They’re a bad idea.
If you have an iOS developer account then you can enable multitasking gestures which allow you to quickly move between applications or back out to the Springboard. Neven Mrgan argues that they’re essential. I disagree.
I’ll agree that they’re very helpful, that they provide a natural way to navigate between application contexts and that they’ve made switching between tasks feel less of a mechanical affair (involving, literally, pushing a physical button) and fit more neatly in the mode of the smooth interaction model of the iPad. They’ve even got cute effects when you’re zooming back out to Springboard or if you reach the very last application while swiping sideways.
These gestures are activated by using four or five fingers on the iPad screen at once. Left and right swipes with four fingers change applications to the previous and next applications as visualized in the task switcher along the bottom. A four finger swipe upwards brings up the task switcher while a four finger pinch will shrink the application window back down to nothing while bringing Springboard back up. In a particularly nice effect you can zoom in and out and control Springboard scattering the icons. So, nice work on that, Springboarders.
This all sounds wonderful but I still think they’re a bad idea and shouldn’t ship enabled by default. The problem isn’t that they’re not handy (zing), rather that they break what I feel is one of the key wonders of iPad — it becomes the application that is running. These multitasking gestures add a set of interactions that relate not to what is on the screen but to an abstract higher-level of functionality. The touch screen is now an input into two systems: the application and the operating system. Despite the utility I believe this is a step backwards and certainly a trade off I’d be hesitant to make so early in what will undoubtedly be a long-lived product’s life cycle.
Enabling this functionality for all applications will break a few (indeed, right now this includes Apple’s own brilliant GarageBand) and limit many others. Games, especially, I can imagine being limited by making these a system wide gestures. What about a game where you and a friend grab a bunch of objects and need to slide them into a bin on the left side of the screen? You grab two, she grabs two, you drag left and find yourself task-switching to Mail. Or even just a GarageBand style keyboard UI where sliding up and down will bend the note as you hold the key down?
There are a few possible solutions. First, design your applications around only allowing for an extremely limited set of touch inputs. The hardware on iPad can support up to eleven simultaneous touches. Now, sure, odds are you’ve got about ten fingers but thinking of iOS being limited to iPad size devices is silly. What’s stopping a Cintiq style iOS device a few years down the road? The second solution to enable these multitasking gestures is to allow applications to opt out. For a number of reasons this is an awful idea, not least of which is you’ve added a meta-command layer to the whole operating system except in a few cases where it won’t do what users expect it to do and they’ll need to work out for themselves where it applies and where it doesn’t.
A third solution is less elegant than the current implementation but presents fewer issues. Keep all these multitasking gestures, enable them by default, but require the Home button to be pressed while performing them. Now, obviously, this breaks some of the magic of the gestures in the first place but it has the benefit of telegraphing the input mode change to the user (nothing but the system ever deals with the Home button), doesn’t break any software and doesn’t limit any future designs.
Ultimately I think bezel gestures or some other style of extra hardware support would be a great match for this sort of functionality. I’m just not with Neven on this one. Cementing something nifty now is too limiting for the future, even if it does have a great visceral and natural feeling.