Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Because all applications aren't going to be 64 bit for 20 years or more as there is no performance benefit in the vast majority of cases.

I mean Microsoft is still going to be selling a 32 bit OS which can't run ANY 64 bit programs at all for the next 5 years at least.

i doubt that. in the last 20 years we've gone from 16 bit to 64 bit architecture. what makes you think this trend will stop all of a sudden?

i'm not sure about there being no performance benefits in most cases, but if all the Mac's are currently running on 64 bit architecture, doesn't it make sense that Mac developers should start making 64 bit applications. not doing so is like buying a Ferrari and filling it with regular gas instead of premium.
 
File System

Everyone also seems to have forgotten about ZFS. Remember that Apple implemented a read only version of ZFS (smart and self-healing file system) into their development tools.

I'd be very surprised if we didn't see the next OS built around it. It's perfect for all data types and especially perfect for consumer tools like TimeMachine and pro tools like FCP.

Thoughts?
 
There is also a suggestion that Apple may move Mac OS X 10.6 to "Cocoa-only", but the full meaning of this remains vague

Apple is working very hard on clang in llvm. Maybe the 'Cocoa only' refers to the compiler will compile Cocoa only, and not Carbon apps. I don't know enough about compilers, or Carbon, to tell if they are implementing Carbon in clang. The source code is public.
clang supports compiling the C language currently which means it could only support Carbon applications (Carbon API is C based). Folks are working on adding Objective-C and C++ to clang but it will be a while before it is ready. ...so no Cocoa.
 
Unix = RISC
Linux = x86

OS X is UNIX 03 certified.
Linux is a Unix-like Operating System that is mostly UNIX 03 compliant.

The Linux Kernel is purely Monolithic.
The OS X Kernel is Micro-Monolithic.

RISC and x86 are Chip Architectures representing two different worlds of Assembly and system designs for basic CPU designs.

As has been pointed out, x86 today is a wrapped hybrid of RISC.
 
And with it they would kill all business support for Macs forever, and probably take the iPhone with it. No sensible business would run business critical software such as CS3 or Office on a depreciated (and therefore unsupported) framework, so they'd switch to Windows.

Baseless observation. iPhone is Cocoa Touch only. The Fortune 500 Industry have all signed on-board and are the reason WWDC is sold out.

Office will soon be faced with OpenOffice 3.0 Cocoa that has ODF 1.2 native, Office2007 OpenXML compliance and much more.

CS3 is a nitch market that has nothing to do with general business systems.

CS3 will continue to run in a Carbon environment but won't continue to get any additions to that environment.

CS3 will be mainly run on Tiger and Leopard systems.

This transition has been known to Adobe and Microsoft for some time.

Third Parties will be given time to transition or be stuck selling on Tiger/Leopard compliant systems only.
 
Getting rid of Carbon would break a hell of a lot of software including most of Apples software. I'm not sure how much I would trust that rumour even if it is only the UI stuff that is removed. Especially seeing as Core Foundation is technically a Carbon API but is part of the Cocoa frameworks as well.

Core Foundation is intentionally shared between Carbon and Cocoa. It forms the basis for both as part of Core Services.

Core Foundation, Core Image, Core Text, etc aren't going anywhere if Carbon dies. What you do get screwed on is that your GUI /must/ go Cocoa (unless some C-based GUI API replaces Carbon's APIs), and you may have to ditch some of the older Carbon APIs still used from the OS 9 -> X transition that were lower level.
 
i doubt that. in the last 20 years we've gone from 16 bit to 64 bit architecture. what makes you think this trend will stop all of a sudden?

i'm not sure about there being no performance benefits in most cases, but if all the Mac's are currently running on 64 bit architecture, doesn't it make sense that Mac developers should start making 64 bit applications. not doing so is like buying a Ferrari and filling it with regular gas instead of premium.

The answer to your question is simple. A 16-bit arch allows for a 64KB flat address space. This is far, far less than most programs today need... less than programs needed even back in 1985. There are ways to address more than that (evil "segments") but it still limits the kinds of problems you can easily solve.

32-bit allows for up to a 4GB flat space (in theory, some of this is usually not available to the application). Most programs today do not work with this much information at once. A complex program like Safari may need a virtual space of a gig or so, but not much more than that.

