For some reason, it seems that nobody ever brings up this point regarding Rosetta:
Only raw PowerPC code will have to be emulated, all API calls to OS X will be executed at full native intel speed.
For example if a PPC program ask OS X to draw a window, then use some Quartz drawing commands to paint its content, those parts will be run at full speed. Quicktime, Quartz, the file system, Core API's, text services etc will all run natively when called from a PPC app. That's how Rosetta can achieve %80 speed for most apps, while PearPC can run at like 5-20% (?) since it has to emulate everything.
It's kind of like the 68k-PPC transition, but with a big difference. The first PPC machine shipped with some version of System 7.1 that was still 80% 68k code. It took Apple years, until Mac OS 9, to reach a point where the whole Mac OS was PPC native. So in the days of System 7-8, emulated apps rarely had the chance to call native PPC API's.
The intel version of OS X is already %100 intel code, and if it's not it will be very soon.
PowerPC apps that won't run well in Rosetta are the usual suspects, games, 3d apps, pro apps in general. Apps that don't rely much on raw PPC code will run surprisingly fast.
The creative applications may not emulate as well as they could, but graphic artists rarely use laptops.
Quote:
Originally posted by wilco
Wrong.
My partner works for a graphic design company and they use Macs. The first thing you'll notice is that they all use CRT monitors - flat screens don't have the same colour quality for matching with printed results. The second thing is that the where they use a laptop (for working with a client) it's a big screen laptop. The other problem with a laptop is the greater difficulty of manipulating graphics, and the lack of drawing instrument.
Of course different companies do different things, and I don't know how indicative her company is of other graphics companies - but the CRT thing is important... hence my thought that a 15" powerbook is quite possible in january (whereas the 17" may be held).
YMMV, of course. In fact, I'd like you to tell me your own experiences of how graphic designers use their computers, it would add to my knowledge. Care to share?
I'm not saying that they wouldn't [use separate apps for db & spreadsheet]. I'm just thinking about cost, and time to market. Unless they took the spreadsheet from Appleworks, and expanded it, they would have to either come up with one from scratch, or buy one (which Apple has done many times before). But which one?
In terms of cost & time-to-market, I think the spreadsheet in iWorks 06 will be relatively sophisticated, while the database (if any) would be relatively simple. Take some of the spreadsheet/reporting tools, slap them on a Core Data foundation, and voila! Very simple, very easy.
[Edit]
It also conveniently coerces users to upgrade to Tiger to get all the functionality. Apple is good at doing that.
[/Edit]
They have to give us something to look forward to in iWorks 07. That might be something more like a FileMaker Lite and more fully functional.
"I didn't say that Filemaker, as it is, should be used as a spreadsheet, but it could easily be, with some additions."
That's very different. I didn't say that "users shouldn't use it that way", meaning that they shouldn't use the program as a spreadsheet AT ALL. I said that with some additions, it could be used that way.
We may not agree on some things, but I grant you the intellegence to understand that.
My use of "as in" before each section of text that was between quote marks meant that I was not quoting you, but rather defining what "should/shouldn't" meant in each context.
For example:
-- read, as in "I read the book yesterday"
-- read, as in "I read the book every day"
Those two examples are not quoting anybody; they're defining the word "read" in each case (pronounced "red" and "reed", respectively). That's what I was doing for "should/shouldn't"--I was defining the terms.
Only raw PowerPC code will have to be emulated, all API calls to OS X will be executed at full native intel speed.
The problem is that many of the larger applications implement their own frameworks to create the UI or even do things like anti-alias their own text. This is usually done for cross platform compatibility so that the code they run on Windows is the same as the Mac.
The problem therefore is that the application vendors have to now produce intel native versions of their frameworks also. That's either a minor or a major task depending on how they've done it but it does mean that they have a larger task than relying on Apple's native API.
Adobe recently finished their intel port of their UI framework although I don't know how much of it's suite uses it. Perhaps melgross has a better feel for that.
