Some theory behind Mac OS X's menubar

Windows keeps the Start button, taskbar and system tray at the bottom of the display and a menubar in every window. Mac OS X keeps one main menubar at the top of the display, with a 'dock' of larger icons that take the place of the Windows taskbar at the bottom of the display. Linux, for the most part, seems to prefer the Windows UI, typically using a taskbar-like system with menubars again in every window, but through the power of Open Source, you can do just about anything you want to the Linux UI to make it feel more like home. Some people find one approach more useful, while others prefer a different side of the fence. While the debate surrounding one's OS preference isn't showing any signs of subsiding, we thought it might be useful to offer at least a little insight and theory into why some fundamentals of Mac OS X are designed so differently.
One of the basic principles that informs the Mac OS X menubar is something called Fitts' Law, which I first learned about from John Gruber of Daring Fireball in a post here. To keep things brief, however, I'll just quote a short introduction from the Wikipedia:
In ergonomics, Fitts' law is a model of human movement, predicting the time required to rapidly move from a starting position to a final target area, as a function of the distance to the target and the size of the target. Fitts' law is used to model the act of pointing, both in the real world, for example, with a hand or finger and on computers, for example, with a mouse.
To summarize: Fitts' Law is about how far you have to travel to hit a target, and how easy that target is to hit. Apple implemented these concepts (and I'm sure plenty of others) when designing their menubar by pinning it to the very top of the display, not only from a hierarchal standpoint (you can always look to the very top left of your display to find out exactly which app you're in), but also from a 'make it easy as possible to hit this' perspective. You can simply fling your mouse 'up' and you're at the menubar; even if you click on the very top-most pixel above File, Edit or Help, you'll still hit that menu item and activate it. It's a seemingly minor detail, but one that can help quite a bit during one's daily computing.
This concept is also present in other major OS interfaces, such as the Windows Start button; fling your mouse 'down and left', click and you'll hit the One Button to Rule Them All. Mac OS X's Apple and Spotlight menus also function the same way: fling your mouse 'up and to the left', click in the furthest pixel up there and you'll activate the Apple menu; 'up and to the right', and you're in Spotlight.
If anything, the main point we want to get across is that there is typically a lot of theory that goes into the design of an OS and how users interact with it. We might not always agree with the approach taken by one camp or another, but at least people are thinking about this stuff, because even in 2007, computers still aren't that intuitive to some users who have yet to hop on board the digital train. The more thought, consistency and intuitiveness OS engineers design into our software, the easier it will be for everyone to come along for the ride, no matter what side of the car they're sitting on.













Comments
17
Subscribe to commentsLordDaManMay 12th 2007 12:45PM
Fritt's law is only in one dimension. The problem with this argument is no one uses thier computer in one dimension only. You have X and Y coordinates that you use, which Fitt's law does not take into account. There is a derivative called the Accot-Zhai steering law , but i have yet to see anyone make an argument suing that
Matt PettyMay 12th 2007 3:05PM
iTunes in Windows doesn't obey this law - even when maximised, your can't just fling your mouse to the top right to shut it down, unlike most Windows apps. Irony?
Mike CohenMay 12th 2007 2:20PM
The next step should be touch feedback - you should "feel" objects on the screen as you move over them. The Wii is a perfect example of this. Anyone who's used a Wii will notice how the wiimote gives a little bump when you navigate to a button on screen.
brianMay 12th 2007 2:30PM
No offence Mike but i'd think that would be awful.
johnnyg0May 12th 2007 2:46PM
"you can always look to the very top left of your display to find out exactly which app you're in"
This is what I don't like about the Mac OSX interface and its "only 1 menu bar for all opened windows" concept, you actually need an icon in there to remind you which window are using..
I really wish the Osx UI could be more customizable.
Dave ChartierMay 12th 2007 2:50PM
#4: To be fair, that is actually (possibly unofficially) called the Application menu, as it also houses a number of standard Mac OS X functions such as quitting or hiding an app, well as the standard Preferences menu, the About window to find out what version it is, etc. So it isn't *just* a 'reminder for dummies', FWIW.
