Daniel Eran Dilger
Random header image... Refresh for more!

WWDC 2008: New in Mac OS X Snow Leopard

OS X Snow Leopard
Daniel Eran Dilger
Apple’s public introduction of Snow Leopard, the next version of Mac OS X, was decidedly brief at WWDC, with only passing public mention of its new feature set. That’s in part because the company is delivering something nearly unheard of in the consumer software industry: Apple is advancing a new software product that improves upon its fundamentals rather than advancing a lot of marketing features.

Nothing New, Everything Newer.
As anticipated by multiple reports, Snow Leopard will not advance any major new consumer-facing marketing features apart from new support for Microsoft’s Exchange Server 2007 push messaging in its new versions of Mail, Address Book, and iCal. That move will match the functionality outlined for the iPhone 2.0 software update on the Mac desktop.

The fact that Apple is building its own Exchange Server support in both Mac OS X and the iPhone indicates that the company will no longer be waiting around reliant on Microsoft to make Macs fit into corporate environments with the dreadful embarrassment that is Entourage, just as I surmised earlier. Apple is taking matters into its own hands.

Additionally, by delivering Exchange support itself Apple will be also be able to offer its own alternatives to Exchange in parallel, much the same as how Microsoft leveraged its Mac Office software to develop Windows as an alternative to the Mac OS. This time around however, its Microsoft’s technology that will be routed around and replaced.

Apple’s first Exchange alternative is Mobile Me, a subscription based web service that serves as the next generation of its existing .Mac service. Beyond Mobile Me, Apple will also be offering an enhanced version of its own server product for businesses interested in hosting their own Exchange alternative.

Apple’s Mobile Me Takes On Exchange, Mobile Mesh
Snow Leopard Server Takes on Exchange, SharePoint

Trickle Down Tech: iPhone Software and the Mac Desktop.
In addition to inheriting the iPhone’s support for Exchange, the Mac OS X desktop is benefiting from investments made for iPhone in other respects as well. Last year, Leopard’s Core Animation sprang from work developed for the iPhone. Originally called the Layer Kit, Core Animation was conceived in order to give the iPhone a slick UI. Layer Kit was essentially a scaled down, mobile version of the desktop window server with new support for a highly animated new user interface explicitly designed to be easy for developers to fully exploit. Apple subsequently reused the technology in Leopard to make it just as easy to add rich interface polish to desktop apps as well.

In Snow Leopard, there will be another example of trickle down tech from the iPhone: QuickTime X (as in ten, not ex). Apple already performed a major overhaul of its flagship media architecture in QuickTime 7, adding support for modern audio and video codecs and ridding it of a lot of old legacy from the media architecture’s early 90s origins. However, QuickTime is still a complex assortment of software components, designed to work with media in nearly any codec imaginable and support everything from simple playback to complex media authoring. It is essentially an operating system for media.

When Apple developed media services for the iPhone, it started fresh with a pared down set of objectives. The iPhone is designed only to play music and video; it doesn’t need to edit or author it. Further, the iPhone uses specialized hardware that allows it to efficiently decompress H.264 video and AAC or MP3 audio; it doesn’t need to play any random files users can scrounge up on the web in archaic or non-standard formats. By limiting the iPhone to playback of modern codecs, Apple could create a really tight, highly efficient subset of QuickTime that performed well on mobile hardware while being conservative with its battery use.

Snow Leopard will make use of that same mobile-optimized playback software when playing any media that uses modern codecs. That gives desktop Macs the same highly efficient playback performance while still allowing them to fall back to the standard QuickTime routines when playing older codecs or anything requiring proprietary plugins. The marketing name for this iPhone-derived boost is QuickTime X.

The Pristine Technology Stack.
Adding marketing features has complicated the performance of Mac OS X. Even though each new reference release has typically boosted the overall speed of existing Macs, the potential for things ever getting too fast has been tempered by demands the new features in each release have required.

Tiger introduced Spotlight searching which while relatively efficient for desktop search has added a significant amount of overhead. Tiger’s Dashboard similarly delivered a far less demanding architecture for running applets than Konfabulator, but it still adds some extra work for the operating system to manage that users see as a hit against overall performance.

Leopard’s Time Machine is similarly designed to back up files at reasonable intervals and avoid wasting too many cycles on iterating through files manually to see what needs to be backed up, but it still has the potential to hammer performance while doing its thing.

