Mac OS X 10.6 Called 'Snow Leopard', All Cocoa?

[unwinded]

I don’t see why anybody would be upset with this.

Honestly, think about it how lean and stable it could be. No more fat binaries compiled for PPC 32-bit, G5 64-bit, x86, and x64. Everything could be one happy x64 binary. They could make the OS better utilize multiple cores since all Core 2 Duos have at least two cores too. No more worrying about legacy hardware either. It would be the best Mac OS ever.

Leopard would be supported and maintained for some time anyways the same way that Tiger is. Tiger isn’t useless either, my wife still runs it just fine on her first-generation Macbook with a non-x64 Core Duo processor. She isn’t going to care when a new OS comes out that won’t even run on her computer. In my opinion, two years is the useful lifespan for ANY computer regardless of how powerful it was when you bought it. Apple probably even sold more non-x64 Intel Macs than G5s anyways. During the G5 era is when consumers shifted to buying more laptops than desktops, one of the main reasons they switched to Intel in the first place (since there was no way G5 was going in a laptop). G4 probably outsold G5, and I’ve seen Leopard on a high-end G4 Powerbook. Not entirely comfortable, even if it is only a little over two years old.

 

[BenRoethig]

Sure, they're a hardware company first, a software company second. No one's debating that.

The sooner they learn that the second lead directly to the first, they better off they are.

 

[Kar98]

The sooner they learn that the second lead directly to the first, they better off they are.

Says the guy who's got $16 billion socked away.
Or...has he?

 

[BenRoethig]

I don’t see why anybody would be upset with this.

Honestly, think about it how lean and stable it could be. No more fat binaries compiled for PPC 32-bit, G5 64-bit, x86, and x64. Everything could be one happy x64 binary. They could make the OS better utilize multiple cores since all Core 2 Duos have at least two cores too. No more worrying about legacy hardware either. It would be the best Mac OS ever.

They will still have universal binaries for most Apps for a while to come. What this does effect is how fast snow leopard will run them.

 

[Manic Mouse]

I guess my point is that Apple are going to have to put a lot of work into making OSX run better.

The rumour is that 10.6 will have no new features, and there are two reasons for this. The first is as to not tick off all the PPC owners who the OS wont support and the second is because they'll be putting all their development R&D into improving the innards of the OS and not adding anything much in terms of what the user sees. I suspect this will mostly involve removing PPC and other legacy code to slim down the OS size (and we can all agree that Leopard is well on it's way to being bloated with it's 15Gb install size, even if languages and drivers take up a good portion of it) and it will also involve optimising the OS as much as possible for Intel processors with their specific quirks and components (such as SSE3 + 4). I'm betting that while the PPC to Intel switch was smooth enough, there is still a lot of work that could be done on the OS to clean it up in terms of Intel support. This was an OS that was primarily developed on PPC for the first 6 years, then ported to Intel don't forget. They may also be putting the hooks for multitouch support into the OS.

As far as I see it all of the above is irrelevant to PPC owners, as all the improvements are either only applicable to Intel machines or for future-proofing the OS. Not that I'm saying the OS couldn't be optimised for PPC machines, but it is probably already fairly optimised for them (having been developed on them from day one) and a lot of effort would be required to "pick those high apples" that, at the end of the day, probably isn't really worth it for Apple when they could spend that time optimising for Intel.

I'm no software engineer however, so it's merely conjecture as to how I see the situation. Snow Leopard will be a clean-up of the OS for current Intel machines and future computers, and would offer very little (over leopard) for PPC users were a PPC version developed. I just don't get why PPC owners would whine about not getting an update that would probably do nothing for them.

 

[jf8]

Totally disagree here. Keeping around legacy crap is what got Microsoft into trouble. They still have very bad legacy APIs laying around. Developers are lazy, I am a developer so I know how it is trust me.

