@globalmoxie wrote a few months ago about his concerns regarding the new iOS5 multitouch gestures. You can read his article here: link
TL;DR of his article
Apple is using the wrong approach by not providing proper differentiation between App and OS gestures. Ideally Apple should have used Edge Gestures, similar to the ones on Windows phones or WebOS. And also, due to the fact you can perform those gestures anywhere on the screen, your app cannot take advantage of them.
That's an interesting and very valid point of view.
I think I can offer a different point of view though.
TL;DR of this article
I think there are also some valid reasons for them to do so.
- enforce simplification of 3rd party apps by reserving complex gestures for non-essential device-wide actions
- uniformization of interactions accross platforms: iOS and Lion share interaction patterns
Canvas real Estate
Regarding Apple's use of the whole app canvas for gestures.
It may not be a bad thing. I'll start by listing a few ideas and show you my point further down below.
Cognitive friction
Gestures like the single finger tap, drag, swipe have very low cognitive friction because very close to the natural interaction we have with objects. We use 2 or more fingers for grasp, to convey a stronger intention, sometimes precision.
Apple has reserved the more complex gestures that require the use of 4 fingers to implement non-essential global gestures.
I don't think it's correct to think about this is a 'logical' way. Perhaps we shouldn't be thinking of the technological details behind the screen, making the separation between App and OS. The reasoning for is that the user doesn't make this distinction. I think it's safe to assume that from his point of view, he's interacting a device as a whole. It has a default state and many extensions he installs.
Apple's global gestures provide an easy way to manage context. Above the App level you have:
- multitask overview - 4-finger drag up / nudge the App
- multitask stack navigation - 4-finger drag / push the App
- home - 4-finger pinch in / shrink the App
If I'm forgetting any, then it probably has more implementation issues than these.
Still in the subject of cognitive friction, I'd like to hear some feedback from people, but I think the 4-finger gestures are so expressive, provide such a great contact level with the device, that anything else than applying an action to the the full canvas wouldn't be natural. That is, outside the context of games and other immersive apps.
Complexity
There's possibly a slight relation to the previous cognitive friction section in here, but yeah, I guess it boils down to it: reducing complexity.
With the complex 4-finger gestures reserved for non-essential interactions, you're left with the most common ones, which is a good thing: simplify your app. Just because you can implement 4-finger drag/swipe/double tap doesn't mean you should abuse it.
Yes, Apple is being very paternalist, again, but I think this might be the resoning behind their decisions.
Also, don't forget there's a greater plan here being put to action by Apple, which is to introduce more direct manipulation gestures in both their mobile and desktop computers.
Lion also leverages 4-finger gestures, also for device-wide actions, like:
- accessing Mission Control - 4-finger drag up
- accessing App Exposé - 4-finger drag down
- switching between Spaces - 4-finger drag
- accessing Launchpad - 4-finger pinch out
- accessing Desktop - 4-finger pinch in
Makes sense they want to reserve these for their own interaction layer. It's hard to miss the similarities between iOS and Lion of these.
Edge UI
As Josh put it, this pattern allows the separation between OS and App gestures, I still stand by my reasoning that this separation shouldn't exist.
From what I've observed so far, this pattern doesn't map very well to the user expectations.
This is based on my observation of people interacting with touch devices. I've noticed people tend to direct the control gestures to the center of the screen, sometimes the bottom if they're lazy, but never initiate a gesture from the edge of the screen.
I don't think there's a proper affordance for this pattern.
Conclusion
Refer to TL;DR.
Input is always welcome.
Do read my other writing on 'Improving gestures in NUIs', where I try to explain the differences between 1:1 and trigger gestures. And how much I prefer 1:1 direct manipulation over triggers AKA button gestures, read the article for more details.