All of these interfaces (Sonic Transducer, Medusa, and de-Medusa) are similar in interaction and instructive to compare, so I’ll be reviewing them together in this post.
After the terrifying dinner party and subsequent chase around the castle, Janet, Brad, and Dr. Scott wind up on a raised platform facing Frank-N-Furter in the laboratory, and, fed up with their nonsense, he activates the Sonic Transducer.
Sonic Transducer
To do this, he pulls a large double-pole knife switch mounted to the panel between the Reactor Power Input and Monitor from an up to a down position.
This has the effect of “freezing in place” the people on the platform. One must presume that the effect lasts as long as the switch is down.
That’s it. That’s the interaction.
The Transducer will seduce ya
Dr. Scott guesses on hearing the name that, “It is, I suppose, some kind of audio-vibratory physio-molecular transport device?” Brad grips his shoulder and says, “You mean…!” Dr. Scott confirms it, “A device which is capable of breaking down solid matter and then projecting it through space and—who knows—maybe even time itself.”
Narrator: It wasn’t.
I mean, in the show it is, but I think that term “sonic transducer” was included in the script because it sounded pseudoscientific, mocking the B-movie shows from which this movie draws some inspiration. But that term does mean a real thing in the real world, and I don’t want any engineers reading to be confused.
Map it, Janet
The mapping here is upside-down. As I noted in the prior post, for vertical controls, up naturally maps to on. There is a competing map here, that [switch down == feet down]. (Yes, yes. Vheels, too.) And if this were the only control of its sort in the panel, I might agree.
But given that this is part of a panel of controls, some of which are also knife switches (see below), consistency wins. In a high-pressure situation, like, say, a monster rampaging out of a vivification vat, Transsexuals (as in from the planet Transsexual, not my awesome trans readers, shout out yo 👋🏻🏳️⚧️) shouldn’t have to remember which knife switch is up for on and which one is down for on. That’s far too error prone.
Add gravity threatening to turn it on accidentally, and the case is closed.
Medusa and de-Medusa
Following a bit of dialogue, Magenta, Riff Raff, Rocky, and Columbia all enter the lab through the wheelchair hole. After some more dialogue and song, Frank-N-Furter glances at Magenta, who stands by a control on the panel labeled Medusa. It’s a large-handled double-pole knife switch mounted to the control panel, covered in the same powder coat red as the rest. When we first see it, the handle is up. During Brad’s threat to Frank, Magenta raises her hand to the handle. When Frank nods her way, she throws it to the down position and immediately back to the up position.
At the moment the switch is at the lowest point, a lightning-bolt-shaped light illuminates green on the front of the switch for a moment, we hear a pewp sound, and Brad turns to stone. Dr. Scott, Janet, Columbia, and even Rocky are turned to stone in the same way.
Quick consistency aside: Why is Dr. Scott’s lap blanket turned to stone when the rest of his clothes disappear, and when Brad’s robe and Janet’s undergarments just fall to their feet? Is there a missing set of controls for what to do with the clothing, and how to make modesty exceptions?
Later,—after Frank dresses them, puts makeup on them, and places them on a proscenium stage (Dr. Scott is left backstage for some reason)—a hand reaches to a control which has a chalk label of “de-Medusa” written on the brick wall to which it is mounted. When it is pulled down and pushed back up again, Columbia turns back to her regular flesh-and-blood self, and the floor show musical number begins.
One by one, the switch is flipped and the other characters are de-Medusa-ed as well, right when their number is up. It is a reminder of the gonzo nature of this movie that not only is a diegetic rationale for the whole floor show never given, but we must accept that somehow when they are turned back from stone, they exhibit no disorientation and immediately begin singing and dancing their parts in a diegetic musical number they have never heard of, much less rehearsed, before that moment. Were the lyrics and choreography somehow beamed into their heads? But let’s not ponder these things, as the (floor) show must go on.
The de-Medusa is very similar to the Medusa control and has the same problems: A double-pole knife switch, which does its thing when at the bottom of its extent. It also targets a single person at a time but there’s no interface to indicate which one. The only thing I can see that’s different is that it is painted powder coat black instead of red. Since they are functionally the same, we can talk about them together.
Missing control
How do the Transsexuals target individual people to Medusa? Using today’s tech, you can imagine an audio classifier determining who spoke most recently (or sung most recently? insulted Frank most recently? who Frank was looking at? Any of these could work.) and a visual classifier identifing where that person was. Then the weapon could automatically aim there.