Maintaing backwards compatibility by merely keeping old APIs around is not the major issue with Windows. Microsoft has continued to develop the same API while keeping backwards compatibility a primary goal. They've designed a new API, .NET, which solves many of the problems inherent to Win32. But .NET is not suitable for all kinds of applications and will not take over as the preferred way to develop most applications (like Cocoa has.) There are some elements of Windows's design that Microsoft has not improved - not all are inherently bad, but some things can simply be done better.

DOS was the foundation for Windows until NT came along...Win2000, XP are all derived from an all new code base. But MS kept the APIs supported from the first versions of Windows (Win16s) on the new codebase. Through virtualization, DOS apps also run, even though XP and Vista have no DOS code.

Backwards compatibility through virtualization hurts nothing. It doesn't impact stability one bit. Windows is a mess because of things it chose NOT to virtualize but run natively on the NT base, like older variants of the Win APIs. Those do nasty things.

Win16 applications are virtualized on Windows NT via ntvdm.exe. 64-bit versions of Windows dropped this capability.

The Classic environment was necessary because Classic applications ran on a significantly different (and inferior) environment. Classic Mac OS was a single-user cooperative multitasking system without protected memory.

Virtualization is not necessary to successfully have different versions of the API - Windows NT has always had support for multiple subsystems. There is a POSIX subsystem which provides a very Unix-like environment (available as an optional install).

 

[BenRoethig]

I guess my point is that Apple are going to have to put a lot of work into making OSX run better.

The rumour is that 10.6 will have no new features, and there are two reasons for this. The first is as to not tick off all the PPC owners who the OS wont support and the second is because they'll be putting all their development R&D into improving the innards of the OS and not adding anything much in terms of what the user sees. I suspect this will mostly involve removing PPC and other legacy code to slim down the OS size (and we can all agree that Leopard is well on it's way to being bloated with it's 15Gb install size, even if languages and drivers take up a good portion of it) and it will also involve optimising the OS as much as possible for Intel processors with their specific quirks and components (such as SSE3 + 4). I'm betting that while the PPC to Intel switch was smooth enough, there is still a lot of work that could be done on the OS to clean it up in terms of Intel support. This was an OS that was primarily developed on PPC for the first 6 years, then ported to Intel don't forget. They may also be putting the hooks for multitouch support into the OS.

As far as I see it all of the above is irrelevant to PPC owners, as all the improvements are either only for Intel machines or for future-proofing the OS. Not that I'm saying the OS couldn't be optimised for PPC machines, but it is probably already fairly optimised for them (having been developed on them from day one) and a lot of effort would probably be required to do this that, at the end of the day, probably isn't really worth it for Apple when they could spend that time optimising for Intel.

I'm no software engineer however, so it's merely conjecture as to how I see the situation. Snow Leopard will be a clean-up of the OS for current Intel machines and future computers, and would offer very little for PPC users were a PPC version developed.

Yep. All snow leopard is a highly optimized version of leopard for intel machine. With the exception of a hopefully vastly improved OpenGL implementation, there isn't much that would translate. What does could be incorporated into patches for regular leopard. The bloating of the OS by legacy PowerPC code and drivers is what they're trying to fix.

 

[milo]

And how would a release designed to improve Intel performance and clean up the code-base for those machines be of any use to PPC owners? If 10.6 isn't going to introduce any new features, you're not going to be missing out on anything (that would reasonably affect you).

That's assuming that it will only be optimization and bugfix, with no new features. That's still a rumor at this point.

And either way, my biggest concern is if we start seeing apps that are 10.6 only. As long as things still run on 10.5, it won't matter that much, but if apple stops supporting PPC, why should anyone else keep supporting it?

And your comparison is a strawman as current macs aren't like PPC, which is a dead architecture (at least as far as Macs go).

First, you need to look up what "strawman" means, because that has nothing to do with anything I have said. (for the record, I'm listing an extreme case in order to make a point)

As for the chips, you can define "dead" however you want. Why shouldn't apple optimize for the intel cpus that are in machines shipping now and dump earlier models? There are instruction sets in newer chips that aren't in previous generation macs? If they're going to be cleaning out legacy, why not go all the way and dump earlier intel as well?