By spending a full release cycle on tuning up existing services rather than tacking on more features, Apple will provide users with far more value in the long run, and be well positioned to eviscerate Windows 7 when it arrives in the ballpark of 2010, likely around the same time Apple will be showing off the release of Mac OS X 10.7.

Handling Processes like Network Packets: Grand Central.
Of course, part of the reason Mac OS X gets faster in each release while tacking on those “200 to 300 new features” (including the more ambitious undertakings such as Spotlight and Dashboard) is that Apple has also been working on its fundamentals all this time, too.

Throughout the development of Mac OS X, Apple has reexamined the old ways of doing things in UNIX and proposed new architectures. One example is launchd, the process that manages the launching, termination, and supervision of other processes in the system. It replaces a variety of existing process managers including init, rc, inetd, xinetd, atd, crond and watchdogd. Few UNIX vendors would bother to engineer an entirely new way to do things, and if undertaken in the FOSS world, such an innovation would rarely be adopted by enough of the Linux community to ever matter.

Apple now has the resources to create its own weather, so it can fix outstanding problems and then reap the benefits of that investment, frequently across multiple product offerings. The latest example of this is Grand Central, a new thread management architecture that greatly simplifies developers’ ability to take advantage of the multiple cores now being used in modern CPUs, as well as the raw processing power available in GPUs (graphical processor units) on the system’s video card.

Rather than expecting each developer to become an expert in the black art of multithreading, Apple has built sophisticated process management into the kernel where it belongs and added language conventions that enable mere mortals to take advantage of a wide variety of different hardware that users might have at their disposal.

Grand Central Dispatch manages processes in a manner analogous to modern networking. Old telephone equipment used to use circuit switching to transmit information over networks; a dedicated circuit path is easy to set up but it is also expensive and potentially fragile. Modern networking uses packet switching, which breaks up data, phone conversations, or video streams into packets and routes each of them independently in a far more efficient way that is also resilient to network outages. Packets get routed around the problems.

Snow Leopard’s Grand Central Dispatch does the same thing for processes, packetizing tasks into Blocks and routing them to available processing cores as efficiently as possible. It can also manage the big picture for the whole system, adjusting how it balances its tasks as the performance load increases. This would be close to impossible for Individual developers to do themselves.

More Bits of Computing: 64 And OpenCL.
It shouldn’t be surprising that Snow Leopard extends support for 64-bit processing, which became possible in Tiger and went mainstream in Leopard. Apple will be making the kernel and nearly all of its user apps 64-bit savvy, and is making it easy for developers to do the same for their own software.

The potential downside to 64-bit computing is that many processes will consume more memory to do the same thing. The upside is a significant, across the board performance improvement. For certain processor intensive apps, moving to 64-bits makes a massive improvement. These both offset the increase in memory needed, just as computers begin running into the 4 GB limit of 32-bit architectures. The move to 64-bit support system wide will allow Macs to eventually load up on an insane 16 terabytes of RAM, but in the shorter term will make better use of the 4 GB in the latest Mac Book Pros and the 32 GB or more available in Mac Pros and Xserves.

Apple is also advancing a new technology called OpenCL (for “open computing language”), which allows developers to spin off computationally heavy tasks into jobs that can be run on the fastest processors available to the system. The name is not accidentally related to OpenGL, which does a similar task for graphics operations to fully leverage the GPU in an abstracted way that works on any graphics card.

OpenCL unlocks the GPUs and the multiple cores in modern CPUs for developers who are writing code that deals with graphics and media processing or just heavy math and physics calculations (such as games). Using just slightly specialized code, OpenCL’s just in time compiler prepares developers’ tasks to run on the most appropriate processing engine available on the system, which is increasingly going to be the graphics card.

OpenCL also integrates with Grand Central to manage all those packetized tasks across multiple GPUs and the multiple cores of multiple CPUs in today’s systems. This will also open the door for Apple to include its own acceleration hardware technology in a way that will be easy if not automatic for developers to use.

