How People Act: Targets, Movement and Control
Podcast · How People Act: Targets, Movement and Control

Physics of Human Movement in UI

by ZumiLabs · 49m
0:00 49m

Transcript

Made automatically from the audio, so it may contain mistakes. Click a line to play from there.

Host 1

Picture this. You are sitting at your desk, fully dialed in, taking one of our AI generated video lessons on the pathway platform.

Host 2

Oh, I already know where this is going.

Host 1

Right. So you realize the instructor is speaking just you know a little too slowly for your liking. So you decide you want to speed the video up. You take your mouse and you glide it diagonally toward the 1.5 times playback speed option in the quarter of the video player.

Host 2

But right before you get there.

Host 1

Right before you get there, snap. The hover menu just vanishes mid-flight. It shuts completely. I mean, it disappears before your cursor can even cross the threshold to actually click it.

Host 2

It is incredibly frustrating. Yeah. And you know the worst part is it completely breaks your learning flow right in that moment. You have to go all the way back to the gear icon, re-trigger the hover state, and try again.

Host 1

Yeah, and usually much slower this time. You find yourself basically holding your breath, right? Treating your mouse like a surgical scalpel just to get to a button.

Host 2

Exactly.

Host 1

Okay, so let's look at scenario two. You are a commuter. You're riding one-handed on a crowded train, holding onto a pole with your left hand, holding your phone with your right. You are looking at that same pathway lesson screen on your mobile app. And you try to tap a microscopic little pencil icon to edit your notes.

Host 2

Sounds dangerous already.

Host 1

Oh, totally. Because right as your thumb comes down, the train jolts over a rough patch of track. Your thumb just misses the pencil and catches the trash can icon instead. You hit the delete button right on that finger down motion.

Host 2

A classic catastrophic slip.

Host 1

I mean, it's the physical environment completely interfering with the digital interface. And before your conscious brain can even fully register what just happened, like before you even realize you hit the wrong button, scenario three hits you. The system realizes you tapped delete, so a confirmation pop-up instantly appears on your screen asking, are you sure you want to delete this?

Host 2

And let me guess.

Host 1

Yep. Your thumb, acting completely out of pure muscle memory, reflexively taps okay. The pop-up clears, your file's just gone. Three everyday motor failures cascading into a total data loss. And it all happens in about five seconds.

Host 2

It really is the absolute definition of a mechanical user experience disaster. I mean, the interface didn't just fail to support you, it actively conspired with your physical environment and your own muscle memory to destroy your work.

Host 1

Exactly. And that brings us to where we are today. Last time in our first deep dive of season 11, episode S11E01, we established our foundation. We looked at how to structure clear choices for you through information sent and smart defaults.

Host 2

Right. Getting away from worshipping those arbitrary click or item limits.

Host 1

Yes, exactly. Focusing on choice architecture. But today, we are shifting our focus to the exact moment after the decision is made. We are examining what happens in the half second after you decide what you want to do.

Host 2

When the cognitive choice is made and the physical reality sets in.

Host 1

Right. What actually happens when a hand, a thumb, a switch device, or a keyboard actually tries to hit the control.

Host 2

And that is the entire mission of this deep dive. We are going to unpack the empirical physics of human movement. We're going to debunk some incredibly stubborn, I mean really pervasive industry myths about touch targets and mouse interactions.

Host 1

Oh, there are so many myths.

Host 2

There really are. And then we're going to apply these predictive laws to step into the repair shop and fix the broken pathway interface across all those different modalities. Because a brilliant decision or, you know, a perfectly structured choice architecture simply doesn't matter if the user physically cannot execute the action.

Host 1

Okay, let's unpack this. And I want to start with the absolute granddaddy of human-computer interaction physics. In our season 11 prologue, we talked a lot about the difference between heuristics, which are essentially just rules of thumb, and actual predictive models. You are telling me there is an actual genuine, mathematically predictive law of physics for how human beings move to interact with something.

Host 2

Yes. There is. It is one of the very few genuine predictive motor laws we have in our entire field. It's called Fitz's Law. It was originally formulated by a psychologist named Paul Fitz back in 1954.

Host 1

1954. I mean, they barely had computers then, let alone graphical user interfaces.

Host 2

Exactly. He wasn't studying mice or touch screens. He was actually studying aviation psychology and ergonomics. He was trying to figure out the mathematical relationship between human movement, distance, and the size of physical targets.

Host 1

Right, for like cockpits and machinery.

Host 2

Exactly. But what's crucial for us is how this was later adapted for human-computer interaction by a researcher named Scott McKenzie. McKenzie refined Fitz's original work using something called the Shannon formulation.

Host 1

The Shannon formulation. Okay. That sounds incredibly intense. And I know we promised our listeners absolutely no raw math in this deep dive, like no reading out logarithms or slashes or equations. So how do we translate this Shannon formulation into plain English? Why does it matter to me as a user or a designer?

Host 2

It matters because of where the Shannon formulation actually comes from. It's rooted in information theory. The work of Claude Shannon, who was looking at how data travels across communication channels.

Host 1

Like telegraphs or internet bandwidth.

Host 2

Exactly. McKenzie's brilliant insight was to look at the human motor system, your brain, your nerves, your arm, your hand, as an information channel. It is a biological wire that transmits data. And just like an internet connection, it has a strict finite bandwidth.