 

[Manic Mouse]

That's assuming that it will only be optimization and bugfix, with no new features. That's still a rumor at this point.

And either way, my biggest concern is if we start seeing apps that are 10.6 only. As long as things still run on 10.5, it won't matter that much, but if apple stops supporting PPC, why should anyone else keep supporting it?

Well this thread is about those rumours, so it would make sense to talk about them. I've made it clear that I'm basing everything I say off those rumours, and if they're untrue everything I've said is obviously rubbish (or moreso than it already is ).

If it is simply an optimised version of Leopard then there wont be any API differences between 10.5 and 10.6 so why would applications be 10.6 only? Everything one OS could do the other would be equally (and identically) capable of doing.

First, you need to look up what "strawman" means, because that has nothing to do with anything I have said. (for the record, I'm listing an extreme case in order to make a point)

a "straw man argument" is to describe a position that superficially resembles an opponent's actual view but is easier to refute, then attribute that position to the opponent (for example, deliberately overstating the opponent's position)

IMO your statement is a strawman rather than hyperbole, as the situations are subtly (but importantly) different. Not that it really matters, semantics are silly.

As for the chips, you can define "dead" however you want. Why shouldn't apple optimize for the intel cpus that are in machines shipping now and dump earlier models? There are instruction sets in newer chips that aren't in previous generation macs? If they're going to be cleaning out legacy, why not go all the way and dump earlier intel as well?

Firstly because current Intel chips support the older SSE1, 2 and 3 instruction sets as well as the most recent SSE4. And secondly because each instruction set offers different optimisations, so all are useful and optimisations to any of their utilisation will improve future performance. There is no issue of "dumping" earlier instruction sets as far as I understand it as none can be considered as "legacy". Each new version adds on new instruction sets, but all the previous ones are still there. In other words SSE4 is all of SSE3 plus a few new tricks. SSE3 is SSE2 plus a few new tricks etc. SSE4 incorporates all previous SSEs.

This, of course, is entirely different than spending time developing for PPC.

 

[jf8]

Honestly, think about it how lean and stable it could be. No more fat binaries compiled for PPC 32-bit, G5 64-bit, x86, and x64. Everything could be one happy x64 binary. They could make the OS better utilize multiple cores since all Core 2 Duos have at least two cores too. No more worrying about legacy hardware either. It would be the best Mac OS ever.

There is often little to no advantage to using x86_64 binaries instead of i386 (or ppc64 instead of ppc). Only applications that significantly benefit from 64-bit need to use it at all. Leopard is 64-bit on Intel - it can run 64-bit applications and address >4GB of memory without problems, but most things are only compiled for x86 and ppc. On my Intel Core 2 Duo Leopard system with many applications installed, only Chess and Xcode are compiled for ppc64 or x86_64... and Xcode uses 32-bit mode by default.

And moving to all x64 binaries doesn't really help the operating system use multiple cores. Applications that are multithreaded will perform better when more cores are available, regardless of the instruction set they are compiled for.

 

[milo]

IMO your statement is a strawman rather than hyperbole, as the situations are subtly (but importantly) different.

A strawman argument is disputing statements that nobody has actually made. Creating an opposing argument that nobody has stated in order to have something easy to dispute. Where did I do that?

 

[Eric S.]

I don’t see why anybody would be upset with this.

Then you fail to see the viewpoint that a lot of us have.

Honestly, think about it how lean and stable it could be. No more fat binaries compiled for PPC 32-bit, G5 64-bit, x86, and x64. Everything could be one happy x64 binary. They could make the OS better utilize multiple cores since all Core 2 Duos have at least two cores too. No more worrying about legacy hardware either. It would be the best Mac OS ever.

I wonder how much time they spend "worrying" about it. This decision, if it comes, will be a marketing decision, not a technical one. The code to support both architectures is already there. It would take them more effort to remove it than to leave it in. I'm sure the code to optimize for multiple cores is also already there.