When you double the size of your pointers to 64 bits, your overhead increases, the memory bandwidth you need increases, etc. This is a hit to performance. Now, if you are loading up a 100GB video, it will be much more efficient to work with it in 64 bits, which is why people are so interested in 64-bit nowadays, but nobody wants to use it for every application.
 
Wow, clam down, did you even read what it said?



It makes sense for it to be Intel only. I hope it will be, so that they can remove all of the PPC code. As for the 'no new features' thing, I don't mind, as long as Leopard has Resolution Independence by then.

It makes absolutely no sense to get rid of PPC. Put yourself in the shoes of people who still have PPC. People with G5's suddenly have a powerful computer that can't run the latest OS, simply because there's no code for their processor, even though the processor itself is fully capable of running it. PPC support will continue at least until 10.6, maybe 10.7 for the upper end G5's. I suspect that 10.7 will be the last to support PPC, and Apple will introduce OS XI, with no PPC support.
 


It makes absolutely no sense to get rid of PPC. Put yourself in the shoes of people who still have PPC. People with G5's suddenly have a powerful computer that can't run the latest OS, simply because there's no code for their processor, even though the processor itself is fully capable of running it. PPC support will continue at least until 10.6, maybe 10.7 for the upper end G5's. I suspect that 10.7 will be the last to support PPC, and Apple will introduce OS XI, with no PPC support.

How is a G5 running Leopard useless? You are out the park on this one. PPC support will be lucky to survive 10.6 and it's certainly not going to be in 10.7. Apple is not going to jump from 10.7 to 11.0 that's nonsensical.
 
Ok so the real question is HOW are they going to make it faster and why haven't they done so already?

It can be described as trying to pick fruit from a tree. Each fruit you pick is an optimization that provides speed. The low-hanging fruit is easy to reach, immediately feeds your hunger, and doesn't require a lot of effort. The stuff at the top is really hard to reach and takes a lot of effort to get at.

All speed optimizations are the same way, some are easy, and take little time to implement, while others take a long time. If I had to choose between 5 easy ones with 5% total improvement, and 1 hard one with 5% total improvement... I take the 5 easy ones every time, because I will still probably have time left over after the 5.
 
Ever heard of OS 9.5? Or .3 and up, for that matter?

Nope. But it borders on insanity to think that Apple would move through 8 successive 10.x upgrades and then inexplicably jump to 11.x.

Clearly for branding reasons Apple loves the Roman Numeral "X" "XI" isn't bad but isn't quite as clean IMO.

The stuck mental process for some here is the old world thinking that the bigger the advance in number the more "special" the application is.
 
It looks like Apple wants to match Windows 7 with their own version 7 of OS X i.e., 10.7, from a marketing perspective. So they will get out 10.6 fast to have enough gap in between and time it to Win7 release. So putting 10.6 in Jan 2009 kind of make sense. They can increase speed and reliability with few perks like ZFS, etc. May be they will surprise and make headlines by giving it a free update, or charge few bucks. Finally, I think instead of 10.6 being intel only, it may rather be last OS X version to be PPC also. Next versions will be PPC-free. That will give time to PPC folks to make a switch in a year and a half.

I should stop dreaming now. :D
 
Finally

I personally can't wait for Snow Leopard.

I'm ready for them to optimize the *fool* out of it.

And I'm waiting for ZFS support.

I want a truly modern UNIX operating system with the easiest to use interface, and I want it to be rock solid and super fast.
 
But does it make sense to do a speed and stability upgrade only? That sounds like a 10.5.X, not a 10.6! Has a new version of Mac OS (or Windows or a Linux distribution for that matter) ever been released with nothing new to offer? I think both the strange two words cat name and the lack of new features are both very unlikely...


youve never used Windows ME.

god that was a joke.
 
OS X is UNIX 03 certified.
Linux is a Unix-like Operating System that is mostly UNIX 03 compliant.

The Linux Kernel is purely Monolithic.
The OS X Kernel is Micro-Monolithic.

RISC and x86 are Chip Architectures representing two different worlds of Assembly and system designs for basic CPU designs.

As has been pointed out, x86 today is a wrapped hybrid of RISC.


I know. I just googled it also.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.