For a while the blogosphere has been running its mouth off about how “Snow Leopard is just a service pack Apple has the gall to charge for.”
TechRadar.com reports, “Despite all of the necessary ‘under the hood’ improvements in Snow Leopard, this release has the inescapable air of a service pack about it.”
Gina Trapani says, “Even though David Pogue says this is an ‘uninformed wisecrack,’ I’m still calling Snow Leopard a service pack. ‘Snow Leopard Fixes Leopard’s Bugs’ is not the headline you’re going to see in the NYT or WSJ.”
Merlin Mann snarks in this tweet, “Wow. 10.6. If some of these ‘refinements’ were any more ‘subtle,’ they’d seem kinda like ‘bug fixes.’”
To a certain degree I can see their point. We’re used to Steve Jobs getting on stage and touting an OS upgrade with 300+ new features as well as moving any purchaser to the front of a liver transplant line. Frankly, the parts of Snow Leopard the average user would notice are incredibly pedestrian. Boring, even. Apple priced it a boring, $29 price point to make it a no-brainer.
I’m going to go through some of the marquee features and break them down into areas I think would be in a service pack, could be in a service pack, and are too invasive for a service pack.
Service Pack Candidates
Just to show some partisanship, I’m going to reach across the aisle and tell you the features I think Apple could have released in a 10.5.x update: the new Exposé and Stacks, signal-strength meters in Airport, quicker Time Machine, QuickTime X, hi-res iChat and improved disk eject. All of these could have been added into the eight sub-releases of Leopard. Maybe there’s something to the theory that Apple held them back just to have something to sell users on. Maybe there were two gunmen in Dallas, too.
Built-in support for Exchange: I know Microsoft dot-released the new Exchange features into Entourage, but I’m labeling this one borderline because I’m not sure the underlying code in Mail, Address Book and iCal would have been feasible for a .x release.
Rewrite of Finder: For the same reasons above, I’m not sure the Finder rewrite could have been done in a service pack. I’m erring on the side of caution and theorizing as a central part of the OS, it needed to be done in an upgrade.
Not Service Pack-able
The 64 bit transition, Grand Central Dispatch, and Open CL: I believe these could not have been done as a service pack. They are too central, too invasive, and are also key parts that require the Intel chipset to work. Since Snow Leopard will not run on Power PC machines, and I doubt Apple would have ever released a dot upgrade to Leopard that didn’t work on both platforms, to me these are the features that required an upgrade.
The perception problem is while these are likely to be the most important parts of Snow Leopard long term (once developers upgrade their apps), to the average user they are seriously unsexy because we don’t see them.
I think Bertrand Serlet’s comment at WWDC ’90 “[Snow Leopard] is a better Leopard” has been misinterpreted as Apple’s way of saying Leopard had some issues and needed some serious love. I don’t read it that way. I think Apple is fairly happy with how Leopard turned out — at least once it got a lot of the Spaces issues worked out. I do agree that Leopard was released too early, but by 10.5.5 or so it was in fine form.
With Snow Leopard, Apple wanted to refine some of the user interface elements while building an underlying architecture that will continue to take advantage of Core Duo 2 (and better) chipsets and decent video chipsets. When you look at the enhancements Open CL can give, it’s no mystery why Apple got away from the crappy integrated chips in even the low-end products — it wanted all Apple products to be able to take advantage of it.
People claiming that Snow Leopard is just a service pack need to look past the handful of interface improvements and focus on the underlying architecture. It may not be a sexy upgrade, but I think it will be a very important upgrade as apps are upgraded to take advantage of its features.