Leopard would be supported and maintained for some time anyways the same way that Tiger is.

Which no one is questioning.

Tiger isn’t useless either, my wife still runs it just fine on her first-generation Macbook with a non-x64 Core Duo processor. She isn’t going to care when a new OS comes out that won’t even run on her computer.

Good for her.

In my opinion, two years is the useful lifespan for ANY computer regardless of how powerful it was when you bought it.

This statement is just wrong. Two years?? I still use the Power Mac G4 that I bought eight years ago. With upgrades to the CPU and graphics card, adding USB 2.0 and a SATA drive (obviously this was back when Apple built expandable systems), this machine is still eminently usable. (My method was to buy a Mac every other generation. But after the G5 came the switch to Intel, and since I couldn't run Classic mode that left me out. So I have made the G4 last all these years.) And, for example, a dual-processor G5 desktop was no small expense a couple years ago. To say that that should just be abandoned, and for no technical reason, would be insensitive to loyal Mac customers.

Apple probably even sold more non-x64 Intel Macs than G5s anyways.

Possibly, but so what?

G4 probably outsold G5, and I’ve seen Leopard on a high-end G4 Powerbook. Not entirely comfortable, even if it is only a little over two years old.

A stock G4 laptop probably has a bit of a slow disk and somewhat low amount of memory for Leopard. Upgrade those and I expect it would perform just fine.

 

[dgdosen]

See, this is exactly what I'm talking about. This guy still loves his PPC machine, has zero reason to replace it. And he shouldn't have to repplace it just to run 10.6 Apple, us PPC users will upgrade eventually, just give us a little time. You need to support us for just a little while longer!

Just because they come out with a new version, doesn't mean they won't continue to support 10.5 on existing PPC hardware, for at least the time being.

It just might mean you can't upgrade to the latest and greatest. Big deal.

 

[Manic Mouse]

A strawman argument is disputing statements that nobody has actually made. Creating an opposing argument that nobody has stated in order to have something easy to dispute. Where did I do that?

You implied that my logic would be applicable to current Intel machines. Which is not the case, that is not my argument, and is thus a strawman. Is it not? (Like I said semantics are silly).

 

[SkyBell]

Leopard would be supported and maintained for some time anyways the same way that Tiger is. Tiger isn’t useless either, my wife still runs it just fine on her first-generation Macbook with a non-x64 Core Duo processor. She isn’t going to care when a new OS comes out that won’t even run on her computer. In my opinion, two years is the useful lifespan for ANY computer regardless of how powerful it was when you bought it. Apple probably even sold more non-x64 Intel Macs than G5s anyways. During the G5 era is when consumers shifted to buying more laptops than desktops, one of the main reasons they switched to Intel in the first place (since there was no way G5 was going in a laptop). G4 probably outsold G5, and I’ve seen Leopard on a high-end G4 Powerbook. Not entirely comfortable, even if it is only a little over two years old.

I understand many people don't care about running the latest OS. I'm one of them. But, I'm going to guess most do care about it, as evidenced by this whole thread. You're right, Tiger isn't useless, it's what I'm running. But, I did have Leopard on my three year old iBook G4 for a while. I missed the way Tiger looked, so I switched. Aside from the wireless problems, and the bit of lag that was caused by me having only 512 MB of RAM, Leopard ran fantastically. G5 is definitely still useable for nearly everything, G4's are still useable for most people. I run Adobe CS3 apps without problems.

PPC is still very usable, and Apple should understand this.

 

[Manic Mouse]

I understand many people don't care about running the latest OS. I'm one of them. But, I'm going to guess most do care about it, as evidenced by this whole thread. You're right, Tiger isn't useless, it's what I'm running. But, I did have Leopard on my three year old iBook G4 for a while. I missed the way Tiger looked, so I switched. Aside from the wireless problems, and the bit of lag that was caused by me having only 512 MB of RAM, Leopard ran fantastically. G5 is definitely still useable for nearly everything, G4's are still useable for most people. I run Adobe CS3 apps without problems.