Apple – Mac OS X Leopard – Snow Leopard

Why Did Apple Buy PA Semi?
How Apple’s PA Semi Acquisition Fits Into Its Chip History

More New Than the Who’s-Who Knew, Too.
Of course, there’s a lot more that all this going on in Snow Leopard. In fact, there’s so much that’s new that Apple’s line about “no new features” is a bit misleading. Snow Leopard pushes things ahead in a way that will confuse and befuddle tech pundits used to arranging Apple’s marketing names like refrigerator magnets. The next article will attempt to clear things up with a look at Myths of Snow Leopard.

I really like to hear from readers. Comment in the
Forum or email me with your ideas.

Like reading RoughlyDrafted? Share articles with your friends, link from your blog, and subscribe to my podcast! Submit to Reddit or Slashdot, or consider making a small donation supporting this site. Thanks!

Technorati Tags: , , , ,

  • Pingback: Apple’s Mobile Me Takes On Exchange, Mobile Mesh — RoughlyDrafted Magazine()

  • Pingback: WWDC 2008: Is Mac OS X 10.6 the Death of PowerPC? — RoughlyDrafted Magazine()

  • Pingback: Snow Leopard Server Takes on Exchange, SharePoint — RoughlyDrafted Magazine()

  • http://johnsessays.blogspot.com John Muir

    Insight into this Snow job of a Leopard at last! Thanks.

    The iPhone 3G explains why 10.6 was swept aside to a private meeting at WWDC instead of receiving the customary keynote preview. Though it does leave those of us outside the sessions at a curious loss. The idea of OS X reworked from the inside out is naturally intriguing. I’m looking forward to how you cover it just as much as the inevitable confusion among the pundit class!

  • elppa

    Microsoft have the Singularity research project which aims to help deal with the problem of parallelism. When this will turn into technology they can put back into Windows® remains to be seen (if at all).

    Interesting to see that Apple looking like having a solution ready to market far earlier. Parallelism one of the biggest problem facing computer science today — and I don’t think I am exaggerating in saying that.

    What I find most interesting of all is that even though Apple uses generic intel chips and generic PC hardware, OS X boxes could in the future perform significantly better than boxes running Windows or UNIX like systems (Ubuntu Linux, FreeBSD), thanks to the advances outlined above.

    I also wouldn’t underestimate the importance of Sun Microsystems D-TRACE technology, which was introduced in Leopard. It can’t all have been for the benefit of 3rd party developers, no doubt Apple will be using it to squeeze out extra performance from their own Applications.

    Finally (I appreciate this was a bit long) thank you to Dan for providing me with the clearest, most informative explanation of the these new technologies (Grand Central and OpenCL) along with their implications for the Mac platform that I have read anywhere.

    Many Mac news sources seem to have glossed over the issue or steered clear of a through explanation. I tried to pose this question on MacRumors (under the title “Connecting the dots: An explanation on recent OS X technologies?”) and got a single response, which told me what I already knew.

  • jdoc

    Pardon my ignorance, but I’ll assume that Apple is licensing MS Exchange Server for it’s iPhone and MobileMe services. What, then, is an “Exchange alternative”?

  • OlsonBW

    “Pardon my ignorance, but I’ll assume that Apple is licensing MS Exchange Server for it’s iPhone and MobileMe services. What, then, is an “Exchange alternative”?”

    It would be using Apple programs of Mail, iCal, etc., instead of Entourage.

  • OlsonBW

    Apple is making programming exciting again.

    I used to program in the early 80s until I was told to learn about PC networks, set them up, and keep all the users working.

    That ended my programming career as I didn’t have time for both.

    I’ve had chances since and did some programming but it was more of consulting to the programming department where I had expertise in systems that they didn’t and I could explain how things worked. But I didn’t do any serious programming.

    Now Apple has the iPhone and I’ve got an idea that I hope someone will pick up and do before I can learn XCode and get it done.

    It’s a program that you use when you are actually golfing. AFTER you take your shot you would tell the program on the iPhone which club you used. It would keep track of that and the GPS coordinates of where you hit the ball at. It would already know that the first shot of the round means you are teeing off. It would also check and store the weather at each hole in case the weather changed during the round (weather.com?).

    If the last club you used was a putter and then an iron or wood it would realize that you holed that last hit or you could manually touch the screen to say you holed it.

    If you are waiting for someone else you could do over head views of what the last hole looked like and draw on the screen a draft of what that hole looked like and see how your shots went.

