Finder is getting rewritten in Cocoa, right? And most other Apple apps are already in Cocoa now, right?
What if....Apple enabled Cocoa to use OpenCL behind the scenes!!!!!!! You write in Cocoa and the Cocoa components are tapping OpenCL making everything insanely...SNAPPIER!
Seriously though, do you know what kind of performance increase that would make for all Cocoa apps???? Windows 7 would have nothing on that!
The point of 10.6 is to replace almost all userspace apps with Cocoa only versions, be Multicore enabled/leveraged and Quartz Extreme on using OpenCL from the low level to the userspace by making the demands on OpenGL offloaded per core whether it's a core on the GPU or a CPU core.
Finder is getting rewritten in Cocoa, right? And most other Apple apps are already in Cocoa now, right?
What if....Apple enabled Cocoa to use OpenCL behind the scenes!!!!!!! You write in Cocoa and the Cocoa components are tapping OpenCL making everything insanely...SNAPPIER!
Seriously though, do you know what kind of performance increase that would make for all Cocoa apps???? Windows 7 would have nothing on that!
This is a definite for Grand Cetral. Which will make attempts at abstracting and automatically handling multiple cores using existing Cocoa programming structures and objects. Which will make all SL Cocoa Apps run faster and more efficient for free! I believe it will be the same case for OpenCL but will require an extension to the API to designate certain tasks as potential processor intensive calculation tasks that are eligible to be offloaded to the GPU if necessary.
If Apple could release Mac OS X 10.6 Snow Leopard at the same time as the new quad-core iMacs in late January, mid-February 2009, the new iMacs would sell like hot cakes!
I like how people always say noooo, develop OS X slower. How do people know how many issues need fixed and optimized and how long they will take? The presentation shows that the Leopard -> Snow Leopard period is 14+ months and some other major releases took this long. 6 releases in 8 years = 16 month average cycle.
Look at Microsoft's efforts. Taking longer doesn't guarantee a better product. Longer times mean that people lose focus on the task.
The focus with Snow Leopard is optimization, possibly removing PPC stuff and compiling for 64-bit. Just because some stuff sounds complicated doesn't mean it takes longer to implement - it's not as if they start working on it as soon as the last version is out the door. The GPU stuff has been going for a while in various incarnations. The multi-threading stuff will take R&D from Xgrid computing and previous multi-threading techniques.
A March release is still 4 months away too. Beta in January, a few betas in Feb, Final version in March.
To be honest, it could have been a typo in the presentation. Q2 is what people were expecting. Nonetheless, I think people should be more open to an early release. It's not as if the people working on it are being given a public deadline and scared that it's not done yet. The release date is not official, right now it's probably a rough target.
Which systems did Apple put out before they were ready? MobileMe perhaps but they know what they are doing most of the time. The simple solution is for Apple to release it when they think it's ready and consumers who know little to nothing of Apple's internal developments who arbitrarily think otherwise can simply wait before using it.
There's an expectation that when a ship date is announced for major software projects, that date will slip. It did for all of Apple's OS releases, except for 10.4, and we learned after the release that Apple wanted it out of the way by the ADC so they could concentrate on the Intel move.
Going by Apple's normal beta time schedule, there isn't enough time for the proper number of beta's, and final master before the release, if it's for end of March. Apple would have to skip a beta at least. This is what MS is expected to do if they release before the beginning of 2010. That's never a good idea.
As Jobs said a year for 10.6's release, there is an expectation that it would slip from that. Apple must know the minimum time it will take for a quality release. Unexpected problems hold that up.
How often have these large projects come in under the deadline? Almost never.
So when I hear that the time has been pushed up, I wonder why. Could this have all been so much easier than they expected?
Or are they, like MS with Vista, dropping features from the release in order to get it here before Windows 7?
Will ZFS be there or not? That's a big question. Sun just added a couple of good new features to the release in Solaris. Would that require that Apple drop it from consideration in 10.6? If so, would that speed up the release?