The problem is that many of the larger applications implement their own frameworks to create the UI or even do things like anti-alias their own text. This is usually done for cross platform compatibility so that the code they run on Windows is the same as the Mac.
Exactly. Although Apple told them not to do that, they didn't listen.
But I am amazed that Jobs was able to convince companies like Adobe to write an OS X Intel version of all of their apps. I thought that would never happen. I was sure that Adobe would just say "Your Intel Macs can run Windows - just let your customers run our Windows Photoshop."
Of course that would require all the Mac users to buy Windows - which might have been enough for Jobs and his RDF to convince Adobe that such a plan would result in lost sales.
Either that or Jobs offered to pay for the Intel port. He could always mention that Apple might polish up the Gimp to make it an OS X app.
The problem is that many of the larger applications implement their own frameworks to create the UI or even do things like anti-alias their own text. This is usually done for cross platform compatibility so that the code they run on Windows is the same as the Mac.
The problem therefore is that the application vendors have to now produce intel native versions of their frameworks also. That's either a minor or a major task depending on how they've done it but it does mean that they have a larger task than relying on Apple's native API.
Adobe recently finished their intel port of their UI framework although I don't know how much of it's suite uses it. Perhaps melgross has a better feel for that.
Microsoft I believe has an equivalent framework although maybe not for cross platform compatibility - more to just be annoying IMHO.
Adobe would like to integrate that across their entire suit. I'm not sure how far along they are in doing that. They want to integrate the structure of each program.
Right now, they seem to be in the throes of introspection about the whole thing.
Of course, Apple has been known to do this as well. It would be easier for developers if Apple stayed withing their own guidelines.
Exactly. Although Apple told them not to do that, they didn't listen.
But I am amazed that Jobs was able to convince companies like Adobe to write an OS X Intel version of all of their apps. I thought that would never happen. I was sure that Adobe would just say "Your Intel Macs can run Windows - just let your customers run our Windows Photoshop."
Of course that would require all the Mac users to buy Windows - which might have been enough for Jobs and his RDF to convince Adobe that such a plan would result in lost sales.
Either that or Jobs offered to pay for the Intel port. He could always mention that Apple might polish up the Gimp to make it an OS X app.
I'm not sure that these companies had a choice at this point in time. If Apple manages to up their marketshare with this switch, then it will be a problem if they didn't. A lot of people simply wouldn't upgrade for some time. That would cost them more than the rewriting would. Plus the bad blood that would ensue.
Remember what happened to Avid when they decided to move over to NT, and leave Apple users stuck.
AI's article is nothing more than a recap of the seed notes. OSx86 are not the only ones with access to that info, and they were not the first ones to break the news.
Btw. at the moment the Altivec support in Rosetta makes Photoshop filters much slower.
I'm sure the x86 versions of OS X on the loose has to have serial numbers which could be easily traced back to the developer who let their copy go.
But in reality Apple doesn't seem to be too concerned about it. Through inaction are allowing hackers to break and use OS X.
What makes logical sense to me is that aspects of the current OSx86 and its security have little relation to how it will operate in the actual production version.
Yes, the Pentium M (Yonah/laptop stuff) is the basis of future Intel chips and I think Apple will follow those chips as they move from low-power versions into the desktop over the coming 1-2 years.
I see no evidence for or against Apple not having the software ready for January. The hardware could certainly be done, the rest depends on OSX stability (and emulation). I wouldn't be surprised if Apple released an Intel based Mac-Mini-TV system that officially doesn't run 3rd-party applications AT ALL.
It depends on the speed of emulation. The G5 systems are FAST, but Apple Laptops are not so fast. If the emulation works well a laptop could feasibly be switched to Intel and run apps in emulation better than a G4 today. And I think Apple will either make that their goal, or wait till more apps are ported.