Host 1

Wow. Okay. So it can only process a certain amount of physical and visual information per second. Right. So my arm moving the mouse isn't just a physical action. It's literally a data transmission from my brain to my hand to the screen.

Host 2

Precisely. And Fitz's law, via this Shannon formulation, states that when you move your hand or a mouse cursor to a target, that movement isn't one smooth, continuous, identical calculation from start to finish. It actually consists of two very distinct phases of movement.

Host 1

Okay. Two phases. What is the first one?

Host 2

The first phase is a rapid, fast, ballistic fling of the pointer or your arm. You see the target, your brain does this massive instantaneous calculation of the general distance and direction, and you just throw your hand in that general direction.

Host 1

You aren't being precise at all.

Host 2

No, not at all. You are just trying to cover the distance as quickly as possible.

Host 1

So the ballistic phase is just pure speed. But I imagine it's pretty inaccurate, right? So what happens next?

Host 2

That is where the second phase takes over. This is the visual feedback homing phase. As your cursor or your finger gets close to the target area, your brain completely shifts gears. It stops relying on that initial gross motor calculation and starts relying heavily on active visual processing.

Host 1

So it's constantly checking its work.

Host 2

Exactly. Your eye looks at the cursor, looks at the target, looks at the gap between them, and literally slams on the biological brakes so you don't overshoot. You enter this intense micro adjustment loop to perfectly land on the target.

Host 1

I am literally moving my mouse around my screen right now while you say this, and I can totally feel it. It sounds like well, think about driving a car. If I need to go 50 miles to another city, how do I drive? I get on the highway, I put it in the highest gear, maybe even set the cruise control, and I just fly. It's fast, it's low effort, I'm just eating up miles.

Host 2

Right. That is your ballistic fling. You are covering vast distances with minimal cognitive overhead. But what happens when you finally reach that city, you get to your destination, and you realize the only parking spot left is a tiny cramped space between two massive trucks, and you have to parallel park.

Host 1

Oh man. I slam on the brakes, I I turn the radio down because I somehow need less audio distraction to see better. I creep forward inch by inch, constantly checking my rear view mirror, checking my side mirrors, adjusting the wheel, terrified I'm gonna scrape the bumper of the truck next to me. The speed completely drops to zero, and the cognitive effort just goes through the roof.

Host 2

That is the perfect analogy.

Host 1

So the homing phase is the parallel parking.

Host 2

That is exactly it. And if we look at the math behind the Shannon formulation, again without reading the actual equation, what it tells us is that this relationship operates on a logarithmic curve.

Host 1

A logarithmic curve. So in practical terms, what does that actually mean for the user sitting at their desk?

Host 2

It means that distance and size do not affect your movement equally. Because the relationship is logarithmic, doubling the physical distance to a target does not double the time it takes you to click it.

Host 1

Wait, really? That feels really counterintuitive. If a submit button is all the way on the other side of my massive ultra-wide monitor, like twice as far away as it was before, it doesn't take me twice as long to reach.

Host 2

Not at all. Because of that initial ballistic fling, your highway driving phase. If the target is twice as far away, your brain just subconsciously tells your arm to fling a little harder, a little faster. Because the logarithmic curve flattens out as it grows, doubling that distance only adds a tiny incremental, almost imperceptible step of difficulty in time.

Host 1

So distance really isn't the enemy here. Distance has a surprisingly low penalty on human movement.

Host 2

Exactly.

Host 1

Okay, so if distance is the highway, and we just drive faster on a longer highway, what about the parallel parking? What about the size of the target?

Host 2

Size is everything.

Host 1

Size.

Host 2

Heavily disproportionately dictates the duration of that slow, cautious homing phase. If a target is tiny, like a microscopic 14 pixel pencil icon, the parallel parking phase takes forever.

Host 1

Because your brain is freaking out.

Host 2

Yes. Your brain has to constantly course correct, processing visual feedback loop after visual feedback loop. It's slowing your hand down to an agonizing crawl to ensure you don't miss.

Host 1

That makes total sense. If the parking space is three car lengths long, I can just swoop right in without even thinking. If it's exactly the length of my car with one inch to spare, I'm sweating and inching back and forth for five minutes.

Host 2

Exactly. So the core takeaway from Fitz's Law is that to make a user interface fast and low effort, you don't necessarily need to cram everything close together. You just need to make the targets big enough that the homing phase is effortless.

Host 1

Okay, but hang on. I'm looking at my notes here and I'm thinking back to our cold open. Our desktop learner wanted to change the playback speed. The gear icon was maybe two inches away from the one point five times speed option. It was a tiny distance.

Host 2

Right.

Host 1

So if Fitz's law says traveling distances across the screen isn't actually that painful, and we just naturally fling our mouse. Why did our desktop learner fail so miserably? They just moved a couple of inches. Why did the menu snap shut on them?

Host 2

That is the perfect question, because it reveals the limitation of Fitz's Law. Fitz's law only applies when you are traveling to a single unconstrained endpoint. But in your scenario, with that hover menu, the user wasn't just trying to hit a final destination. They were navigating a constrained path to get there.

Host 1

A constrained path.

Host 2

Right. And that brings us to our second major physical principle. ACAT and Jai's steering law.

Host 1

The steering law.

Host 2

Yeah.

Host 1

So steering like staying in a lane?