This is all a headache to think about.
As long as Apple isn't rushing this, it will be fine. The only way we'll know is by how stable it is when it does get released, and which features that we do know about, that made it in.
If Apple could release Mac OS X 10.6 Snow Leopard at the same time as the new quad-core iMacs in late January, mid-February 2009, the new iMacs would sell like hot cakes!
There's an expectation that when a ship date is announced for major software projects, that date will slip. It did for all of Apple's OS releases, except for 10.4, and we learned after the release that Apple wanted it out of the way by the ADC so they could concentrate on the Intel move.
Going by Apple's normal beta time schedule, there isn't enough time for the proper number of beta's, and final master before the release, if it's for end of March. Apple would have to skip a beta at least. This is what MS is expected to do if they release before the beginning of 2010. That's never a good idea.
As Jobs said a year for 10.6's release, there is an expectation that it would slip from that. Apple must know the minimum time it will take for a quality release. Unexpected problems hold that up.
How often have these large projects come in under the deadline? Almost never.
So when I hear that the time has been pushed up, I wonder why. Could this have all been so much easier than they expected?
Or are they, like MS with Vista, dropping features from the release in order to get it here before Windows 7?
Will ZFS be there or not? That's a big question. Sun just added a couple of good new features to the release in Solaris. Would that require that Apple drop it from consideration in 10.6? If so, would that speed up the release?
This is all a headache to think about.
As long as Apple isn't rushing this, it will be fine. The only way we'll know is by how stable it is when it does get released, and which features that we do know about, that made it in.
To be honest, it could have been a typo in the presentation. Q2 is what people were expecting. Nonetheless, I think people should be more open to an early release. It's not as if the people working on it are being given a public deadline and scared that it's not done yet. The release date is not official, right now it's probably a rough target.
I also thought it could be a typo, but it appears on two slides (#5 and #6) and it also says 14+ months delta, and 14 months is January.
I also thought it could be a typo, but it appears on two slides (#5 and #6) and it also says 14+ months delta, and 14 months is January.
and the + in 14+ means it could be January, or February, or June for that matter
I just hope they don't rush it, a buggy SL would be a BAD move, besides, if they do release fast, a lot of people will be suspicious and just wait it out anyways so what's the difference?
I like how people always say noooo, develop OS X slower. How do people know how many issues need fixed and optimized and how long they will take? The presentation shows that the Leopard -> Snow Leopard period is 14+ months and some other major releases took this long. 6 releases in 8 years = 16 month average cycle.
That average is dragged down by the shorter intervals of the earlier releases. There weren't as many ambitious features, and the OS was a lot smaller and simpler then too. As the OS got bigger and more complex, the interval grew. The more complex features generally can't just be dropped in. Stuff like Grand Central and OpenCL sound like one of Apple's most ambitious features to date. Even if it draws from existing knowledge, refactoring them to be their own set of system services and APIs sounds like too much to expect in such a short period of time.
Vista's problems were completely unrelated, any comparison there is tenuous at best.
Quote:
Which systems did Apple put out before they were ready? MobileMe perhaps but they know what they are doing most of the time. The simple solution is for Apple to release it when they think it's ready and consumers who know little to nothing of Apple's internal developments who arbitrarily think otherwise can simply wait before using it.
I don't think Leopard was ready. iPhone 2.0 doesn't have Apple's typical level of stability either.
Comments
OH OH OH! I just thought of something cool.
Finder is getting rewritten in Cocoa, right? And most other Apple apps are already in Cocoa now, right?
What if....Apple enabled Cocoa to use OpenCL behind the scenes!!!!!!! You write in Cocoa and the Cocoa components are tapping OpenCL making everything insanely...SNAPPIER!
Seriously though, do you know what kind of performance increase that would make for all Cocoa apps???? Windows 7 would have nothing on that!
