Three principles form a foundation to the graphical user interface: Discoverability, Stability, and Visibility. They stand in stark contrast to MS-DOS and the earlier generation of interfaces, and their presence swept all of those others away. All three principles were so ingrained in the culture, so absolutely inviolate, that I eventually dropped all of them entirely from my list of core principles as no longer necessary to mention. (I also don’t mention the need to breathe. Some things you just figure people know.) I eventually started talking about Visibility again, but only in regards to the web, where navigation is inherently invisible.
The old interfaces persisted as long as they did because they worked just fine for the people who created them with their wrongly assuming that everyone else in the world shared their exceptional memories, off-scale IQs, and unbridled joy at the challenge of overcoming abstract, invisible interfaces. For many years, interfaces-for-the-rest-of-us ruled the world, and we let our guard down. Now, a new generation of people, of equally high IQ and lack of understanding of how different they are from the rest of the population, are once again creating interfaces that are a joy for them and a continuing frustration for others. Stability, Discoverability, and even Visibility are now both being widely violated. I can’t blame the newcomers because, after all, I pulled it out of my own “history book.” All three will soon be back in a newly-revised version of my First Principles.
Meanwhile, this month, let’s look at just one victim of this movement away from the central underpinnings of visual design, the Predictable Target, starting with an old friend, Fitts’s Law.
Fitts’s Law can accurately predict the time it will take a person to move their pointer, be it a glowing arrow on a screen or a finger tip attached to their hand, from its current position to the target they have chosen to hit. You can learn all about it in my riveting article, “A Quiz Designed to Give You Fitts,” but what the article doesn’t cover is a bit of history.
Paul Fitts was not a computer guy. He was working on military cockpit design when he discovered his famous Law. Paul Fitts never had to deal with the the issue of stability because stuff inside aircraft cockpits is inherently stable. The few things that do move only do so because the pilot moved them, as when he or she pushes a control stick to the side or advances the throttle forward. The rest of the targets the pilot must acquire—the pressure adjustment on the altitude indicator, the Gatling gun arm switch, the frequency dial on the radio, and fuel pump kill switch—stay exactly where they were originally installed. Everything in that cockpit is a predictable target, either always in the same place or, in the case of things like the throttle, within a fixed area and exactly where you left it. Once you become familiar with the cockpit and settle into flying the same plane hour after hour after hour, you hardly look at your intended targets at all. Your motor memory carries your hand right to the target, with touch zeroing you in.
The early graphical user interface researchers came to the conclusion that GUIs, regardless of their inherent ability to slide objects around at every turn, should maintain targets with all the stability of an airplane cockpit. For years we followed that course with great success.
Let’s look at something that’s quite the opposite. Ever had someone do you a favor and “straighten up” your desk, room, or, worse, entire house? When you straighten things up yourself, you have a reasonable chance of remembering where you moved everything. When someone else does it, you haven’t a clue. And that should be a clue to us. When we unnecessarily move stuff around behind our users’ backs, we are causing trouble. It’s going on now, and it’s time to stop.
Case study: Safari and Firefox tabbed browsers
Firefox tabs on the Mac at the time of this writing start out equally sized, only shrinking when there is no more room across the width to fit, and then only so much as is needed. For the longest time, if you want to hit the second tab, you could return to the exact same spot on the screen and do so without fear of hitting either the first or third tab in the process. That second tab is a predictable target.
By the time habitual Firefox users consciously elect to hit the second tab with their mouse, they’ll find the mouse already hovering either over or almost right over the target, their subconscious having jabbed the mouse in the direction their motor memory suggested to bring them there, similarly to the way a pilot’s hand will head for a control. Only the absence of a physical device at the destination then requires the user look at the screen and make a small conscious correction, as necessary, to bring the mouse perfectly over the target before clicking down with the mouse. Both that initial, habitual jab and the correction-to-target follow the parameters of Fitts’s Law. Because it’s both a predictable and large target, the acquisition time it very short.
Now, let’s look at Safari. What Apple does is to immediately divide the entire available space between the currently-opened tabs. If there’s one tab, it stretches the full width. When you add the second, it opens on the opposite side of the window from the first, as seen below, only moving into the position seen in the Firefox example after two more tabs have been added.
If you just look at this from the standpoint of Fitts’s Law, this seems like the ideal solution, as you always have the largest possible target, and the bigger the target the faster you can acquire it. However, the Firefox target is already plenty big, and, much more importantly, before you can acquire any target, you must first know where it is, and that’s where this scheme fails badly: The location of the target is not predictable because the second tab moves around so radically dependent on the number of tabs on the screen.
“Ah,” you might say, “but it is predictable as long as you know the rule! Yes, it is. And that rule ends up dictating the following four-step procedure which replaces Firefox’s one step (step three, below) procedure:
- Cease thinking about your task entirely.
- Look carefully at the tab bar and find the second tab wherever it may be right now.
- Go there and click it.
