When it comes to Apple's next operating system, there are known knowns, known unknowns and -- well, you get the idea.
Apple is gearing up to release Mac OS X 10.6, also known as Snow Leopard, on Friday -- beating its own announced plans to ship the operating system in September. And in a break from the recent litany of whiz-bang, arresting-gizmo OS releases like Leopard and Microsoft's Windows Vista, Apple readily acknowledges that its new software will offer few new user features. Instead, it will pretty much use the existing face of Mac OS X and make all of the big changes underneath. (It's the same tack Microsoft has taken with Windows 7.)
Perhaps in an effort to quiet skeptics, Apple has been a lot more forthcoming than usual about what's coming in Snow Leopard. Buyers can reasonably expect what the company has outlined both at keynote presentations and on its Web site since announcing the new operating system last year.
What we know
Apple unveiled Snow Leopard at its Worldwide Developers Conference last year, and it released an updated beta at this year's show. Granted, those betas are for developers and testers only -- and they're covered by tight nondisclosure agreements. But that hasn't stopped leaks to Apple fan sites in recent months. Keeping the OS pretty much under wraps is a smarter move than going the public beta route, as Microsoft has done with Windows. Most of the changes in Snow Leopard involve seriously under-the-hood technologies; Apple doesn't want people judging an insufficiently tested, possibly buggy operating system in public beta form, especially one that looks no different than the one they're already using.
Good luck drumming up sales that way.
In fact, Apple last released a public beta of an operating system in 2000, when it gave users a glimpse of the very first Mac OS X. (The final, formal version debuted in March 2001.)
Eight years later, Bertrand Serlet, Apple's senior vice president of software engineering, is already looking ahead, stressing that "Snow Leopard lays the foundation for thousands more [new features]." He also said that Apple "hit the Pause button on new features."
What he means is that snazzy new end-user features, such as Time Machine in Leopard (and the new Aero UI in Vista), are missing this time around. That doesn't mean Snow Leopard won't be an interesting, and perhaps vital, upgrade for many -- especially for just $29. There are a number of small tweaks that users will appreciate.
Deepest under the hood are the twin technologies of Grand Central and OpenCL। Both should be invisible to users, yet both should be able to leverage pervasive hardware changes of the last few years to make everything just a bit बेत्टर

The first, which Apple has also been referring to lately as Grand Central Dispatch, moves the responsibility of handling threads -- different program processes -- away from the application itself to the operating system. This ends up being much more efficient, allowing more things to get done concurrently, with fewer bottlenecks at the application level. In addition, the OS-level threading means the computer itself can spread the work out to all available cores; the computer knows how many processors and cores it's sporting. The end result should be smoother applications and some measure of speed.
Of course, applications will have to be updated for this. But Apple has built Grand Central-aware tools into Xcode, the company's development environment for Mac OS X. Sure, the scope of reworking software will vary for each application, but the early buzz is positive and it seems that Apple has committed to this technology going forward. This is one of those tweaks that should pay dividends long into the future.
OpenCL (Open Computing Language -- I know, how generic), in theory, parallelizes the computing environment of any Snow Leopard Mac with qualifying video cards by offloading computing-intensive tasks to a computer's graphics card. Adding them to the mix taps horsepower that was previously left unused. Apple notes that some of those cards may be "capable of over 1 teraflop -- as much [computing power] as the room-size ASCI RED supercomputer of just 12 years ago."
Fortunately, as Apple points out, developers will only need to rewrite the most intensive parts of their applications, such as data modeling or video rendering. Tools for this are also built into the new version of Xcode, and OpenCL is based on the familiar C programming language.
Those are both unseen changes. But Snow Leopard will also feature QuickTime X, a welcome modernization of Apple's venerable media wrapper technology. Not only does it have a name like a supervillain, but it sports a new minimalist interface and should be able to take advantage of GPU acceleration, capture video from cameras, stream online content better, convert media for iPod/iPhone use, and maybe offer enough editing features to eliminate the need for a $30 QuickTime Pro license.
The big news for enterprises is that Snow Leopard will include baked-in support for Microsoft Exchange (though it's only the 2007 version, so far). This is something no version of Windows has, I should add. Sysadmins rejoice.
When properly configured, this should allow Snow Leopard users to see all their corporate Exchange data -- e-mail, schedules and contacts -- in Mac OS X's Mail, iCal and Address Book। And the standard Mac OS X technologies should apply to all imported info: Spotlight can be used to find a "lost" e-mail message, and QuickLook will open it on the fly when you find it.
