Fitts' Law and Infinite Width

blog
21

Whatever happened to force feedback mice? It sounded like a great idea. When you rolled over an application edge, there was slight resistance bringing physicality to the UI.

Seems to me that the “edge” problem with multi-monitors and applications could be improved by simulating a slight bit of physics (with or without force feedback). When mousing over an app, for example, perhaps the mouse slows down slightly over edges and if you keep moving pops over. Or something similar.

Of course, the only way to know if that would be better is to usability test test test!

22

With WinXP they made the Taskbar effectively infinite height. Too bad the buttons still look like they’re offset from the edge, so your brain still thinks “better mouse precisely”.

I ran into a really bad violation of Fitt’s law the other day. I tried out Cyworld because I’m a sucker for cutesy isometric avatar chats. Check out the pagination in this screenshot: http://www.newshutch.com/blog/wp-content/assets/cyworld.png

First of all they only show you 10 items per page, so you have to remember which “page” the option you liked is on. But the absolute worst is that only the number text of the page is clickable, not the entire page icon. After struggling with this on a trackpad, I gave up.

Why are social networking sites always so poorly designed?

23

When mousing over an app, for example, perhaps the mouse slows down slightly over edges and if you keep moving pops over. Or something similar.

Consider WinAmp’s “snap to edge” behavior. As you drag the app, if it gets within ± 5px of an edge, it snaps over to that edge automagically.

I think manipulating the user’s cursor movement speed would be more of a negative than a positive. If the edges of the screen were like “molasses” you had to punch through with your cursor, that would get annoying fast.

24

although I presume [the mac menubar] is focus sensitive

Yes, it is. If you look at the screen shot you will see the word ‘TextEdit’ in the left side of the MenuBar. That’s because TextEdit is the current foreground application. (Much like Windows, the Mac only allows one foreground window and, thus, one foreground application.) If you switch to a different app, the new foreground app’s menus will magically replace the old app’s menus. This prevents the system from wasting valuable screen real estate displaying the useless menus of background applications. :slight_smile:

[S]ince they are no where the window I am working with, I don’t think of them being related.

That’s an adjustment you have to make moving between the systems. It’s like Windows having the close box on the ‘wrong’ end of the title bar. Or driving on the ‘wrong’ side of the road in England.

Second, I have no way of knowing if the menu applies to the dialog box I am working with or the application as a whole.

Menu Items, and entire menus, that don’t apply to the currently front-most window (or even the widget with current focus) are disabled: their name dims and disabled items aren’t highlighted when you mouse over them. For example, the Close, Save and Print items in the File menu in the screenshot. Thus, it is very simple to tell at a glance which menus and menu items apply to the current item/window/task/widget/etc.

Third, I cannot see the menu for the application I want work with unless it is already active. That means moving the cursor all the application, then all the way back to the menu. I cannot imagine doing this on a multi-screen system.

Fitts Law. It takes less time and effort to fling the mouse to the top of the screen and choose a Mac menu than it does to choose a Windows menu. The Windows menu has to be located in the X and Y axes, the Mac menu only has to be located in the X axis, because as Jeff said in the post, Mac menus are infinitely tall.

After some computing experience, a user learns to multi-task, and the Mac menus can be really frustrating.

It’s been my experience that most novice users become acquainted with keyboard equivalents (Alt-F4, Ctrl-W, etc.) long before they start trying to carry out more than one task at a time. But then I’ve never taught intro computer courses, so my results are likely biased or skewed.

25

Except you missed the whole point: to get to the menu, you first have to find and click on the application the menu belongs to, which cuts out a lot of the benefit of having the menu at the top. There’s more than Fitts’ Law involved in this case.

Okay, but is this an actual workflow we’re talking about? Or is it just a hypothetical worst-case scenario? I switch applications with command-tab (alt-tab), not the mouse. I use mainly keyboard shortcuts instead of the menus themselves. Is my heavy use of the keyboard abnormal?

26

Except you missed the whole point: to get to the menu, you first have to find and click on the application the menu belongs to, which cuts out a lot of the benefit of having the menu at the top. There’s more than Fitts’ Law involved in this case.

Also, if we apply Fitt’s Law to both actions (click the application to bring it forward, choose a menu) we get the following:

Windows: click a large section of the X-Y plane to bring the window forward (fast), click a small section of the X-Y plane to choose he menu (slow).

Mac: click a large section of the X-Y plane to bring the window forward (fast), click a small section of the X plane to choose he menu (fast).

Assume that the menu is visible on the Windows window. Then the Windows actions become: click a small section of the X-Y plane to bring the window forward (slow), click it again to choose the menu (very fast).

The problem is, you throw out the ability to click anywhere on the window to bring it forward in order to not have to aim for the menu once the window is active. You haven’t eliminated the slow process of targeting the menu, you’ve just shifted it over to the activate action.

Now, actual lab test would be required to see if (slow + very fast) (fast + fast) or vice versa. :slight_smile:

27

[Peter Palludan]
I haven’t tried Vista, but is it possible to rearrange your applications in the taskbar? The “group by application” was never something I fell for.

You can rearrange the order of your applications in the taskbar through this freeware non-open-source program (available in Japanese and English): Taskbar++

a href="http://hp.vector.co.jp/authors/VA013430/program/taskbarpp/main.html"http://hp.vector.co.jp/authors/VA013430/program/taskbarpp/main.html/a

You mouse over the taskbar, hold down Alt, drag the order of the taskbar app buttons, then release the Alt. Then Taskbar++ does a “trick” to rearrange taskbar app buttons.

It works at least on Windows XP Pro, and is mentioned to run on WindowsNT/2000/XP/2003.

[Jeff Atwood]
I just wish we had more stuff along the top edge!