The point of 10.6 is to replace almost all userspace apps with Cocoa only versions, be Multicore enabled/leveraged and Quartz Extreme on using OpenCL from the low level to the userspace by making the demands on OpenGL offloaded per core whether it's a core on the GPU or a CPU core.
OH OH OH! I just thought of something cool.
Finder is getting rewritten in Cocoa, right? And most other Apple apps are already in Cocoa now, right?
What if....Apple enabled Cocoa to use OpenCL behind the scenes!!!!!!! You write in Cocoa and the Cocoa components are tapping OpenCL making everything insanely...SNAPPIER!
Seriously though, do you know what kind of performance increase that would make for all Cocoa apps???? Windows 7 would have nothing on that!
This is a definite for Grand Cetral. Which will make attempts at abstracting and automatically handling multiple cores using existing Cocoa programming structures and objects. Which will make all SL Cocoa Apps run faster and more efficient for free! I believe it will be the same case for OpenCL but will require an extension to the API to designate certain tasks as potential processor intensive calculation tasks that are eligible to be offloaded to the GPU if necessary.
Great news! Go, Apple, go!
You state that like a fact. At this point, I think we're all just speculating.
SL can be whatever Apple wants it to be.
/Daniel
This is what Jobs said. So far, the beta's aren't showing anything major, though no doubt announced features haven't been added yet.
From the description of all the major re-writes to the OS they are doing, they would be very foolish to add too many new features.
I like how people always say noooo, develop OS X slower. How do people know how many issues need fixed and optimized and how long they will take? The presentation shows that the Leopard -> Snow Leopard period is 14+ months and some other major releases took this long. 6 releases in 8 years = 16 month average cycle.
Look at Microsoft's efforts. Taking longer doesn't guarantee a better product. Longer times mean that people lose focus on the task.
The focus with Snow Leopard is optimization, possibly removing PPC stuff and compiling for 64-bit. Just because some stuff sounds complicated doesn't mean it takes longer to implement - it's not as if they start working on it as soon as the last version is out the door. The GPU stuff has been going for a while in various incarnations. The multi-threading stuff will take R&D from Xgrid computing and previous multi-threading techniques.
A March release is still 4 months away too. Beta in January, a few betas in Feb, Final version in March.
To be honest, it could have been a typo in the presentation. Q2 is what people were expecting. Nonetheless, I think people should be more open to an early release. It's not as if the people working on it are being given a public deadline and scared that it's not done yet. The release date is not official, right now it's probably a rough target.
Which systems did Apple put out before they were ready? MobileMe perhaps but they know what they are doing most of the time. The simple solution is for Apple to release it when they think it's ready and consumers who know little to nothing of Apple's internal developments who arbitrarily think otherwise can simply wait before using it.
There's an expectation that when a ship date is announced for major software projects, that date will slip. It did for all of Apple's OS releases, except for 10.4, and we learned after the release that Apple wanted it out of the way by the ADC so they could concentrate on the Intel move.
Going by Apple's normal beta time schedule, there isn't enough time for the proper number of beta's, and final master before the release, if it's for end of March. Apple would have to skip a beta at least. This is what MS is expected to do if they release before the beginning of 2010. That's never a good idea.
As Jobs said a year for 10.6's release, there is an expectation that it would slip from that. Apple must know the minimum time it will take for a quality release. Unexpected problems hold that up.
How often have these large projects come in under the deadline? Almost never.
So when I hear that the time has been pushed up, I wonder why. Could this have all been so much easier than they expected?
Or are they, like MS with Vista, dropping features from the release in order to get it here before Windows 7?
Will ZFS be there or not? That's a big question. Sun just added a couple of good new features to the release in Solaris. Would that require that Apple drop it from consideration in 10.6? If so, would that speed up the release?
This is all a headache to think about.
As long as Apple isn't rushing this, it will be fine. The only way we'll know is by how stable it is when it does get released, and which features that we do know about, that made it in.