Host 2

Precisely. Fitz's law applies to throwing a dart at a dartboard. The steering law applies when you force a user to navigate a cursor through a narrow corridor. Think about that hover menu for the playback speed. When you hover over the gear, a little menu pops up above it. To reach the 1.5 times option, you have to move your cursor up, but you have to keep your cursor perfectly inside the boundaries of that menu.

Host 1

Right. If you stray one pixel outside the box, the system assumes you are no longer interested, and the whole thing just vanishes.

Host 2

Exactly. Or think of a razor thin video scrubber, where if my mouse drifts a millimeter up or down while I'm dragging it, it drops the video playhead and I lose my spot. It's like trying to trace a wire without letting your hand shake.

Host 1

Oh man, that is so stressful.

Host 2

It is. And mathematically, the ACATJ steering law reveals something terrifying for designers. We just established that Fitz's Law operates on a forgiving logarithmic curve, where distance doesn't add much penalty. But the steering law dictates that forcing a cursor through a narrow corridor multiplies difficulty linearly.

Host 1

Linearly. So it says straight line up on the graph, a direct multiplier?

Host 2

Yes. The longer and narrower the tunnel you force the user to navigate, the exponentially higher the error rate and the time it takes. You are effectively stripping away their ability to use the ballistic fling.

Host 1

You are taking away the highway.

Host 2

You're taking it entirely away. You are forcing the user to stay in the homing phase, the terrifying parallel parking phase for the entire journey.

Host 1

So instead of driving down a wide open highway and then parallel parking at the end, the steering law means you are parallel parking for 50 miles straight.

Host 2

Yes. Imagine driving 50 miles down a highway where the concrete barriers are exactly one inch wider than your side mirrors. You couldn't drive 60 miles an hour. You'd be creeping along at two miles an hour, dripping with sweat, entirely consumed by the cognitive effort of not crashing.

Host 1

That sounds absolutely exhausting.

Host 2

It is exhausting.

Host 1

I would hate that drive.

Host 2

It is exhausting.

Host 1

But wait, there are sites that have massive menus that don't do this. I I'm thinking of Amazon. Amazon has that famous mega drop-down menu on their left sidebar. You hover over a category like electronics, and the massive sub-menu opens up to the right.

Host 2

Right, the mega menu.

Host 1

But I can whip my mouse diagonally across all the other menu categories below electronics to get to the sub-menu, and the menu doesn't accidentally switch to home and garden. It just waits for me. It feels incredibly stable. How are they defeating the steering law?

Host 2

That is the Amazon Cure. It's this legendary piece of interaction design that was brilliantly analyzed and broken down by a developer named Ben Cammons. To understand why it's so good, you have to understand how a normal, bad hover menu works.

Host 1

Usually designers try to fix the steering law penalty by adding a simple hover delay.

Host 2

Meaning the system just waits a fraction of a second before closing the menu.

Host 1

Right.

Host 2

The code basically says if the cursor leaves the menu, wait 300 milliseconds. If it doesn't come back, then close it. It's an attempt to guess the user's intent. But it feels sluggish, and if you move too slowly, it still vanishes. Amazon didn't use a timer. Amazon used directional tracking.

Host 1

Directional tracking. Okay. How does that actually work under the hood without using raw math?

Host 2

Imagine your cursor is currently resting on the word electronics on the left side of the screen. The sub-menu is opened up on the right. Amazon's code draws an invisible triangle, a cone, originating from the exact tip of your cursor.

Host 1

Okay, a cone.

Host 2

The cone expands outward, and the two far corners of the cone are anchored to the top right corner and the bottom right corner of the open sub-menu.

Host 1

Oh wow. So it's like a flashlight beam shining from my cursor, illuminating the entire sub-menu panel.

Host 2

That is a perfect visualization.

Host 1

Now here is the genius part. As long as your mouse is moving anywhere within that invisible flashlight beam, even if you are technically dragging your cursor directly over other menu items like home and garden or books. The system mathematically calculates that your trajectory is aimed at the submenu.

Host 2

So it ignores the fact that I'm hovering over other things because my trajectory proves my intent.

Host 1

Exactly.

Host 2

It assumes you are steering toward that sub-menu and it keeps it open. By doing this, Amazon mathematically cured the linear tunnel penalty of the steering law. They took that narrow, restrictive corridor where you had to move perfectly horizontally, and they dramatically widened the safe path into a massive diagonal cone.

Host 1

So by utilizing directional tracking, Amazon isn't changing the user's motor behavior at all. They are just mathematically altering the boundaries of the safe zone to match natural human movement. They accommodated the ballistic fling. They let me drive on the highway again.

Host 2

Precisely. It is a brilliant mathematical solution to a physical problem. But you know, if we step back and connect this to the bigger picture of UX, there is a much, much simpler cure for the steering law penalty. You don't need complex trigonometry or directional tracking cones.

Host 1

What's the simpler cure?

Host 2

Just use click menus instead of hover menus.

Host 1

Oh. Just make them click it.

Host 2

Yes. Require an explicit click on the gear icon to open the playback speed menu. Once it's clicked, the menu stays open permanently until they click something else. Suddenly the corridor doesn't exist anymore. The user can fling their cursor anywhere they want on the screen, take their hand off the mouse, take a sip of coffee, and the menu is still there waiting for them.

Host 1

That is hilarious, but so true. Sometimes the best way to solve a deeply complex physics problem is just to remove the physics problem entirely. Don't build a better tunnel, just blow up the mountain. Okay, so we've established the baseline mechanics here. We have Fitz's law governing how we hit targets, and the Ecodsai steering law governing how we navigate paths. But there is so much bad advice out there about how to actually apply these laws. I want to spend some serious time busting some industry myths.

