Profile cover photo
Profile photo
Max Fellmuth

Max's posts

Post has attachment
Max Fellmuth was in a video call. <a class='ot-hashtag' href=''>#hangoutsonair</a>

Post has shared content
Mission Control (in MacOS X Lion) is maddeningly inconsistent, particularly with multiple displays.

I don't believe that any novice users ever used spaces, so by definition, this is a feature for more advanced users. Problem is that they've made this feature harder to use for the intended audience, and I'm not convinced it has any positive impact on usability for the beginner audience.

My main gripe: Multiple displays is a total clusterfuck.

In the old spaces implementation, multiple displays were shown as adjoining thumbnails on each display. I could see, on one screen, what all my displays looked like. I could easily drag any window thumbnail from any space/display to any other space/display. Mission control makes this basic task maddeningly difficult.

With mission control, each display has thumbnails unto itself. This seems reasonable until you consider a common use-case: placing apps on secondary monitors next to other apps.

I've classically dedicated space #4 to "communications": email, twitter, chat, and so on. From the outset, mission control makes that use-case harder by removing my ability to pin specific apps to a space—but I digress.

I have my twitter and email clients open on space #4, both positioned on my primary display. I'm now in space #1, and I realize that I forgot to open iChat. I launch iChat, and then want to position it in the same space as my email client, on the secondary display.

Accomplishing this simple task used to be

1. Open spaces
2. Drag the window thumbnail

And that's assuming I hadn't already pinned iChat to space #4, which would obviate the need for any action on my part.

With mission control, Apple has made this experience hellishly complicated, requiring lots of my brainpower and attention to successfully execute.

First, spaces no longer "stay put" by default. Lion automatically rearranges spaces "based on most recent use," which means that I must seek out space #4 visually, based on my knowledge of which apps are open and how they look, and then mentally transform that into how they would look when scaled down to thumbnail size. Between my existing spaces and fullscreen apps (which get their own pseudo-space), Apple is forcing me to perform a tricky visual search every time I want to find a space.

Fortunately, there's a preference to disable the space-switcheroo behavior, but you can't reorder spaces (another feature lost) and you still have the pseudo-spaces added for each fullscreen app. The visual list of spaces is dynamic, and your mental map of where things are is broken every time you add/remove a fullscreen app.

Assuming that I've identified the correct space, I now want to put iChat's windows on the same space, but the secondary display. Here's where the fun really starts.

Each display shows a list of space thumbnails for itself. The order of spaces is arbitrary, but consistent across displays, e.g. the 3rd space from the left in the horizontal list of thumbnails on your primary display correlates to the 3rd space from the left on your secondary display—regardless of what those spaces are named (Desktop 2, Desktop 3, et al.)

In order to put thing on the secondary display next to a given app, you must now:

1. Identify which space your target app is located on by remembering whether it is on the primary or secondary display, and finding the appropriate thumbnail on the appropriate monitor.
2. Identify the desired space's position in the list of spaces, or alternatively, note the space's name (e.g. "Desktop 4").
3. On the alternate display, identify the matching space, by position or name—a task with significant cognitive cost.
4. Drag iChat across displays to the target space.

Oops! That doesn't work. As soon as I drop the iChat window on the desired space, it snaps back to its current position. There's no visual cue that this operation isn't permitted. It actually provides positive feedback when I try to drag windows into spaces on other displays: as soon as the window enters the space thumbnail, the window shrinks to "fit" into the space, and the space darkens to indicate selection. Unsurprisingly, this is the exact feedback it offers when dragging windows into spaces on the same display—a legal operation that will succeed.

So basically, mission control leads me on, only to fail at the last moment. If that weren't failsome enough, it commits the further sin of offering zero feedback about why it won't permit me to take that action. Double-u-tee-eff.

Apparently, Mission Control won't let you drag windows across displays _at all_—not even within the same space.

Take a moment to let that sink in, and maybe try it out yourself. That's right—you can't move windows across displays in mission control. In order to do this, you have to exit MC and drag the window across.

Mission control jails windows to the display that they're on. You can't drag them to the other display, let alone dragging them to different spaces on the other display.

Back to our simple-in-theory task of moving iChat: here's the simplest sequence of steps I found for achieving the desired result:

4. Drag iChat to the desired space on the same display.
5. Click that space to switch to it and exit mission control.
6. Locate and drag the window across to the secondary display.

Six high-cognitive-cost steps compared to two. I don't think that it's just a matter of "getting used" to the new workflow. The degree of cognitive involvement required to move windows around spaces is so high as to be dangerous to whatever else I was thinking about at the time. It's a flow-wrecker.

Mission control is subtly deficient in other ways compared to spaces. Gone is the ability to drag windows from any space to any space; attempting to grab a window in another space thumbnail will drop you into that space. A rude surprise the first time I tried it, and an annoyance ever since.

In order to drag a window from one space to another, I must be in mission control for the space where the target window is located. If I only know how to get to spaces using Mission control, I now have three operations: enter mission control, click on the desired space (which exits MC), and then reënter mission control. Depending on which space I'm currently in, this series of operations might actually be equivalent to the number of swipes required to reach the desired space.

Once I'm in mission control for the correct space, I can finally move my window. Once again, mission control adds a lot of cognitive bureaucracy to a common task.

Mission control is not all thorns; there are small things that I do find agreeable, but even there success is tinged with UX failure. The ability to dynamically add spaces is welcome, and the implementation isn't bad (phantom space slides in from right when dragging a window). In fact, this is the only cue which hints at my inability to drag windows across displays—the phantom "new space" doesn't appear on other displays.

Unfortunately, controlling the layout of these spaces is not permitted. Wholesale rearrangement of spaces, previously a drag'n'drop affair, is now impossible. The sole way to accomplish this today is by manually rearranging apps. Besides the laborious process of moving each window from space to space, you must subsequently reposition each window you moved.

There's a lot to like about Lion, but Mission Control feels half-baked to me. There's a (weak) argument to be made about how the high-level functionality changes make "spaces" more accessible to the general public, and time will tell if that's true (I don't think so.) Still, The affordance of misleading cues when dragging windows across spaces is a very un-Apple-like UX fumble, and stands out to me as a glaring indicator of some internal failure at Apple. I can only infer that the design team hadn't thought this through to the end, or that the engineering and design teams weren't on the same wavelength. Good design is Hard, and Mission Control is sufficiently different as to be given the "1.0 product" benfit-of-the-doubt.

I don't see this as a parallel to FCPX. None of the breakage I've identified is really in the service of "making things easier for beginners." It's not sacrifices made in the name of appealing to a wider audience. It's just broken, and poorly thought-out.

Mostly, I'm just sad that I have to live with this for a while now; given Apple's past behavior, I don't think we'll see changes to mission control in the near term.


Pistachio is looking for alpha testers. You have to be in Berlin and should be able to swing by our office next week. Interested? Ping me!

Post has shared content
A few Google+ tips:
- You can click on profile pictures to rotate through them. Nice find by +MG Siegler on that one.
- In the stream, you can click 'j' to navigate down to the next item or 'k' to navigate up. I think it's the same keys that Gmail uses, which is probably in turn because +Paul Buchheit uses vi and created Gmail. :)
- If you're sharing a post with a small circle of people, you can prevent resharing. Click the arrow at the top-right of the post and choose "Disable reshare."
- One more: if you're looking for more fun things in your stream, the "Incoming" stream is stuff from people who are sharing with you, but who you haven't added to a circle.

What are your favorite Google+ tips?

Post has attachment
This thing needs serous Twitter integration.

Post has attachment
Max Fellmuth changed his profile photo.
Wait while more posts are being loaded