Beta releases to Premier Developers are always dozens of releases behind internal builds.
That doesn't matter. Those internal builds, they are not really betas, aren't of a quality to release to developers for trial.
It's the announced betas, the one that Apple puts out to the developers, that matter.
Those are the ones that are monitored to see how far Apple has come.
If Apple could release Mac OS X 10.6 Snow Leopard at the same time as the new quad-core iMacs in late January, mid-February 2009, the new iMacs would sell like hot cakes!
Great news! Go, Apple, go!
Yeah, great!
Then userland will become crashland.
with snow leopard coming out and the chance of hardware upgrades, should I wait to buy a mac book pro sometime before summer 2009?
Here's the end of an article about Nehalem (i7), the newest chips to be arriving.
Concentrate on what he says about Lynnfield.
Then you can decide.
http://www.anandtech.com/cpuchipsets...spx?i=3461&p=3
There's an expectation that when a ship date is announced for major software projects, that date will slip. It did for all of Apple's OS releases, except for 10.4, and we learned after the release that Apple wanted it out of the way by the ADC so they could concentrate on the Intel move.
Going by Apple's normal beta time schedule, there isn't enough time for the proper number of beta's, and final master before the release, if it's for end of March. Apple would have to skip a beta at least. This is what MS is expected to do if they release before the beginning of 2010. That's never a good idea.
As Jobs said a year for 10.6's release, there is an expectation that it would slip from that. Apple must know the minimum time it will take for a quality release. Unexpected problems hold that up.
How often have these large projects come in under the deadline? Almost never.
So when I hear that the time has been pushed up, I wonder why. Could this have all been so much easier than they expected?
Or are they, like MS with Vista, dropping features from the release in order to get it here before Windows 7?
Will ZFS be there or not? That's a big question. Sun just added a couple of good new features to the release in Solaris. Would that require that Apple drop it from consideration in 10.6? If so, would that speed up the release?
This is all a headache to think about.
As long as Apple isn't rushing this, it will be fine. The only way we'll know is by how stable it is when it does get released, and which features that we do know about, that made it in.
Very well said.
...
To be honest, it could have been a typo in the presentation. Q2 is what people were expecting. Nonetheless, I think people should be more open to an early release. It's not as if the people working on it are being given a public deadline and scared that it's not done yet. The release date is not official, right now it's probably a rough target.
I also thought it could be a typo, but it appears on two slides (#5 and #6) and it also says 14+ months delta, and 14 months is January.
I also thought it could be a typo, but it appears on two slides (#5 and #6) and it also says 14+ months delta, and 14 months is January.
and the + in 14+ means it could be January, or February, or June for that matter
I just hope they don't rush it, a buggy SL would be a BAD move, besides, if they do release fast, a lot of people will be suspicious and just wait it out anyways so what's the difference?
I like how people always say noooo, develop OS X slower. How do people know how many issues need fixed and optimized and how long they will take? The presentation shows that the Leopard -> Snow Leopard period is 14+ months and some other major releases took this long. 6 releases in 8 years = 16 month average cycle.
That average is dragged down by the shorter intervals of the earlier releases. There weren't as many ambitious features, and the OS was a lot smaller and simpler then too. As the OS got bigger and more complex, the interval grew. The more complex features generally can't just be dropped in. Stuff like Grand Central and OpenCL sound like one of Apple's most ambitious features to date. Even if it draws from existing knowledge, refactoring them to be their own set of system services and APIs sounds like too much to expect in such a short period of time.
Vista's problems were completely unrelated, any comparison there is tenuous at best.
Which systems did Apple put out before they were ready? MobileMe perhaps but they know what they are doing most of the time. The simple solution is for Apple to release it when they think it's ready and consumers who know little to nothing of Apple's internal developments who arbitrarily think otherwise can simply wait before using it.
I don't think Leopard was ready. iPhone 2.0 doesn't have Apple's typical level of stability either.