Thursday, October 6, 2022

Stage Manager in iOS 16.1 Beta 3

Joe Rossignol:

In the latest iPadOS 16 beta seeded earlier this week, developer Steve Troughton-Smith and MacStories editor-in-chief Federico Viticci highlighted various user interface issues they continue to face from time to time while using Stage Manager, including the dock disappearing when rotating the iPad, content failing to scale properly when a window is resized, keyboard input failing to register in certain apps, and more.

Steve Troughton-Smith:

Today’s iPadOS beta: still able to bring down SpringBoard in the first 5 minutes


No joke, if Stage Manager ships the way it is in this build, it will break iPadOS. I only put a fraction of the stuff I come across on Twitter, but so many little things across the OS and system apps break in weird ways.

Federico Viticci:

I understand that some folks think “this is good enough” for them. If so, great.

Personally, if I open an app and can’t type, or use keyboard shortcuts, or can’t click into a text field, that’s a problem that prevents me from working with my iPad.

Jason Snell:

When I try to use Stage Manager on my iPad Pro, I almost end up with a single configuration: two windows, more or less equally sized, next to each other. I end up so frustrated with Stage Manager that I essentially re-create Split View!


I’m troubled by the fact that Apple doesn’t seem to have a real vision for Stage Manager–or, if what we have right now is its vision, that it’s decided on some weird halfway step.

Steve Troughton-Smith:

I don’t understand why I can create a much more complex/powerful app w/ UIKit on macOS than I can on iPadOS. They built all the APIs, means to make this stuff great — then utilized none of them for Stage Manager. It’s not that they haven’t done the work — they chose to ignore it.

Damien Petrilli:

Stage Manager requires more power than a real windowing system like on desktop OSes, is ultra limited, unstable and with confusing / bad UX.

It’s a disaster engineering and design wise.


Update (2022-10-14): Federico Viticci:

Within the first 10 minutes of the latest Stage Manager beta:

  • Magic Keyboard input does not work unless I disable QuickType suggestions
  • Got app windows being randomly thrown into a different workspace after ⌘-Tab
  • App windows still randomly decide to cover the dock


Sean Heber:

The ongoing Stage Manager drama has been something to see. I can’t remember any major iOS feature of this size in the entire history of the operating system having so many problems for so long so close to a release. 😬

Steve Troughton-Smith:

From another perspective, Stage Manager has delayed every other team on iPadOS from shipping by at least a month, and has barely had any tangible improvement to show for it. It’s still rough and not ready to ship, even if you wave away all of its fatal design issues.

11 Comments RSS · Twitter

I have a theory that Stage Manager started out as a part of the interface for the 'Reality OS'/AR glasses OS, and Apple decided to adapt it for iPadOS and MacOS as well, to allow users to get used to the paradigm before introducing the glasses. If true, this might explain why the Stage Manager implementation in iPadOS and MacOS feels so at odds with those interfaces.

So Macs had multiple windows on 9" 512 × 342 screens, yeah? Wonder why iPad multi windows have proven so difficult? Even in the early days of OS X, computing resources were pretty constrained compared to the bonanza available now. The iOS paradigm of computing was truly built around largely single tasking (iOS had limited multi tasking early on, but it was certainly not a free for all), single app, single screen size interface, with no third party app support, hence the constant rearchitecting around newer use cases, third party apps, larger screens, more flexible multitasking options, and now windowing support on larger screen iOS devices, maybe?

I can tell you that LG (RIP, LG is out of the phone business now) and Samsung Android devices with USB C ports that support DisplayPort over alt mode have the ability to go into a desktop view with multiple windows and everything, even with apps that are not made to understand windowing modes on Android. I think on a smaller screen portable device maybe tiling/snapping windows really is the best compromises. So 1-4 windows snapped into places would be enough but when docked into an external display, then maybe full desktop mode with stacking window manager? It does seem this current Apple solution seems overengineered for what is a pretty simple problem on any modern system, yes?

@Gord L.

So teething pains for the next big thing as it's shoehorned into the existing interface? I mean, that's not a first time. OS X rendering was really slow at first because of all the Aqua stuff, but it rounded up into shape. Somewhat simplified because there have been ups and down along the way, but it was really slow, then quite useful on fairly low end devices.