Greg
ps. The creative applications may not emulate as well as they could, but graphic artists rarely use laptops. They also like big screens, and the colour matching isn't available on LCDs. If graphics professionals can still buy a G4 17inch Powerbook and G5 PowerMacs for all of 2006 I think that part of Apple's customer-base will be fine.
True; but then again, I just had a second look, and it appears that some third party companies are already shipping universal binaries or are in the last phase of testing - so I guess things are as bleak as they could be; and hopefully with companies taking advantage of Apples 'optimisation in a framework', we'll see applications take advantage of SSE3 earlier than what occurs in Windows.
What I'd love to see is this, however; a free Jabber based service provided by Apple, something like Apple iJabber or something, and set it up with MSN, AIM and Yahoo transport agents so that people can do away with their multiple clients in favour of using one (iChat).
There's a lot more going on than we ever had with IBM or Moto and Freescale.
Tomshardware probably has the best article of the lot; those 2 and 4 core beasts look great - what I hope, however, is Apple spend alot of time making their operating system mega threaded so that as end users, we reep the rewards of having multi-core processors.
Not that the Tom's Hardware article isn't interesting but it really does say nothing outside of a few codenames, that cores will be added and expected cache. There's a lot more to be said. Even Merom and Conroe have had very few details actually released about the chip's design.
Comments
Only raw PowerPC code will have to be emulated, all API calls to OS X will be executed at full native intel speed.
For example if a PPC program ask OS X to draw a window, then use some Quartz drawing commands to paint its content, those parts will be run at full speed. Quicktime, Quartz, the file system, Core API's, text services etc will all run natively when called from a PPC app. That's how Rosetta can achieve %80 speed for most apps, while PearPC can run at like 5-20% (?) since it has to emulate everything.
It's kind of like the 68k-PPC transition, but with a big difference. The first PPC machine shipped with some version of System 7.1 that was still 80% 68k code. It took Apple years, until Mac OS 9, to reach a point where the whole Mac OS was PPC native. So in the days of System 7-8, emulated apps rarely had the chance to call native PPC API's.
The intel version of OS X is already %100 intel code, and if it's not it will be very soon.
PowerPC apps that won't run well in Rosetta are the usual suspects, games, 3d apps, pro apps in general. Apps that don't rely much on raw PPC code will run surprisingly fast.
Originally posted by GregAlexander
The creative applications may not emulate as well as they could, but graphic artists rarely use laptops.
Originally posted by wilco
Wrong.
My partner works for a graphic design company and they use Macs. The first thing you'll notice is that they all use CRT monitors - flat screens don't have the same colour quality for matching with printed results. The second thing is that the where they use a laptop (for working with a client) it's a big screen laptop. The other problem with a laptop is the greater difficulty of manipulating graphics, and the lack of drawing instrument.
Of course different companies do different things, and I don't know how indicative her company is of other graphics companies - but the CRT thing is important... hence my thought that a 15" powerbook is quite possible in january (whereas the 17" may be held).
YMMV, of course. In fact, I'd like you to tell me your own experiences of how graphic designers use their computers, it would add to my knowledge. Care to share?
Originally posted by melgross
I'm not saying that they wouldn't [use separate apps for db & spreadsheet]. I'm just thinking about cost, and time to market. Unless they took the spreadsheet from Appleworks, and expanded it, they would have to either come up with one from scratch, or buy one (which Apple has done many times before). But which one?
In terms of cost & time-to-market, I think the spreadsheet in iWorks 06 will be relatively sophisticated, while the database (if any) would be relatively simple. Take some of the spreadsheet/reporting tools, slap them on a Core Data foundation, and voila! Very simple, very easy.
[Edit]
It also conveniently coerces users to upgrade to Tiger to get all the functionality. Apple is good at doing that.
[/Edit]
They have to give us something to look forward to in iWorks 07.
Originally posted by VL-Tone
Only raw PowerPC code will have to be emulated, all API calls to OS X will be executed at full native intel speed.
Excellent point.
Illustrates the advantage of a robust API such as Carbon, Cocoa and the POSIX calls.