    Or if you preloaded that course by scanning it in from the score card you could see the hole and your shots with distances and clubs showing the round you are playing or previous rounds or stats as to how you’ve done before and how you are doing today.

    When you get done and go home and sync your phone (or better yes wirelessly) you could see your rounds and all your stats on your iPhone or your Mac (maybe a crippled version for Windows … maybe).

  • MikieV


    “What I find most interesting of all is that even though Apple uses generic intel chips and generic PC hardware, OS X boxes could in the future perform significantly better than boxes running Windows or UNIX like systems (Ubuntu Linux, FreeBSD),…”


    Here I had been thinking that Apple would try to find some way to keep OSX from running on genetic hardware – to protect their hardware sales.

    Funny to think they may be able to sell an OS which would run significantly faster – and possibly more stable – on the same hardware as other OSes.

    How much markwet share could they gain from “switchers” who would gain a faster computer – and still be able to interact with the data/services running on MS Exchange servers?

    Would Apple be willing to risk the loss of hardware sales to get a siginificant increase in market-share?

  • roz

    If Apple can deliver 10.6 and have great Exchange support built-in it would mean to me that they are getting serious about delivering a competing enterprise client. That is very exciting.

  • roz

    @MikieV: I think that elppa meant that the guts of a mac are now very similar to any generic Intel based machine in terms of cpu and logic board.

    The question you raise though, at least to me is a valid one even though its been hashed and rehashed on forums like this ad nauseam.

    Clearly today Apple makes almost all its rev from hardware sales and very little from OS, software and service sales. Many think that its unlikely that Apple would want to put that business in jeopardy and they point to the fact that when Jobs arrived on the scene his first act was to kill the clone business.

    I would say that while no one really knows what they are up to, the company is now in a significantly different position. They have an apparent technical lead in OS development and a better platform for software development of all kinds. They also have built a much more rich ecosystem for their hardware sales, particularly on the retail side and stand-out hardware designs that differential their devices for the consumer market.

    I also think that regardless of what anyone says the Windows world is 20x larger than the mac install-base. Whenever Apple has torn down walls and opened their technology to the wider Windows-based tech market their business has exploded: iTunes, iPod, iPhone, Bootcamp. I have got to think they watch these dynamics very closely and work to find ways to talk to bridge their business into the larger world as they clearly have in MobileMe and seem to be intent to do with 10.6′ focus on Exchange support. It will be really exciting to see what they come up with…

  • http://www.roughlydrafted.com danieleran


    “Pardon my ignorance, but I’ll assume that Apple is licensing MS Exchange Server for it’s iPhone and MobileMe services. What, then, is an “Exchange alternative”?”

    Apple licensed ActiveSync, which is Microsoft’s proprietary push protocol, in order to support Exchange connectivity on the iPhone. Apple doesn’t license Exchange Server; your company has to be running that itself.

    Apple doesn’t need to use the ActiveSync protocol to send messages between Mobile Me and the iPhones or Macs. It certainly isn’t using Exchange Server to implement Mobile Me, which is a series of WebObjects applications serving up web pages and web services (push email to the iPhone, for example).

    Mac OS X Server also won’t be running Exchange Server to provide push email/calendar/contacts, and Apple does not need to implement ActiveSync in order to provide push email from its own server to its own clients.

    So Exchange alternative #1 is Mobile Me, which offers 80% of what a hosted Exchange Server account would, but adds other features of more interest to consumers and bundles it all up in a nicer package at a better price with more storage.

    Exchange alternative #2 is Apple’s own server, which similarly offers perhaps 80% of what Exchange Server does, but does it in a standards-based, interoperable, better scaling fashion at a tiny fraction of the cost of setting up an Exchange infrastructure.

    Apple’s efforts to support Exchange have greatly helped its ability to offer both alternative, as Exchange support will sell iPhones and Macs right now, funding the ongoing development of Apple’s two Exchange alternatives. Had Apple only offered its own Exchange alternatives, it would be harder to get Exchange Server users to switch and would be far harder to sell its products to shops already invested in Exchange.

    There is another related technology that I’ll describe in a future article that ties into the same efforts. Essentially, Exchange support will help fund the development of Apple’s assault on Exchange. But Apple would never say that.