Host 2

I am fully ready for this. Let's dismantle them.

Host 1

Let's start with myth number one. I call this the infinite edge on smartphones. I remember reading years ago about this classic desktop design principle, popularized by Bruce Tognazini, who was one of the early interface designers at Apple. Tog's rule was that if you put an interactive element like a start button or an application menu perfectly on the absolute edge or the absolute corner of a computer monitor, it essentially becomes infinitely taller, infinitely wide.

Host 2

Yes, the infinite edge principle. It is a cornerstone of desktop UI design.

Host 1

Right. And the logic makes total sense to me because no matter how fast I fling my physical mouse across my mouse pad, the cursor physically stops the exact moment it hits the boundary of the screen. I cannot physically overshoot the target. My cursor just gets trapped in the corner.

Host 2

It acts like a backstop.

Host 1

Exactly. Therefore, because the homing phase requires zero precision, a target in the corner is mathematically the easiest, fastest target in the world to hit. It's a brilliant hack. I can just throw my mouse blindly. So naturally, designers take this principle and apply it to mobile phones. They stick hamburger menus in the extreme top left corner and floating action buttons in the extreme bottom right corner. But you are saying this is a myth.

Host 2

On a traditional desktop computer with a physical mouse and a physical monitor bezel, Tognazini is absolutely correct. The corners and edges of a desktop monitor are prime infinite real estate. But the myth is that this physical law translates to touch screens. Applying that exact same logic to a smartphone completely backfires.

Host 1

Why does it backfire? A smartphone screen still has a physical edge, right? My finger still stops at the edge of the class.

Host 2

Because the physical mechanics of holding a touchscreen device are entirely different from pushing a plastic mouse across a horizontal desk. On a smartphone, targeting the extreme edges is incredibly hostile to the user. First, let's look at it from a software ecosystem perspective. Modern mobile operating systems have literally claimed the edges for themselves.

Host 1

Oh, right. The gestures.

Host 2

Exactly. Okay. If you try to swipe near the left or right edge, you trigger a system level backswipe, dragging you out of the app entirely. If you reach too high, you pull down the notification center. If you tap too low, you activate the home bar or the app switcher.

Host 1

Okay. That's true. The edges are overloaded with gestures, but what about the physical hardware?

Host 2

From a purely physical hardware perspective, the user's finger often cannot reach the absolute edge of the pixels because it bumps into the thick raised ridge of their protective phone case.

Host 1

Oh, yes. I can't tell you how many times I have been trying to read an article on my phone, and a full-page pop-up ad appears. The ad puts a microscopic X to close it in the absolute top right pixel of the screen. I tried to reach up and tap it, but my thumb just grazes the rubber lip of my phone case and the tap doesn't register.

Host 2

It happens constantly.

Host 1

Or worse, my thumb flattens out against the case, and the meaty part of my thumb registers a tap just below the X, and suddenly I'm being redirected to buy a pair of shoes I don't want.

Host 2

Exactly. The infinite edge becomes an infinite frustration. And we have massive, undeniable empirical data to back this up. A researcher named Stephen Hoover conducted one of the most extensive field observations of mobile usage ever done. He literally went out into the real world on streets, in coffee shops, on subway trains, and observed over 1,300 people interacting with their mobile phones.

Host 1

1,300 people out in the wild, not just sitting in a sterile usability lab. That is huge. What did he find?

Host 2

He found that nearly half of all people hold their phones one-handed. And when you hold a phone one-handed, you do not interact with the screen using a precise floating pointer. You interact with it by sweeping your thumb in a bottom center arc.

Host 1

A thumb arc. Okay. So visualize this. The base of the thumb is anchored roughly at the bottom corner of the phone where it rests in your palm, and the tip of the thumb swings back and forth. So the thumb acts like a windshield wiper anchored at the bottom corner of the glass.

Host 2

Yes. That is the biomechanical reality of mobile usage. The comfortable safe zone for a user is a sweeping arc in the lower middle half of the screen. And because of the physical anatomy of the human hand, the extreme corners of the screen, especially the top corner on the opposite side of the hand holding the phone, are the hardest, most physically straining, and most error-prone areas to reach.

Host 1

That completely flips the desktop logic on its head.

Host 2

It does. Huber's data absolutely destroys the infinite edge myth for mobile. The edges aren't infinite targets, they are physical and software minefields.

Host 1

Okay, so we just established that the physical edges of a mobile screen are minefields for the thumb. But that entirely assumes the user is relying on touch or a physical mouse. This is a perfect transition to myth number two. I'm going to plead devil's advocate here and push back on this whole framework.

Host 2

Bring it on.

Host 1

If Fitz's law is strictly about physical distance and physical size, flinging arms and sweeping thumbs. Does this whole physics framework just completely collapse the second a user drops the mouse and switches to an assistive technology? Does this stuff even matter if a user has a motor impairment and navigates a website using sequential keyboard tapping or a sip and puff switch device, or even voice control? Like, if there is no physical cursor moving across physical pixels, do the laws of movement just disappear?

Host 2

That is a brilliant question and a very, very common misconception. The myth is that Fitz's law completely stops applying to alternative modalities like keyboard, switches, and voice. But the answer is that the physics still entirely govern the interaction. And this brings us back to what we discussed at the very beginning regarding the Shannon formulation. Fitz's law is not actually about mice or fingers or glass screens, it is about the information capacity of the human motor system.