One thing missing in Snow Leopard: support for pre-Intel Macs. Come on, you knew it was going to happen sometime, and it's been more than three and a half years since Apple made the CPU switch. Still, I'm sad that the perfectly capable Power Mac G5 I've used for years won't be joining this bright, shining future. And I know there are many such workhorses in graphic design studios, video production houses and other businesses still happily chugging along. That's progress, I guess.
What does it mean?
There's a lot packed into Serlet's statement about Snow Leopard being a base for future development.
Certainly Grand Central and OpenCL could help bust open performance benchmarks down the road. Various developers who have talked obliquely about the changes have said that Grand Central-specific changes in Apple's Xcode have made the onerous task of programming for multiple cores much easier. That bodes well for speed bumps on current Intel-based Macs (at least, all but the earliest ones) and huge boosts on future hardware.
OpenCL's recruitment of GPUs for general computing tasks is a popular idea for both Macs and PCs; we'll see how that plays out. Perhaps this could finally give Apple the poke it needs to get state-of-the-art GPUs on Macs. (Historically, gaming has driven this, and Apple has lagged far behind the Windows world in that area.)
It's worth noting that Adobe has said that the next versions of its Creative Suite and Lightroom will be Intel-only, also. John Nack, who follows Adobe, puts it succinctly: "By the time the next version of the Suite ships, the very youngest PPC-based Macs will be roughly four years old. They're still great systems, but if you haven't upgraded your workstation in four years, you're probably not in a rush to upgrade your software, either."
As always, if every picosecond counts, you have to weigh your own cost-benefit ratio before buying Apple's latest and greatest.
And how great the latest will be remains unknown for now. A lot of implementation has to take place before performance gains appear. But it's good to see these technologies going into consumer systems, along with the developer tools for them.
Exchange support
I can't say much about what to expect in terms of Exchange support and what that could mean, since I don't use it and have never had to deal with it in an enterprise/corporate environment. It should be a boon to Mac users in Exchange-centric companies, but whether it helps Apple penetrate further into that market will depend a lot on the implementation. Macs gained built-in support for Active Directory a while back, but I know IT managers who have had serious problems with it. Then again, many IT managers have also had problems getting individual Windows boxes well integrated into their networks.
Tim Bajarin, longtime Mac analyst and president of Creative Strategies, supports the idea that this is a Big Deal for Macs in the enterprise.
"For many, especially potential enterprise users, the integration of ActiveSync and Exchange server support is going to be the most interesting [change]. Now Active Sync is part of Snow Leopard itself and delivers full synching functionality of Exchange server with all of Apple's key applications such as iCal, [Address Book] and, most importantly, Mail," he said. These two additions "will make the Mac more attractive to IT and business users at all levels and should help Apple gain some ground with the Mac in mainstream business markets.
This could be one reason why Microsoft has announced that the next version of Office for Mac, due next year, will include Outlook for Mac, replacing the Mac version of Entourage. Those hooked on Entourage, the company said, can use the Web version.
No doubt, Apple's decision to focus on stability and speed sounds better than saying, "We're fixing mistakes we made in Leopard." But a lot of these changes fall under what has long been called on certain message boards FTFF: Fix the -- ahem -- Finder. Some of the complaints historically have been technical -- why isn't Apple's crown jewel of a user interface written in its top-seeded API, Cocoa? And some have been functional -- why can't we have this feature that worked oh so well in Mac OS 9?
Apple's answer: Snow Leopard will finally have a Cocoa-based Finder. Apple intimates that this will improve responsiveness (bring more of "the snappy"); technically, this will make the Finder 64-bit-aware and enable it to take advantage of Grand Central, at least on systems with the minimum GPUs. The switch to Cocoa also brings some welcome services to the Finder, such as bidirectional text support.
Snow Leopard won't be the second coming for those who miss the halcyon days of the spatial finder. (Google those two words to get a sense of the discontent Mac OS sparked.) But it will dial back some of the less successful, we'll say, interface tweaks in recent OS revisions, and add a few that are useful: a Put Back command for the Trash; easier assignment of applications to Spaces; and navigation of folders in Stacks and the ability to scroll through their complete contents in grid view. Neither of those will replace pop-up folders, but both are welcome fixes.
A foundation for growth
Personally, I'm glad Apple put the brakes on marketing-driven feature bloat. Going in and overhauling the foundations and frameworks is a great idea -- and something Microsoft is partly doing with Windows 7 in an attempt to woo disgruntled Vista and satisfied XP users.
I'm also glad that Apple is charging a relatively nominal price. Buyers might feel stung if they had to pay $129 for what a lot of users would see as nothing more than a lot of plumbing fixes.
The proof of the pudding is in the eating, of course। If Apple's goal with Snow Leopard is to lay a foundation for future growth, a lot will depend on adoption, both by users and developers. The more of each, it seems, the more benefits to all.