  • dicklacara


    I know some guys that just might be able to write that program– they are iPhone developers about to release with an exciting app. Email me at:


    and i’ll connect you to them!

  • dicklacara


    They are at WWDC until Sat AM.

  • labrats5

    “Would Apple be willing to risk the loss of hardware sales to get a siginificant increase in market-share?”

    Not yet. The history of computers has taught us that even if you technology is vastly superior to that of your competitors, it won’t always mean you’ll win. Apple’s largest barrier as a platform is enterprise, and there biggest barrier into enterprise is mindshare. People simply don’t even consider Apple as an option, so having better technology won’t help them.

  • gus2000

    Dan has once again explained a new concept so that even I could understand it. Kudos.

    Back in the early 90’s, PCs still had “IBM XT Compatibility Mode” where you could force the motherboard back to 4.77MHz for legacy software or ISA cards. There was frequently a dedicated front-panel button to turn down the speed that we jokingly referred to as the “Macintosh Compatibility Button”.

    Fifteen years later, the Mac is already the best hardware for running Vista, and OS X is poised to overtake Windows entirely in term of speed. This is probably because all the code from 15 years ago is still in some Vista DLL somewhere.

    Maybe Macs should ship with an underclocking button for “Vista Compatibility”.

  • http://opinion-nation.blogspot.com/ philipmach

    gus2000, I like this one: “This is probably because all the code from 15 years ago is still in some Vista DLL somewhere.” Since Mac OS X is based on UNIX, there’s probably some code in there that’s 30 years old. And probably a good thing too. Software development is going backwards. It’s too easy to pass off wrapping layers of cruft around an old idea as innovation. Microsoft is of course the master at that.

    I don’t quite believe the story about each release of OS X being faster. I suspect some of the cleanups an install does creates that impression. I’ve genuinely seen this with Linux, but not so convinced Apple has it right yet.

  • MikieV

    @roz & labrats5

    Thanks, for the replies

  • anonymous500r

    Quote from Apple’s website: “Taking a break from adding new features”. Doesn’t adding an entirely new Exchange-style component seem to be a new feature? I think Apple’s fully thinking of delivering plenty of new features in 10.6 but simply doesn’t want any of us to know about them just yet. I think a revision of Aqua is the very least they will do – and it would be nice to mount SMB shares on logon.

  • http://coderad.net StrictNon-Conformist

    There’s absolutely no logical reason to expect that “Grand Central” will remain in the non-portable version of OS X only: it seems the way things are going, even portable devices are going multicore, because that’s more power-efficient (in most cases) for the amount of computational power. So, too, OpenCL will also end up going into the iPhone/iPod Touch: GPUs for the mobile space are likely to continue to get more powerful, within reason, and are very well-suited to the tasks they do. Combine that with the PA Semi purchase, who knows what customized hardware will end up in the mobile space that’s ideally suited to various tasks, all transparently used (regardless of implementation details) from various system and user applications?

  • http://johnsessays.blogspot.com John Muir


    You’re quite right about the enterprise. I often agree with the general mood out there that Apple scarcely even consider it a viable target, and for the most part they don’t lead me to reconsider. But ActiveSync, although a direct consumer feature, is a step in the right direction and I think Daniel may be right on this one after all.

    Build it and they will come. Well, some of the time. Better yet: bring the people with everything else you are doing. That is Apple’s ongoing strategy, from one success to the next.


    I hope OlsonBW took you up on that. Now’s the perfect time for iPhone development. As I’m out in Scotland, I’m taking the more hands on approach myself, slowly figuring out for myself this whole Cocoa thing along with Objective-C and even frankly object oriented code in general! So far so good. Fortunately my own top app idea is easier to pull off as a singleton, and for that reason I’ll keep it under my hat.


    Differentiating hardware as well as software is where it’s at, now that Apple have PA Semi in house. Future iPhones could turn out to be so bespoke that copycats could be left years behind. If there’s something that Grand Central sounds like to me: it’s a unifying measure to suit OS X to a very wide range of hardware. With Nehalem coming soon and Atom and ARM likely to be in a close fight in a few speedy generations, there is great advantage in owning just such a technology.

    Alas it won’t run on my G4. But I’m not complaining: it’s been a good five years at the leading edge. Wayne Gretsky and all of that!