Kevin M.May 12th 2007 10:19PM
I still love the classic CLI.
BanzaiMay 12th 2007 8:23PM
Too bad the OSX dock breaks just about every rule of Fitt's law, especially when magnification or hiding is turned on.
http://www.asktog.com/columns/044top10docksucks.html
KMay 13th 2007 4:07AM
Ummm #1 -
you're nor really DaMan, now, are ya? hehehehehe
Accot-Zhai is DERIVED from FITT'S law (not Fritt's, buddy) - so without Fitt's, there is no Accot-Zhai. Get it?
And you thought you were so smart. Stick to talking about what you know, OK?
collapsibletankMay 13th 2007 5:41AM
Hmmm. The Start button obeys the law, but none of the minimised apps. Does that infer design, or accident?
hazardMay 13th 2007 7:39AM
"you can always look to the very top left of your display to find out exactly which app you're in"
.. that's because Apple apps are so [visually] derivative it's can be quite confusing to work out which window is active and sometimes even what appliction you looking at!
Dave ChartierMay 13th 2007 11:27AM
#7: That's probably at least part of the reason why hiding and magnification are not turned on by default. These design laws aren't always meant to be followed to the letter; they're typically theories on how some of these concepts should work, and given the young nature of computing and interface design, we have a lot more theorizing and designing to go.
#10: In my opinion, accident; I honestly don't think Microsoft pays much attention to much design theory or standards at all. One only needs to read through the Mini-Microsoft blog to get a feel for how bureaucratic and disjointed their divisions have become.
LordDaManMay 13th 2007 3:45PM
#10
Reading right from the wikipedia link.
"It applies only to movement in a single dimension and not to movement in two dimensions (though it is successfully extended to two dimensions in the Accot-Zhai steering law); "
So even the linked article says the same thing i did. This does not apply to 2d motion, which is the very foundation of all GUI usage.
Even if you just ignored what the law says, in reality it just doesn't fully apply to any sort of real life usage. Do we just target the menu bar? no, we target that and select a menu item. That's two very distinctive actions. Flicking it up there is good according to fitt's law because you are hitting an infinite target, but then the infinite target disappear when you got from one very specific point to another very specific point to select a menu item
kpMay 13th 2007 10:21PM
I always found the Mac OS more confusing then Microsofts when you are multiple applications.
Also, if you are one of those people that likes the start bar at the top of the screen, in windows you can just click and drag it up there, or to the sides. I don't know if you can do the same on a mac.
morganMay 14th 2007 12:19AM
I don't have a cool theory name to back me up, but if you're using the mouse and a drop-down menu bar on any system, you are going turd-slow anyway.
ColinMay 17th 2007 10:42PM
@6
No, iTunes for Windows is a terrible abomination, and Apple should know better. While consistency in Apple apps is definitely better than MS ones, Apple should at least respect MS's GUI rather than foisting OS X onto Windows. (On the other hand, MS spends a lot of time "skinning" their apps, so it could be argued that iTunes is just a terrible skin of a Windows app.)
@10
In W2K and earlier, nothing obey Fitt's Law. When XP came along, as long as you were using the Luna themes, the Start Menu/Quick Launch/Task Bar obeyed. Strangely, the System Tray never did. I'm not sure how it works now.
@13
It made much more sense when the Apple Menu was actually customizable, eg, Mac OS Classic. Windows' Start Menu is somewhat derivative of this old behaviour.
The five quickest locations on a GUI to get to are the four corners and whatever is directly under the mouse. Apple makes use of the four corners as "hot corners" (triggering actions).
Interestingly, while the area under the mouse is utilized by contextual menus, they introduce another level of confusion to GUI design: You never know what's in them until you actually right-click.
John SawyerJun 15th 2007 11:28PM
I'm not sure about the one-dimensional description--doesn't the entire screen constitute a 2D area? I realize that moving the mouse pointer from one point to another onscreen is traversing an (ideally) straight line, but that's not one dimension--that's two, involving an XY axis. Technically, "one-dimensional" is a single point.