I was going to suggest moving a Taskbar toolbar to the top edge of the screen, but unfortunately the top most pixel of the bar is not active for clicks. Fitts’ law broken again!

But really, what can Windows programmers do in light of this Fitts law? Make buttons bigger? Because menus can’t be fixed.

28

Vista has that big-ass shiny circle of a start menu that takes up the entire left hand corner

Hooray :slight_smile:

I agree more stuff along the top would be good - can’t you make a new toolbar and put commonly used shortcuts etc up the top too?

29

Mentioned by some other posters — keyboard shortcuts…

I’m no UI designer, and I have no human-computer-interaction training. But I find it terribly inefficient to use a pointing device for everything. Moving between controls using tab and using common keyboard shortcuts is much faster and probably helps with repetitive movement to/from the mouse, not to mention the fine movement to get to the desired button…

(pressing tab all day long isn’t so good either though. I’m going to try a foot pedal next, for that)

30

Here’s a nice article by Jensen Harris on the same topic

http://blogs.msdn.com/jensenh/archive/2006/08/22/711808.aspx

31

One interesting and useful feature that seems to help Fitt’s Law issues is mouse (touchpad) acceleration - the quicker you move, the further the on-screen icon moves when covering the same distance.

This alters the equations so that things which are far away are “just as close” as stuff nearer - if you move a bit faster when going there.

32

Windows: click a large section of the X-Y plane to bring the window forward (fast), click a small section of the X-Y plane to choose he menu (slow).

You don’t need to bring a Windows app into focus before using it’s menus. If you have an application on screen (but without focus), you just click the button you need and it focuses the application and does whatever you clicked in a single operation. Surely it’s much slower to work with an OSX multi-monitor setup in which you need to switch application focus to get the menu to appear.

It’s nice that OSX uses this infinite space for their menus, but this article seems to ignore that Windows uses the same space for quite a few handy things on a maximized application: the title bar, minimize, restore, close buttons, vertical scrollbars. All of these items on OSX float in mid-air and thus make simple window manipulation so much harder.

33

Excellent visual exploration of Fitt’s Law

http://particletree.com/features/visualizing-fittss-law/

34

Ever tried DejaMenu on Mac OS X?

It will create the frontmost application’s menubar as a pop-up menu at the mouse’s current location!

http://www.macosxhints.com/article.php?story=20040426064258161query=dejamenu

35

To take better advantage of Fitt’s Law, you could do as I do and move the Quick Launch tool bar over to the left edge of screen. You can then slam the cursor left to access the Quick Launch items, plus it frees up space for more task bar icons. Also, with Quick Launch now separate from the task bar, you’re free to fill up Quick Launch with many more buttons. I hardly ever use the Start menu anymore because every tool that I run is on the Quick Launch tool bar.

36

“But one unfortunate side-effect of multiple monitors is the removal of some natural edges between adjoining monitors. The cursor now flows freely between monitors; it’s painful to stop the cursor on the left and right edges of the app on the center monitor.”

This is why I take advantage of Windows’ ability to specify my right monitor is 16-32 pixels lower than my center monitor - it maintains the top-right corner boundry and makes it easy for me to click on the close button for windows. I would love a utility that enforced a ‘boundry’ among those first couple dozen pixels at the corner of each display, to allow the corners to work properly, but still allow the monitors to line up nicely. (I have seen… interesting effects when playing games at different resolutions than my main desktop res.)

37

“Fitts’ law indicates that the most quickly accessed targets on any computer display are the four corners of the screen”

Actually, the fastest pixel to click on is the pixel the mouse is already on. Check out the CAD program Liquid PCB (http://www.liquidpcb.org) for a very fast access menu system.

Hugo

38

While I am a die hard mac user, I am often frustrated with the dock. When using a program where I have to bring the mouse to the edge of the screen (i.e. photoshop), it is more often than not that another program is accidentally opened.

Windows: click a large section of the X-Y plane to bring the window forward (fast), click a small section of the X-Y plane to choose he menu (slow).

You don’t need to bring a Windows app into focus before using it’s menus. If you have an application on screen (but without focus), you just click the button you need and it focuses the application and does whatever you clicked in a single operation.

I know this is a discussion of Fitts’ law, but I feel that a few important facts are being left out. First, for those of us who only have one monitor, having more than one window occupying the screen is infinitely annoying, meaning that clicking on the window itself isn’t an option, regardless of Operating system. Second, while Windows does it better, both OS’s have a taskbar (okay, the ‘dock’), giving a screen edge to change applications. And third, it is infinitely faster to use alt-tab/command-tab to cycle between applications. In fact, I can’t remember the last time I actually clicked to switch apps.

39

Wyatt, that’s an interesting argument.

First, I only have one monitor, and it annoys me to have only 1 window open/maximized. I like having multiple windows open (e.g., different browser windows for different categories of information, or visual studio nunit - no, I don’t want the nunit plugin to muck up my ide, I prefer it being a separate application).

Second/Third - I almost never use alt+tab because it is too slow. It is much faster for me to click on the application in the taskbar. The only exception being if I’m toggling between exactly two application windows often (e.g., copy paste notes into Word when I’m writing a research paper).

40

one so-far unmentioned advantage to the Windows menu system is that Very Often one has to move BACK to the application window after doing something on the menu. By putting the menu bar atn the top (vs. in the same window) this move-back has a much larger distance, and the app window, while large, is not inifite height. More specifically, the pint in the Application Window which needs to be access is potentially quite small - I don’t want to go back to the window just anywhere.

So the nifty inifinite-height thing seems ok until you are done with the menu, at which point you have pout the user into a Very Bad state, Fitt’s Law-wise