Host 1

Information capacity. You mentioned that earlier, the idea of cognitive and physical bandwidth. How does that actually translate to a keyboard?

Host 2

Let's redefine the variables of the equation. The underlying physics of distance to target versus size of target remains a universal predictive model, regardless of the modality. For a sequential keyboard user, someone relying on the tab key to move through an interface, the distance is no longer measured in physical screen pixels or inches. It is measured in the number of tab stops.

Host 1

The number of keystrokes required to get from point A to point B.

Host 2

Exactly. If you put your main content at the bottom of the page and you force a keyboard user to press the tab key 50 times to navigate through a massive, unskippable drop-down menu just to reach the article they want to read, well, that is an enormous exhausting distance. You have forced them to travel a massive cognitive and physical distance.

Host 1

Okay. So if the distance is the number of tab keystrokes, what is the size of the target for a keyboard user? They aren't trying to land a cursor inside a box.

Host 2

For a keyboard user, the size of the target is the visual clarity and the predictability of the focus state. When you press tab, a little outline usually appears around the active element. If a designer has removed that focus outline because they think it's ugly, or if the focus state jumps unpredictably around the page, the user loses their place.

Host 1

That makes total sense.

Host 2

If they don't know where they are, the target is effectively microscopic. The cognitive homing phase slows to a crawl because they have to scan the entire screen after every single keystroke just to find their cursor.

Host 1

That is fascinating. What about a switch user? Someone using a sip and puff device or a single button switch mounted on their wheelchair. How do they experience distance?

Host 2

Switch devices often use a method called row column scanning. The interface highlights the top row of the screen. If the user does nothing, it highlights the next row down. When the row they want is highlighted, they trigger the switch. Then it starts highlighting individual columns within that row, one by one. They wait and trigger it again when it reaches the specific button they want.

Host 1

So for a switch user, distance is literally just time. It's the agonizing way for the scanner to iterate through every single preceding element on the screen until it finally reaches the one they want.

Host 2

Exactly. Time is distance. And even for voice control, there is cognitive target acquisition. Imagine a user navigating a complex dashboard by speaking commands. They have to visually scan the interface, locate the target they want, read its label, and then speak that label aloud to the system.

Host 1

So if a button has a tiny low contrast label or an ambiguous icon with no text label at all, they don't even know what to say.

Host 2

Right. If the label is visually ambiguous, the homing phase of their cognitive movement slows down. They have to guess the label or ask the system to display a grid overlay. The core physics, how much effort it takes to cross the distance and how much clarity is provided to acquire the target, still heavily govern the interaction. Fitz's law is truly universal.

Host 1

That completely flips how I think about layout. Distance isn't just physical space, it's effort. It's time, it's keystrokes. Okay, myth number three. And this one is crucial because I see brilliant designers do this all the time. Let's talk about the invisible tap box loophole.

Host 2

Oh, this is a pervasive one.

Host 1

I'm gonna defend this practice for a moment. From a visual design perspective, minimal interfaces are the current standard. Users like clean, uncluttered aesthetics. Let's say I am designing a mobile app and I have a microscopic minimalist icon on my interface. It's just a tiny little three-dot menu icon, maybe 14 pixels wide. But I know my accessibility guidelines, I know I have to pass compliance. So in the code, I wrap that tiny 14 pixel icon in an invisible transparent bounding box that is 44 pixels wide.

Host 2

The invisible safety net.

Host 1

Right. The user touches anywhere in that massive invisible box and the button clicks. I get to keep my beautiful minimalist aesthetic. And I pass the mechanical accessibility checks. That is a win win, right? Why is that a myth?

Host 2

Because you are designing for a compliance script, not a human being. You might pass an automated accessibility checker looking for a mechanical 24 pixel double A bounding box or a 44 pixel triple A bounding box according to the web content accessibility guidelines. The script scans your code, sees the 44 pixel bounding box, and gives you a green check mark, but you have completely failed the human motor system.

Host 1

But the physical target is big enough. The phone will register the tap. If the invisible padding allows the thumb to physically hit the target, why does the visual size actually matter?

Host 2

Because of the homing phase we talked about earlier. Remember, Fitz's law dictates that the second phase of movement, the parallel parking phase, is visually guided. The brain constantly adjusts the speed of the hand based on what the eye sees, but the user's eye cannot see your invisible HTML padding.

Host 1

Oh. They don't know the padding is there. They don't know they have a 44 pixel safety net.

Host 2

Exactly. The user's brain looks at the screen, sees a tiny, cramped 14-pixel target, and it basically panics. It assumes the target is microscopic. So it still drastically slows down their visual homing phase, cautiously and painstakingly trying to pinpoint that microscopic 14-pixel icon. They will hover their finger, hesitate, and tense up their muscles to aim perfectly at the pixels they can actually see.

Host 1

Wow. So all that invisible padding does nothing for their cognitive load.

Host 2

You haven't saved them any effort at all. You force them to parallel park in a space they think is tiny, even if it's technically three car lengths long.

Host 1

That is a massive disconnect between the code and the cognitive experience. The code is generous, but the visual cues are threatening.