Originally posted by melgross
You're changing my words, and meaning. I said:
"I didn't say that Filemaker, as it is, should be used as a spreadsheet, but it could easily be, with some additions."
That's very different. I didn't say that "users shouldn't use it that way", meaning that they shouldn't use the program as a spreadsheet AT ALL. I said that with some additions, it could be used that way.
We may not agree on some things, but I grant you the intellegence to understand that.
My use of "as in" before each section of text that was between quote marks meant that I was not quoting you, but rather defining what "should/shouldn't" meant in each context.
For example:
-- read, as in "I read the book yesterday"
-- read, as in "I read the book every day"
Those two examples are not quoting anybody; they're defining the word "read" in each case (pronounced "red" and "reed", respectively). That's what I was doing for "should/shouldn't"--I was defining the terms.
Originally posted by VL-Tone
Only raw PowerPC code will have to be emulated, all API calls to OS X will be executed at full native intel speed.
The problem is that many of the larger applications implement their own frameworks to create the UI or even do things like anti-alias their own text. This is usually done for cross platform compatibility so that the code they run on Windows is the same as the Mac.
The problem therefore is that the application vendors have to now produce intel native versions of their frameworks also. That's either a minor or a major task depending on how they've done it but it does mean that they have a larger task than relying on Apple's native API.
Adobe recently finished their intel port of their UI framework although I don't know how much of it's suite uses it. Perhaps melgross has a better feel for that.
http://opensource.adobe.com/
Microsoft I believe has an equivalent framework although maybe not for cross platform compatibility - more to just be annoying IMHO.
The problem is that many of the larger applications implement their own frameworks to create the UI or even do things like anti-alias their own text. This is usually done for cross platform compatibility so that the code they run on Windows is the same as the Mac.
Exactly. Although Apple told them not to do that, they didn't listen.
But I am amazed that Jobs was able to convince companies like Adobe to write an OS X Intel version of all of their apps. I thought that would never happen. I was sure that Adobe would just say "Your Intel Macs can run Windows - just let your customers run our Windows Photoshop."
Of course that would require all the Mac users to buy Windows - which might have been enough for Jobs and his RDF to convince Adobe that such a plan would result in lost sales.
Either that or Jobs offered to pay for the Intel port. He could always mention that Apple might polish up the Gimp to make it an OS X app.
Originally posted by aegisdesign
The problem is that many of the larger applications implement their own frameworks to create the UI or even do things like anti-alias their own text. This is usually done for cross platform compatibility so that the code they run on Windows is the same as the Mac.
The problem therefore is that the application vendors have to now produce intel native versions of their frameworks also. That's either a minor or a major task depending on how they've done it but it does mean that they have a larger task than relying on Apple's native API.
Adobe recently finished their intel port of their UI framework although I don't know how much of it's suite uses it. Perhaps melgross has a better feel for that.
http://opensource.adobe.com/
Microsoft I believe has an equivalent framework although maybe not for cross platform compatibility - more to just be annoying IMHO.
Adobe would like to integrate that across their entire suit. I'm not sure how far along they are in doing that. They want to integrate the structure of each program.
Right now, they seem to be in the throes of introspection about the whole thing.
Of course, Apple has been known to do this as well. It would be easier for developers if Apple stayed withing their own guidelines.
Originally posted by lundy
Exactly. Although Apple told them not to do that, they didn't listen.
But I am amazed that Jobs was able to convince companies like Adobe to write an OS X Intel version of all of their apps. I thought that would never happen. I was sure that Adobe would just say "Your Intel Macs can run Windows - just let your customers run our Windows Photoshop."
Of course that would require all the Mac users to buy Windows - which might have been enough for Jobs and his RDF to convince Adobe that such a plan would result in lost sales.
Either that or Jobs offered to pay for the Intel port. He could always mention that Apple might polish up the Gimp to make it an OS X app.