The makers of the film in 1975 didn’t have access to those technologies. But they could have solved it manually. Riff Raff—who is right there, not doing anything—could have aimed a pivoting barrel, like a spotlight, or fiddled with a joystick to aim one of the overhead lights. Something to convey this is how they set the area of effect.
Mapping. Again with the bad mapping.
For very similar reasons to the Sonic Transducer, this is the wrong mapping. Up maps more naturally to on. Gravity threatens to activate it accidentally. And what happens if it did fall and just hangs there at the bottom? This leads us to the next section, where I try to convince you that it’s exactly the wrong control.
A knife switch is the wrong control
It is an axiom of interaction design that the control should map as well as possible to its effect. To illustrate this, consider two common household controls.
Doorbell buzzer: Push this and it rings a bell inside the house for as long as you keep it pushed. The moment you let up, the ringing stops. This is called a momentary switch.
Light switch: Flip this up and see the light it controls illuminate. It’ll stay that way until the switch is flipped down again. Continuing to press up does nothing. This is called a toggle switch.
Now consider what happens if you switch these two.
Doorbell switch: Flip this up, and the bell rings for as long as it remains up. The opportunity for abuse is nightmare fuel.
Light buzzer: Press this and the light comes on, but only as long as you stand there and hold the thing. The lamp is dim because you refuse to stand there and press it constantly. Because you only think of yourself.
These examples serve to illustrate the point: Some controls fit their effect, others don’t.
Where it works
In the case of the Sonic Transducer, Medusa, and de-Medusa, the control is a toggle switch like the light switch, which fits persistent states. That’s exactly right for the Sonic Transducer effect. It is a state that the Transsexuals want to control, and have it be persistent as long as that control stays on.
Where it don’t
But in the case of the Medusa and de-Medusa, it is being used to say “turn them to stone now”. The event is instant, but the control is not. I mean, sure, that point at the bottom can be considered a momentary switch the way Magenta uses it, but it’s not germanely so. And she doesn’t have to treat it like a momentary switch. She could just leave it at the bottom, and it’s unclear exactly what would happen if she did. So it’s just wrong.
Clever readers will note that the stone/flesh states are persistent and reversible, just like light on/off. So doesn’t that befit a toggle? It would, if there were one thing to toggle. But the state lives in the people, not the panel, and there are five of them petrified at different times. A single toggle can’t Medusa the next person without de‑Medusa‑ing the prior, even with some selection mechanism. You can imagine one toggle per victim, somehow “imprinted” on each as they’re turned, but then you have to explain how that association gets made, and that’s its own pain. Targeting plus activation is narratively simplest.
Much more fitting than a toggle switch or a knife switch would be a big button on the panel, which, like the lever on a toaster or the “start” button on a microwave, sets a process in motion. So better would have been: Riff Raff works the joystick to target the spotlight on the victim and at Frank’s nod, Magenta slams the giant button on the panel. If a button wouldn’t deliver the dramatic motion, then a spring-return lever—like on a punch clock or slot machine—would work better than the knife switch. That would preserve the drama without being nonsensical, and preserve the design principle at the heart of this post:
Controls should match their effects as best they can.
When Georgiou escorts Noe into his hotel room, she activates and tosses a device onto the container he’s carrying. It’s a palm-sized, metallic saucer-dome shape with intricate detailing, and sports three rounded signal lights that glow white. It magnetically grips to the surface of Noe’s container. He cautiously asks what it is and she pulls out another one to show him, explaining it is a “phase pod”. She attaches hers to her belt, and a beat later its white lights turn green. Her appearance becomes blurry and shifting and there is an audible low purr. A beat later the lights on the pod attached to the container turn green, and it gains that phasing appearance as it slips from his hand to land with a thud on the floor. He tries to shoot her with a phaser, and the blast passes through her to hit the wall behind. He tries to hit her physically and passes through. When San enters through the wall a little later, we see a similar pod attached to his belt and the phase-fight begins.
There’s a moment where Georgiou pulls her phase pod off to render herself immune to the knife he’s slicing at her while phased. San pulls his off to reengage her in unphased-space. A beat later we see them crash through the glass window separating the room from the nightclub floor.
When the fight takes them both onto a raised dias, Georgiou taps her pod to turn it on again. We see its lights and the lights on the case pod instantly turn green. She’s phased. San turns his on again, too, and the fight continues. About midway through the fight, San stabs her pod. He doesn’t quite disable it, but it starts to malfunction, its status lights flickering between green and white.
San drops the case and kicks it through a wall. Georgiou tries to run through the wall to retrieve it, but the malfunctioning pod fails to phase her left shoe. It “catches” and won’t pull through the wall. The status lights of the pod flicker green and red. In frustration, Georgiou smashes the device a few times to no avail.