Host 2

And it gets worse. Let's say you have a toolbar and you pack a bunch of these tiny 14-pixel icons tightly together. Because you need them to pass compliance, you wrap all of them in 44 pixel invisible padding. Now those invisible hit areas are crammed right up against each other, edge to edge.

Host 1

Oh no. I see where this is going.

Host 2

You have created a massive risk for neighbor slips, what we commonly call fat finger errors. Because the visual targets are small and visually crowded, the brain cannot confidently separate them. The user tries to hit one, their thumb lands slightly off center, still within the invisible padding of the wrong button, and they accidentally trigger an action they didn't want.

Host 1

So visual size matters just as much as physical tap size. You cannot fake Fitz's law with invisible boxes. The eye and the hand are a single system. If it looks hard to hit, it is hard to hit, regardless of what the CSS says. The myth that habitual confirmation pop-ups save users from making terrible mistakes. The idea that if an action is dangerous, we should just ask the user, are you sure?

Host 2

This is one of the most pervasive, historically entrenched, and frankly damaging myths in interface design. It assumes that more friction always equals more safety. And this myth was thoroughly dismantled by a brilliant interface designer named Oza Raskin.

Host 1

Let's look at the commuter. So the system did exactly what it was programmed to do. Right? It threw up a warning asking, Are you sure? But the commuter hit okay anyway. Why do we do that? Why did their brain just completely ignore the warning?

Host 2

Because of a powerful psychological force called human habituation. Let's step away from screens for a second. Think about your morning alarm clock. If you set the exact same alarm sound every single day for a year, what eventually happens?

Host 1

I start sleeping right through it. Or I reach over and hit snooze without ever fully waking up or remembering that I did it. My brain just filters it out as background noise.

Host 2

Exactly. When a system constantly interrupts you with repetitive stimuli for routine actions, your brain adapts. It is incredibly efficient. It realizes that reading the pop-up, evaluating the text, and making a conscious decision takes up precious cognitive bandwidth. So to save energy, it offloads the task to muscle memory. It bypasses the prefrontal cortex, the thinking part of your brain, entirely.

Host 1

We see this with cookie consent banners on websites or terms and conditions. I don't read them. I just scan for the big green accept button and click it as fast as humanly possible so I can get to the article I actually wanted to read.

Host 2

Right. We build a physical reflex. Click delete, click OK. Click save, click OK. Click submit, click OK. It is no longer two separate decisions. It is a single thoughtless combo move. The warning becomes cognitively invisible.

Host 1

So Aza Raskin's argument is that because we form these habits, the safety net actually has a giant hole in it.

Host 2

Raskin proved that habitual confirmations completely fail to protect us when we actually make a genuine mistake. Because the action has become a reflex by the time your conscious brain catches up to the visual feedback, by the time you realize you didn't actually want to hit delete, your muscle memory has already dismissed the pop-up. You confirm the mistake before you even realized it was a mistake.

Host 1

Okay, but I want to push back on this. Data loss is a massive issue. If a user accidentally deletes a critical file, they are going to blame the software, not themselves. If we can't warn people, if we can't ask, are you sure? How do we protect their data? We can't just let them blindly delete things.

Host 2

You are right. We have to protect them. But the solution is a historical shift in how we approach error prevention. You don't warn them before the action. You let them undo the action after the fact. The vastly superior approach combines two specific design mechanisms. First, we implement a concept from the web content accessibility guidelines known as pointer up event cancellation.

Host 1

Pointer up event cancellation. Okay, let's translate that out of the WCAG documentation and into reality. What does that actually mean for my finger touching the screen?

Host 2

It means an action should never ever trigger the exact moment your finger touches the glass or the moment the mouse button presses down. That is called a down event. Instead, the action should only trigger when you release the pressure, when the finger lifts up off the glass or the mouse button clicks back up.

Host 1

So when my commuter was on the train and the train jolted and their thumb slammed down onto the delete button, if the system used up event cancellation, the file wouldn't delete at that exact millisecond.

Host 2

Correct. The button might visually depress or change color to show it's being touched, but the destructive action is held in suspense. This gives the user a crucial built-in escape hatch. If you realize your finger is on the wrong button, you don't lift it straight up. You keep your finger pressed against the glass, you slide your finger away from the button, moving it outside of the target's physical boundaries into a safe neutral space on the screen, and then you release. The action is safely canceled.

Host 1

That is so elegant. You catch the mistake mid press. It gives the user agency. It's like realizing you are about to drop a glass and managing to catch it before it hits the floor.

Host 2

It relies on the user's immediate physical realization of a slip. But what if they don't realize they slipped until after they lift their finger? What if they fully execute the click? That is where the second mechanism comes in. You replace the annoying confirmation pop-up with a highly visible, non-blocking, 10-second undo function.

Host 1

Like when I archive an email and Gmail. It just instantly vanishes from my inbox, but a little black toast banner pops up at the bottom of the screen that says conversation archived with a bold undo button next to it.

Host 2

Exactly. As Raskin said, never use a warning when you mean undo. The warning blocks the user and forces them to read. The undo respects their intention, executes the action instantly, but provides a temporary safety net if their intention was a mistake.

Host 1

Because with an undo, I don't have to break my flow. I can confidently and rapidly delete five old draft notes in a row, knowing I don't have to swat away five confirmation pop-ups. But I also know I have a safety net if I mess up on the sixth one. You are giving the user their cognitive bandwidth back.

Host 2

Precisely.

