecture, that if Apple decided to include a very powerful OS/2 type emulator for 32bit Windows in OSX ... it would still force any developer who wanted to write killer 64bit version of their program to write it to OSX, rather than to the 32bit Windows emulator? <hr></blockquote>
I disagree. I have large doubts that software developers would adopt 64-bits for such a small platform if that platform can emulate their optimized 32-bit x86 version of the software. It won't cost them a penny more and they can save an assload by halting OS X development, but still making it available for OS X, running under emulation. Then they can then concentrate all of their effort on IA-32 and cover all bases. When they already have a good product for 95% of the market that runs fine in 32-bits, why put money into making it that much better for 3% of the market with 64-bit code. You don't have to believe me, but think for a bit about what _developers_ would do (and what would be in the developors' best interests) and you might come to the same conclusion. Maybe not, but I think my point holds SOME water.
[quote] As to whether this will benefit Apple more, than it will hurt Apple by the effects of the, as earlier mentioned "OS/2 Trap" is something better speculated, than tested in the real world at this point. <hr></blockquote>
That pretty much holds true for all of this speculation, but by examining similar events in the past and working out _logical_ scenarios, I think prediction of what would happen in the real world becomes a bit more clear.