So I could see Apple looking at the next big thing and kind of forcing people to muddle along until the technology is really ready.

macOS has supported window resizing forever. Before Autolayout and Trait Collections were a thing. And long before SwiftUI. But on the Mac the developer controls the window size not Apple; this makes things so much easier for everyone involved. Developers can set a min size, a max size, and constrain a window to aspect ratio in a way that makes sense for their app.

On iPadOS you have to work with Apple’s predefined window sizes which are subject to change at any time. IMO too much has been invested in “magic solutions” to handle layout that supposedly you will be able to take with you anywhere. Does it ever work?

I haven’t used Stage Manager yet but I’ve heard complaints about windows snapping to certain sizes when users feel like they should be able to stretch them more? If true my guess is for most apps the UI breaks if they don’t maintain certain aspect ratios….so they have to snap the window on you because the Autolayout constraints, like manual layout code, only really work if you plan your design for the actual screen sizes the app will be run on (with the exception of a basic screen with a label positioned in the center, or a scrolling table view). The idea that you could just create these magic constraint objects and they’ll future proof you for new screen sizes was always a joke. SwiftUI probably won’t be able to handle it any better, most likely worse.

In my theory, Apple was looking to come up with a windowing-type multi-tasking interface for iPad, and instead of creating one from scratch, they said, hey let's bring over this interface that our AR glasses team came up with, and bolt that on to iPadOS. And oh while we're at it, why not bolt it on to MacOS too—where it's not really needed in any sense whatsoever, since the Mac already has multi-windowing, 'Mission Control' AND 'Spaces'—just because we can. To get folks used to the paradigm, yes, and also to beta-test this new interface idea as much as possible before rolling out the AR glasses.

Just one theory. Whatever the real story is, there's no way that Stage Manager was a 'Mac-first' idea, since: 1) basically no Apple features that have recently been introduced for more than one of their platforms (or none that I can think of) have been Mac-first (they tend to go in the other direction); and 2) I don't see what Stage Manager brings to the table that 'Spaces' can't already do on the Mac. So it was either designed with iPad in mind, or possibly with their next big platform (AR glasses) in mind. Stage Manager seems so kludgy and borked on iPad, that I gotta think that iPad was an afterthought.

> Whatever the real story is, there's no way that Stage Manager was a 'Mac-first' idea

"I have a theory that Stage Manager started out as a part of the interface for the 'Reality OS'/AR glasses OS"

AR glasses create a huge virtual screen that wraps all around you, which this UI seems like a poor match for.

Wouldn't that mean that apples vision for AR are a bunch of windows?

Sounds bleak.

> there's no way that Stage Manager was a 'Mac-first' idea

Originally, Stage Manager was in fact a Mac-first idea.

I really do think that they should have simply created Mission Control for iPad, Stage manager is clearly a solution looking for a problem.

With mission control arbitrary window size and position would be truly supportable as it brings with it Expose which lets you see all windows in each space. This means that there would be no need for the clunky jumping around that happens with stage manager.

Mission control can provide an integrated environment that supports fullscreen, split-view, slide-over, and resizable apps all at once!

Mission control should be adapted to have the current space (or split view or fullscreen app) always visible when invoked (unlike the collapsed desktop row on the Mac)
Show the apps/spaces/split views in most recent order at the top (older to the left) and then slide-over apps to the right.

Moving windows between spaces/split-views/fullscreen can be done via drag and drop (the iPadOS15 era way of viewing all windows in an app by long pressing its icon in the dock while in a space would also let you drag out an app window from somewhere else).

App windows should always have one and only one home, either in a split-view, in fullscreen, in slide-over, or in a resizable space...

Stage manager is reinventing the wheel, never mind its broken behaviour while using a resizable space, it also breaks the multitasking switcher so badly that I couldn't use it even if it was stable. Stacking app windows was stupid when they tried it on the Mac and its still stupid here.

> Originally, Stage Manager was in fact a Mac-first idea.

Hmm... that's quite a stretch. Cosmetically, that 15-year-old mock-up looks similar to Stage Manager. But in the mockup, each separate "stack" consists of windows of the same app. Not a grouping of windows from various apps. I could see it maybe serving as a visual inspiration for Stage Manager (assuming anyone from the Apple of 2021/2022 even remembered an ancient, never-implemented mock-up), but not a conceptual one.

Point is, in the development cycle of Stage Manager, which is on the order of 1-2 years, not 15 years, the feature as we know it was not primarily introduced to solve an interface problem on the Mac. No way.

Leave a Comment