  • retnuh

    I’ve been waiting for Apple to do something custom to enhance at least video encoding, like custom chips or using the GPU. This idea was reinforced when they bought PA Semi, since they control the whole stack there’s no reason why a mac shouldn’t have hardware h.264 encoding. They’re going in a similar direction with OpenCL which is probably the smarter solution as GPUs get faster and if it’ll automatically tune for multiple cpu cores as well. I’m curious if OpenCL is based off of Intel’s Threading Building Blocks framework, I wouldn’t be surprised. And they might be able to take it further with changes to Objective-C. I hope they release a lot more info, non NDA, soon.

  • Michael

    I have to hand it to you Daniel… you’re amazing to be able to keep up this effort to constantly defend Apple’s position in the past few years… because it seems no one else will :( anyway, thanks for insightful comments on all the different decisions Apple makes, it makes living without Windows a paradise :)

  • dicklacara

    @John Muir
    OlsonBW and I are in Contact & I passed his idea to the developers… waiting to hear back, as they were attending WWDC. I sure hope they can do this.

    I agree that the stars seem to be aligning and NOW is the perfect time to be an iPhone Developer– rather a Developer for the first member(s) of the Apple Personal Companion platform.

    I believe that in 3 years (4 at most) we will look around and the laptops and desktops of today will be gone, just gone! Not replaced with a more powerful computer– rather, with with a better way of doing things…

    The WWDC demo of medical imaging is a prime example: The CEO said that the “platform” caused them to change their business– what they produced and how they delivered their product to their customers and to the end users.

    That is profound!

    There is something about the way this amazing platform works:

    1) I want to do something, tap, tap
    2) It gets in my face, monopolizing my attention
    3) I tell it what to do, tap, tap
    4) It does what I want and backs off

    Simple, Elegant, Productive!

  • obiwan

    Grand Central may help developers to better exploit the advantages of multicore/multithreaded hardware, so we may see faster applications.

    It probably will not make the OS itself automagically run faster than other OSs on the same hardware. Apples OS engineers already know, how to make best use of multicore HW, like most other OS makers do too (at least they should). Solaris for examples scales well beyond 64 CPU systems. If Apple manages to make 10.6 significantly faster, it will probably be due to more conventional optimization technics.

    Anyway, if they take their time to focus on stability and performance, this is a good thing!

  • Pingback: Mac Fanatic » Why Snow Leopard is Indeed a Major Release()

  • Pingback: What to expect from Snow Leopard « Bassheadtech.com()

  • Pingback: Apple’s open secret: SproutCore is Cocoa for the Web - Apple.CooOne.Com()

  • Pingback: Blog Mac - Notícias e rumores que você quer()

  • Pingback: Myths of Snow Leopard 1: PowerPC Support — RoughlyDrafted Magazine()

  • http://ephilei.blogspot.com Ephilei

    Why assume there are no new end user features just because they’re not announced yet? Being so marketing focused, I can’t believe Apple won’t have something – even if it’s as simple as pure eye candy.

    Why QT X instead of 8? Marketing?

  • Pingback: Myths of Snow Leopard 3: Mac Sidelined for iPhone — RoughlyDrafted Magazine()

  • Pingback: Snow Leopard - koniec z PowerPC - Strona 8 - MyApple.pl()

  • Pingback: R2X2 - Blogs and News for real nerds.. » Details on OS X Snow Leopard Emerge()

  • Pingback: Ten Big New Features in Mac OS X Snow Leopard — RoughlyDrafted Magazine()

  • Pingback: Myths of Snow Leopard 4: Exchange is the Only New Feature! — RoughlyDrafted Magazine()

  • Pingback: Myths of Snow Leopard 2: 32-bit Support — RoughlyDrafted Magazine()

  • Pingback: It’s a Linky-Dinky Day « Nuke ‘m Hill()

  • Pingback: AppleMania.info » Mac OS X Snow Leopard terá, sim, muitos novos recursos()

  • Pingback: Introducing Snow Leoprad | Яблоки Arwen()

  • Pingback: Distribute This » Virtualizing Mac OS X Server under VMWare Fusion 2 (Beta 2)()

  • Pingback: Inside iPhone 2.0: the new iPhone 3G Software — RoughlyDrafted Magazine()

  • Pingback: What’s new in Mac OS X Snow Leopard? « Bryan Gess()