Host 1

Okay, we have covered an immense amount of ground here. We have the physics of Fitz's law and the Ako-Jai steering law. We've busted the infinite edge myth. We've proved that modalities don't erase physics. We've exposed the danger of invisible tap boxes, and we've dismantled the habituation of confirmation pop-ups. Now, I want to invite you, the listener, to step into the repair shop with us. Let's take the broken pathway video lesson interface we experienced in the cold open, and let's actually fix it. Let's apply everything we just learned across every interaction modality to build something that actually works.

Host 2

Let's do it.

Host 1

It's time to translate the physics into design. Where should we start?

Host 2

Let's start with the very first failure, the speed submenu. The one where you try to move your mouse to the 1.5 times speed option, and because of the steering law, it snaps shut on you mid-flight.

Host 1

Right, the classic steering law trap.

Host 2

To fix this, we are going to completely abandon the concept of hover intent. We are going to convert that fragile hover tunnel into a stable, explicit click popover.

Host 1

So no more hovering over the gear icon to reveal the menu. The user has to physically click the gear.

Host 2

Correct. The user clicks the settings gear. The exact moment they click, a large static menu pops open. And more importantly, it stays open. Because it is triggered by an explicit click state rather than a transient hover state, the corridor penalty vanishes completely. The user can freely move their pointer from the gear icon to the 1.5 times speed option without navigating a tight rope.

Host 1

Because the menu isn't good at vanish if they accidentally drift outside the box.

Host 2

Exactly. They can take a direct, sloppy diagonal, ballistic path to the target without fear. We have removed the linear penalty entirely. We let them drive on the highway.

Host 1

Okay. Next up, let's fix the video scrubber. On the pathway platform right now, the timeline scrubber at the bottom of the video is this gorgeous, minimalistic three pixel tall wire. It looks incredibly sleek. And the play pause and volume icons are cramped little 18 pixel dots right next to each other.

Host 2

And it is a classic textbook violation of Fitz's law and WCAG target sizing. From a physics perspective, a three-pixel vertical target forces an agonizingly slow, highly stressful homing phase. The user has to pinch their mouse or carefully aim their thumb to grab that tiny wire without accidentally tapping the background and pausing the video.

Host 1

So how do we fix it without making the interface look like a clunky child's toy?

Host 2

We are gonna radically widen that scrubber. We transform it from a three-pixel wire into a thick, visually distinct touch track, perhaps 24 or 32 pixels tall. And those cramped 18 pixel icons, we are expanding them into visually distinct, well-spaced 44 pixel targets.

Host 1

Visually distinct. Meaning we aren't just taking the tiny dots and wrapping them in invisible 44 pixel padding to cheat the compliance checker.

Host 2

Exactly. We make the actual visible footprint of the button large enough that the user's brain intuitively trusts it can hit it without slowing down. But we can take this a step further, specifically for motor accessibility.

Host 1

How so? What's wrong with just a big scrubber?

Host 2

Scrubbing a video, meaning grabbing a tiny playhead dot and dragging it horizontally left or right to find a specific timestamp, is an incredibly demanding physical task. It requires sustained pressure to hold the click or the touch, combined with precise, sustained steering across a horizontal axis. For a user with a motor impairment like Parkinson's or a severe tremor, sustained dragging is nearly impossible.

Host 1

Because if the cursor shakes, they might accidentally release the click mid-drag and they lose their spot entirely.

Host 2

Exactly. So alongside our newly widened scrubber, we are adding dedicated jump forward 10 seconds and jump back 10 seconds buttons right next to the play button.

Host 1

Oh, that's brilliant.

Host 2

So instead of forcing a user to drag a playhead to rewind, they can just tap a massive stationary button three times to jump back 30 seconds.

Host 1

Yes.

Host 2

We completely bypass the need for fine motor dragging by providing an alternative discrete interaction model. We replace a complex sustained steering task with a simple ballistic targeting task.

Host 1

Okay, let's move down the screen. Below the video player on pathway is the playlist sorter. It's a list of all the upcoming video modules. Right now, the only way to reorder those lesson modules is by grabbing a card with your mouse and dragging it up or down the list. Visually, it's very slick, great with a mouse, but a total nightmare if you're using a keyboard or a switch device.

Host 2

It is entirely inaccessible. We just talked about how hard dragging is for motor impairments. Drag and drop requires a complex sequence of holding, moving, and releasing that many assistive technologies simply cannot replicate easily. A switch user using row column scanning literally cannot perform a drag and drop gesture. Our repair here is going to provide single pointer and keyboard parity.

Host 1

So what do we add? How do we replace drag and drop?

Host 2

We don't necessarily remove drag and drop for the mouse users who like it, but we add explicit move up and move down chevron buttons directly beside every single playlist card.

Host 1

So there's a dedicated button for every direction for every item.

Host 2

Yes. Now a keyboard user doesn't have to try and emulate a mouse drag. They can simply press tab to navigate to the move up button and hit enter. A switch user can scan to it and select it. A mobile commuter on a bumpy train can just tap it with their thumb.

Host 1

And because we are adding all these dedicated buttons to the interface, we have to make sure they play nice with the rest of the layout, right? We can't just cram them in and break the design.

Host 2

Absolutely. We have to ensure that when a visually impaired user zooms their interface to 400% magnification, the layout reflows cleanly. The move up and move down buttons must stack logically, either beside or below the text without overlapping the content or shrinking below that critical 44 pixel target sizing. We don't just add targets, we ensure they survive extreme responsive reflowing.