PPC is still very usable, and Apple should understand this.

And, if Snow Leopard is just an Intel optimised and cleaned up Leopard, what will it offer PPC users? It seems fairly pointless to me...

 

[iMpathetic]

An extra word to say while their voices are dripping with newfound lust?

 

[kolax]

I understand many people don't care about running the latest OS. I'm one of them. But, I'm going to guess most do care about it, as evidenced by this whole thread. You're right, Tiger isn't useless, it's what I'm running. But, I did have Leopard on my three year old iBook G4 for a while. I missed the way Tiger looked, so I switched. Aside from the wireless problems, and the bit of lag that was caused by me having only 512 MB of RAM, Leopard ran fantastically. G5 is definitely still useable for nearly everything, G4's are still useable for most people. I run Adobe CS3 apps without problems.

PPC is still very usable, and Apple should understand this.

They don't have to upgrade to 10.6. For Apple to make advances, they need to focus on the future rather than supporting technology from the past.

When 10.6 is released, and if it doesn't support PPC, people with PPC Mac's will still be able to do exactly what they done before. 10.6. being released isn't going to cause all PPC Mac's to die.

The latest software should aim for what the current and future hardware is going to be.

 

[Riemann Zeta]

I have to believe that this rumour of Apple releasing many different strains of OS X that only work on certain machines is false...but every news source on the web has sure run with it. I hope it is not true, because if it is, it could the stupidest move they have ever made: forking the OS X codebase and trying to simultaneously develop a bunch of slightly different, incompatible variations.

Code forking is a real pain in the ass, a strain on developers and a complete marketing clusterf*ck. I think Microsoft has learned this the hard way with the dozens of different "flavors" of Windows XP and Vista. There is Windows Home Basic, Windows Home Advanced, Windows Business Premium, Windows Server Advanced Web Database, Windows Media TV Center Home Server, Windows for Tablets and Computers with Pens, Windows for Computers Smaller than Laptops That Maybe Have a Pen, Windows Ultimate Extreme Super-Terrific Happy Fun Edition...it's gotten just as bad as video card product lines, with things like the NVIDIA GeForce9 9990XTX Pro GTS-2X Extreme! I mean, after all these years of Windows [Adjective Orgy] Edition, they learned their lesson and are just going back to a damn number: Windows 7.

Having all those different product names and versions and sub-builds is just overcomplicating things. The whole point of Mac OS is that there is a single Mac OS. It runs on Macs. 'Nuff said.

 

[shaunymac]

I echo some of the PPC users out there, having two, that what is the purpose for apple to stop supporting our macs? Only to have a faster and more stable OS for us on Intel machines? Something about that doesnt quite sound right to me. All of these people that have PPC's and for those that still use them have made apple what they are today. I know that it is invetiable that some macs will not be able to keep up the minimum requirements required to upgrade to the the new and feature packed OS, but if it all possibly, I think it would be wise for apple to continue to support the PPC users. IE, I am able to run tiger on my blue and white g3 . For everyone complaining about all of the coding that PPC's require, which I know nothing about, why not put the "snow leopard" or whatever it is on separate discs. If this were to happen, everyone would benefit from the new OS. Just my 2 cents.

 

[SkyBell]

And, if Snow Leopard is just an Intel optimised and cleaned up Leopard, what will it offer PPC users? It seems fairly pointless to me...

Really, it's more about the principle then anything else. "Sorry, this new release has nothing new to offer so it wouldn't put more of a strain on your system in any way, but we've decided to not let you use it anyway."

They don't have to upgrade to 10.6. For Apple to make advances, they need to focus on the future rather than supporting technology from the past.

When 10.6 is released, and if it doesn't support PPC, people with PPC Mac's will still be able to do exactly what they done before. 10.6. being released isn't going to cause all PPC Mac's to die.

The latest software should aim for what the current and future hardware is going to be.