- See if you can pick up the thread of your thoughts somewhere near where you left off and continue working.
This, of course, is not the end of the world, more like a minor, if constant annoyance. And it wouldn’t be a big deal if it were an isolated example. But more and more of these interruptions to high-level cognitive processing are popping up all the time.
The Apple Dock has such complex rules for the location of objects that I abandoned using it as a test case because the explanation was going to be so lengthy. And I can guarantee you that few Apple users, including myself, could recite all its rules to you. Instead, the “working rule” is that “stuff dances around all the time, so if you want something, just start scrubbing with the mouse from one end to the other and you’ll find it in there somewhere. Probably.” (Apple makes the titles of even identical-looking documents invisible until you scrub over them so the dock will look prettier.)
In Safari, it took me two years to learn the rule for printing a PDF document. You have to click on a target that is always in the exact same place which would make it completely predictable except its invisible I only discovered it even existed when I finally set aside an hour of my time and chased the solution down on Google. Invisibility and predictability don’t exactly mix.
[box]Why pick on Apple?
Two reasons: First, I use Apple almost exclusively, so I am subjected to examples every day. Second, Apple was the first to break the Predictable Target rule. Because their designs are so otherwise brilliant, they’ve been able to get away with it, frustrating their users, but not driving them off. However, others far less skilled are now following their bad example, seriously rolling back the clock on user experience. If you want to send me other bad examples from other people big enough to pick on, please do.[/box]
Several fundamental problems arise when you replace Predictable Targets with the kind of shifting objects governed by rules we’re seeing today:
- Users are often expected to infer these rules on their own, sometimes with nothing but a Google search to come to their aid. Users won’t do this.
- Users are always expected to memorize the rules, keeping all the conflicts straight in their heads. Users can’t do this.
- User’s “hands” don’t “understand” rules. Users “heads” can understand rules, but “hands” work from habit, not intellect, so having shifting locations driven by rules will guarantee slowdowns as users set aside what they are working on to consciously work through the problem of shifting target acquisition.
These problems are why the pioneers of graphical user interface adapted the exact same principles as airplane designers before them. If you want people to hit targets, you, of course, make them visible, but you never, ever move them around. (It’s also why the first thing fighter pilots do, when entering combat, is move their planes around as rapidly and as randomly as possible—kind of like objects in the Apple Dock—so their enemy can’t smite them.)
Supplying predictable targets does not preclude you from having dynamic, exciting screens. It just means that the things users have to find and touch or click, over and over again, should not move around. It means that when a user is finished using something the first time, the user should decide where it will stay in the future, not you. It means that the target is visible at all times. They don’t have to get near it before it mysteriously appears. It means putting the Continue button or its equivalent in the same place on every page. It also means, when you want the user not to continue without really considering what they are about to do, that you put it in a different place.
Both Safari and Firefox compress their tab widths beyond a certain count so that more tabs can appear on the screen. This goes against Predictable Targets, but is also an example of good design. Why? Because the time hit from having to go fetch a tab that is currently off screen is greater than the time hit from having to make a small course correction to get to an existing tab that is on screen. If you do the study, you’ll find there is no appreciable time advantage to getting to the half-window-sized Safari tab shown above vs. the smaller, but still very good sized Firefox tab, but there is a considerable loss that an experienced user will face because of Predictable Target. That’s why Safari’s decision to start at full-width is bad and both browsers’s electing to compress tab size to enable more tabs to appear on the screen is good.
Predictable Target should appear high on your list of mandatory rules, only to be violated when it can be proven that another consideration, in a particular circumstance, will result in even greater productivity.
It’s becoming popular now to speak of the visual interface as one whose time is passing, to be replaced by voice and glance and who knows what. I supposed that’s going to happen just the way the mouse so successfully displaced the keyboard and TV got rid of radio as well as movies.
Yes, the fact is that old technologies do tend to persist, with the new taking their place beside them, and funeral plans for the graphical user interface may be a bit premature. Until such time as either the GUI passes on to its reward or humans spontaneously evolve to all think like engineers, Discoverability, Visibility, and Stability will continue to be vital to people’s comfort and success with visual interfaces. Predictable Targets, lying as it does at the confluence of these three principles, will likewise continue to be vital to people’s comfort and success.
If you happen to cross paths with one of those people who seems to not understand Predictable Targets, please send them a link to this article. They may switch occupations one day. You wouldn’t want them working on an fighter plane cockpit, deciding between sorties it would be a great idea to flip the position of the Gatling gun arm switch and the fuel pump kill switch. Could lead to trouble.
When you visit a forum, you are visiting my home. You will not see personal attacks on myself or other writers here because Siri automatically forwards them to the writer’s mom, along with a letter of explanation. My apologies. She’s rather strict about this. My long-time editor, John Scribblemonger, will then publish comments that are on point, but may edit for brevity and clarity. This being “asktog,” I will then often chime in, even if not explicitly asked.
I hope you will find the result worth reading, as well as joining.