Host 1

Okay, final fix. The deletion flow. Let's go back to our commuter on the train. They tap the delete button by mistake on a finger down motion when the train jolted, and then they reflexively confirmed it when the pop-up appeared. How are we rebuilding this?

Host 2

We are ripping that entire flow out by the roots. First, based on WCAG 2.5.2 pointer cancellation, we swap the immediate finger down deletion for finger up activation.

Host 1

So if the user touches the delete button by mistake, it's the train jolts, they realize their thumb is on the wrong icon. They hold their finger on the glass, they drag it a few pixels away from the button into the white space and release, no action is taken.

Host 2

The ultimate escape hatch. They have saved themselves. But let's say they don't catch it. They lift their finger directly on the delete button, the action fires.

Host 1

Do we show them the confirmation dialogue?

Host 2

We do not show them a nagging confirmation dialogue. The item instantly disappears from the list, the UI shifts up to fill the gap, and a polite banner appears at the bottom of the screen. The banner says item deleted, and it has a massive, highly visible 10 second undo button right next to it. And crucially, we ensure that banner is announced immediately by screen readers, so visually impaired users know exactly what just happened and how to reverse it.

Host 1

Okay, I have to push back on this one last time. I understand the logic, but what does this all mean for the extremes? Do we just abolish all friction forever? Do we never ever ask the user to confirm anything? What if they are deleting something massive?

Host 2

That is the vital nuance here. This raises an important question about exceptions. The undo model is vastly superior for routine, reversible actions, deleting a note, archiving an email, removing a song from a playlist. But there is an exception for truly irreversible, highly destructive high-stakes actions.

Host 1

Like what? What is the threshold for an exception?

Host 2

Imagine a user is in their pathway account settings, and they are about to wipe their entire multi-year certification record from the database. All their degrees, all their progress, gone. Something that legally requires data purging and cannot be recovered from a server backup. In that specific, highly destructive scenario, deliberate friction actually belongs there. But we still do not use a simple pop-up with an OK button.

Host 1

Right, because muscle memory will just reflexively click OK without reading the pop-up.

Host 2

Exactly. For truly destructive nuclear actions, you want to violently break cognitive habituation. You must force the user out of their reflex loop and back into their prefrontal cortex. You do this by asking them to manually type a specific phrase, like the word delete in all caps or the exact name of the project they are destroying into a text box before the final red button even becomes clickable.

Host 1

Oh, love that. Routine actions get the invisible safety net of undo. Nuclear actions get the massive concrete speed bump of manual typing. That mechanical friction forces the conscious brain to re-engage, read the prompt, and physically spell out the destruction they are about to cause. It's incredibly respectful of the user's time and attention.

Host 2

It is all about managing their cognitive and motor bandwidth appropriately. Friction is a tool. Use it sparingly and only when the consequences demand it.

Host 1

This has been an absolutely fascinating deep dive. We have taken abstract mathematical laws of physics and turn them into highly tangible, empathetic design decisions. Which brings us to our diagnostic exercise. We don't want you to just listen to this. We want you to test this out in the real world. I am giving you a 10-minute homework challenge.

Host 2

A chance to apply the physics to your own environment.

Host 1

Exactly. I want you to open your web browser, go to the pathway platform, or better yet, open up your own company's app or website, open up the developer tools, and spend 10 minutes doing three specific tasks.

Host 2

Task one. Turn on the layout overlays or inspect the elements in your developer tools. I want you to compare the visible size of your interactive icons against the real mechanical HTML tap boxes. Are you faking Fitz's law? Are you relying on invisible padding that the compliance checker sees, but the user's brain cannot? If your visual target is smaller than 44 pixels, you have a problem.

Host 1

Task two. Or did it prematurely fire the action the exact millisecond you made contact?

Host 2

And task three. Unplug your physical mouse entirely. Try to reorder a list or navigate your primary dashboard purely with keyboard using the tab and enter keys. No drag and drop mouse gestures allowed. Can you accomplish the core tasks of your application using sequential targeting? How many keystrokes does it take? Is the focus state visible?

Host 1

Take ten minutes today and see exactly how your interface holds up against the unforgiving physics of human movement. And that perfectly tees us up for our next deep dive in this series.

Host 2

Indeed. We have now covered choice architecture, and today we covered physical execution. Now that we know what happens in the half second a user actually tries to hit one single target, we need to ask a broader question. What happens in the milliseconds before they start moving? What happens when their brain opens a page and is just trying to make sense of the overall layout in the first place?

Host 1

Next time in episode S11E3, we are exploring how people perceive grouping, simplicity, and distinction. We are going to zoom out from the physics of hitting a single target to explore how the human eye and brain organize an entire screen of information in mere milliseconds, using Gestalt principles, visual hierarchy, and contrast.

Host 2

But before we go, I want to leave you with a final thought to mull over. If human motor output, your hands, your eyes, your nervous system, is an information channel with a strict, finite bandwidth, then every single unnecessary click, every microscopic target, every narrow, frustrating hover menu you would design is literally stealing your user's finite processing power. You are draining their battery before they even get to the content. What could your users achieve if you redesigned your interfaces to give them that cognitive bandwidth back?

Host 1

Think about that next time you are riding the train trying to tap a tiny icon with one hand as the car jolts forward. We'll see you next time.

Save to playlist

Manage playlists →

Share