I'm not sure that these companies had a choice at this point in time. If Apple manages to up their marketshare with this switch, then it will be a problem if they didn't. A lot of people simply wouldn't upgrade for some time. That would cost them more than the rewriting would. Plus the bad blood that would ensue.
Remember what happened to Avid when they decided to move over to NT, and leave Apple users stuck.
Originally posted by curiousuburb
Perhaps if you wanted more information, you could ask the folks who originally broke the stories
Ouch.
Originally posted by curiousuburb
Perhaps if you wanted more information, you could ask the folks who originally broke the stories
Yes, you're right. It was them who I meant when I said that it was published last week.
Originally posted by curiousuburb
Perhaps if you wanted more information, you could ask the folks who originally broke the stories
AI's article is nothing more than a recap of the seed notes. OSx86 are not the only ones with access to that info, and they were not the first ones to break the news.
Btw. at the moment the Altivec support in Rosetta makes Photoshop filters much slower.
Originally posted by curiousuburb
Perhaps if you wanted more information, you could ask the folks who originally broke the stories
It amazes me that Apple's lawyers haven't swooped on that bunch. There's a whole load of NDAs broken there and quite possibly laws too.
I'm sure the x86 versions of OS X on the loose has to have serial numbers which could be easily traced back to the developer who let their copy go.
But in reality Apple doesn't seem to be too concerned about it. Through inaction are allowing hackers to break and use OS X.
What makes logical sense to me is that aspects of the current OSx86 and its security have little relation to how it will operate in the actual production version.
Originally posted by GregAlexander
Yes, the Pentium M (Yonah/laptop stuff) is the basis of future Intel chips and I think Apple will follow those chips as they move from low-power versions into the desktop over the coming 1-2 years.
I see no evidence for or against Apple not having the software ready for January. The hardware could certainly be done, the rest depends on OSX stability (and emulation). I wouldn't be surprised if Apple released an Intel based Mac-Mini-TV system that officially doesn't run 3rd-party applications AT ALL.
It depends on the speed of emulation. The G5 systems are FAST, but Apple Laptops are not so fast. If the emulation works well a laptop could feasibly be switched to Intel and run apps in emulation better than a G4 today. And I think Apple will either make that their goal, or wait till more apps are ported.
Greg
ps. The creative applications may not emulate as well as they could, but graphic artists rarely use laptops. They also like big screens, and the colour matching isn't available on LCDs. If graphics professionals can still buy a G4 17inch Powerbook and G5 PowerMacs for all of 2006 I think that part of Apple's customer-base will be fine.
True; but then again, I just had a second look, and it appears that some third party companies are already shipping universal binaries or are in the last phase of testing - so I guess things are as bleak as they could be; and hopefully with companies taking advantage of Apples 'optimisation in a framework', we'll see applications take advantage of SSE3 earlier than what occurs in Windows.
What I'd love to see is this, however; a free Jabber based service provided by Apple, something like Apple iJabber or something, and set it up with MSN, AIM and Yahoo transport agents so that people can do away with their multiple clients in favour of using one (iChat).
This gives a good look at Intel's plans. It is again encouraging.
I guess that we now have to root for Intel.
http://www.tomshardware.com/cpu/20051203/index.html
and:
http://www.tgdaily.com/2005/12/04/in...005/index.html
and, software as well,
http://www.tgdaily.com/2005/12/02/in...ils/index.html
There's a lot more going on than we ever had with IBM or Moto and Freescale.
Originally posted by melgross
Meanwhile, back at the hardware front, there is this new article from Tom's Hardware.
This gives a good look at Intel's plans. It is again encouraging.
I guess that we now have to root for Intel.
http://www.tomshardware.com/cpu/20051203/index.html
There's a lot more going on than we ever had with IBM or Moto and Freescale.
Tomshardware probably has the best article of the lot; those 2 and 4 core beasts look great - what I hope, however, is Apple spend alot of time making their operating system mega threaded so that as end users, we reep the rewards of having multi-core processors.