Yes, that is right. I'm not saying that PPC users will be able to do anything less. And yes, Apple does need to focus on the present and future to make advances. But at the same time, they need to focus on the people who bought before the Intel switch, because these people were loyal to Macs even before all of the improvements Intel brought. You can't forget about the people who helped make you a successful company before the "switchers" came along. Sure, a lot of people who were loyal to Apple before Intel came out, have already upgraded to Intel machines. But there's also a lot of switchers mixed in there. People who use PPC, around 90-95% of them were long time Apple users, the ones who were always loyal to the company.

I'm just saying, it would really be a disappointment to long time Mac users if Apple dropped PPC so soon.

 

[Manic Mouse]

Really, it's more about the principle then anything else. "Sorry, this new release has nothing new to offer so it wouldn't put more of a strain on your system in any way, but we've decided to not let you use it anyway."

But the PPC version of 10.6 (in this scenario) WOULD BE 10.5, which you already have available on PPC. You might as well get Apple to give you a blank DVD with 10.6 written on it by Steve Jobs with a crayon.

An OS release focused solely on under-the-hood Intel improvements would be literally pointless on a PPC machine.

Were Snow Leopard supposed to add big new features then I could understand why the PPC crowd are up in arms. But the rumours say it's nothing more than a cleaned up, Intel optimised Leopard.

 

[milo]

You implied that my logic would be applicable to current Intel machines. Which is not the case, that is not my argument, and is thus a strawman. Is it not? (Like I said semantics are silly).

Nope, it's not.

If I had said that YOU used that logic on intel machines (which you didn't), that would be a strawman. But I didn't say that. You're right, it is semantics...but you still shouldn't make accusations using terms you don't really understand.

And, if Snow Leopard is just an Intel optimised and cleaned up Leopard, what will it offer PPC users? It seems fairly pointless to me...

People are skeptical about this rumor - I think most of the concern comes from the notion that PPC support may be dropped, but there WILL be new features. Do you really think apple would do a major OS release with no new features?

Personally, I think this rumor is wrong (the bit about dropping carbon support has already been disputed), but I'm still concerned about PPC support in 10.6, especially if it ships so soon.

They don't have to upgrade to 10.6. For Apple to make advances, they need to focus on the future rather than supporting technology from the past.

If supporting hardware isn't a priority for apple any more, and the length of support is getting shorter, doesn't that make the purchase of ANY new machine riskier? How do we know that intel won't ship Core 3 in a year or two and apple won't dump support for current machines if it helps them optimize more and slim their codebase? After all, there has already been rumblings that 10.6 could dump support for 32 bit intel along with PPC.

 

[nuckinfutz]

Folks

The reality is PPC is a dead end and while that doesn't give Apple the right to strand PPC users I think it's a tough decsion to make and next week we'll see exactly what Apple's strategy is.

Yes the tools are there to compile for both but that doesn't obviate the need for tweaking, debugging and optimizing the Intel and PPC portions of your code as a developer.

I want Leopard to continue to be optimized and support PPC and Carbon fully. However I'd like 10.6 to break a bit with the legacy support and pull the platform over the hump and get it moving forward on an Intel/ARM platform. With PPC you'll never see more than a dual core processor. Nehalem next year will deliver up to 8-cores with 2 threads per core.

Intel will be delivering Larrabee their discrete GPU product which will offer a derivative x86 intruction set for shaders and offer GPGPU functionality.

That doesn't really matter if you have a Quad G5 and want good support so I agree...Apple has to hand this next transition well and keep both sides happy.

 

[milo]

But the PPC version of 10.6 (in this scenario) WOULD BE 10.5

That's what the rumor says, but I don't buy it. It really makes no sense if you think about it. Apple is really going to release a new OS, tell intel folks they need it (and should pay for it) but at the same time tell PPC folks that it's the exact same thing as 10.5 so it's unnecessary? That's too ridiculous to even imagine.

10.6 being intel only is only acceptable to a PPC user if there are no new features, or anything else to break compatibility with 10.5 apps. But there's no way I believe that will be the case.