Caught, she watches as San appears, opens the case, removes the Godsend, and teleports away. Too late, the pod sorts its shit out, and Georgiou is able to pull her foot through and stand up to be confronted by Sahar.
Caveat: Some things are unexplained, here
It’s not clear how the floor holds a phased thing, while the walls do not. Maybe it’s some aspect of the artificial gravity? It’s also not clear why Baraam doesn’t have the equivalent of a red alert when the fight starts, which would prevent San’s beaming away. Most casinos have outrageous levels of security and Georgiou is paranoid, so one would expect it. These are script questions, admittedly, not related to the interfaces, but things that the skeptic in me must voice.
One production gotcha
Before they reach the dias (below) we can see an unphased Georgiou holding the case. Its signal lights are green, but since she’s holding it, it must be unphased, too, which breaks one of the two diegetric rules already established.
Green means Thing is Phased.
Phased and non-phased things do not interact.
Rather than trying to backworld this (which would get complicated fast) I’m going to presume this is just a production mistake.
Evaluation as a wearable
On this blog I’ve established some guidelines for what makes a good wearable. I’ve used those to evaluate this interface.
Sartorial? Yes.
Yep. Palm-sized, one flat side, other side rounded. Lovely textures and shape. All fit being worn.
Easy to access and use: Tap to activate
I wonder about accidental activation. The fight has a lot of bumping around. If it’s a dumb momentary button, capacitance sensor, or accelerometer, some accident of the fight might accidentally turn it on or off, so let’s presume it’s not that. If it’s a biometric signature requiring the authorized user to tap it, it still seems riskily “out there”, ripe for an accidental touch that could have the wearer slamming into a wall they thought was passible, or dephasing in the middle of something. A long-touch, as seen with the mission briefer, might make more sense. Long-touch trades discoverability (how is anyone supposed to know to long-touch a thing) for increased certainty (accidental long-touches are less likely than accidental brushes.)
A long-touch introduces some challenges mid-fight, but I suspect this is primarily meant as an infiltration tool, not a combat one.
Social and Apposite I/O: The glow is almost fine
You might think that having the glow there gives too much away, diegetically. It would draw the attention of onlookers and raise suspicions if not alarms. Generally speaking, a covert wearable announcing itself brightly runs counter to its purpose. But the visual and audio effects of the underlying tech are far more conspicuous, and I presume, unavoidable. So subtlety is not an option using this tech. And the signal lights help convey to the user the states we see.
Status
Display
Off
Dark
Ready
White
Active
Green
Damaged
Green and white
Error
Green and red
That’s quite useful, even if they require a glance down. There’s an additional status which is “these two pods are paired” which might be accomplished with a synchronized blink, but that might also give too much information away to assailants.
As we see in the film, one problem is when phase-ees want to kick other phase-ees out of phase-space. Then the light on the surface provides a helpful signal of exactly where to target. Let’s trust that the jerky visual effects we see from our non-phased vantage point are still in effect when phasing, so that little target’s going to be jumping around anyway. Also, as long as all pods have the same lights, it’s not granting an advantage to anyone in particular. Georgiou could just as easily targeted San’s pod.
Haptics, probably
The use case of most concern for me is when the pod is malfunctioning. It looks like the tech prioritizes its effects for living matter, so bodies don’t unphase inside of solid matter. That would make for graceful degrading that could inform the user that the tech is trending in an unsafe direction. “Hey, why is my scarf stuck. Oh crap!”
We can’t rely on the phasing side-effects, because those seemed to continue as usual even during the half-functioning. Whatever that signal is shouldn’t rely on her looking at the device, either. A haptic feedback seems most fit, and specifically where it vibrates when the device is working, pulses when it is half-working, and buzzes or goes still when it is not working at all. Since film generally does not convey haptic feedback well, it might actually be part of the device we see. That would explain why Georgiou was willing to risk passing through the wall. The stakes were high, and she could tell, without looking, there was a good chance she would make it.
Between the tap that should be a long-tap, the lights that clearly signal mode, and the haptics that we’re going to presume are there anyway, the phase pod makes for a plausible wearable that meets basic usability and story needs, too.
The Fritzes award honors the best interfaces in a full-length motion picture in the past year. Interfaces play a special role in our movie-going experience, and are a craft all their own that does not otherwise receive focused recognition.
Today we’ll be covering Best Interfaces. The movies nominated for Best Interfaces manage the extraordinary challenge of being believable and helping to paint a picture of the world of the story. They advance the state of the art in telling stories with speculative technology.
The winner of the Best Interfaces award for 2025 is Section 31.
As you’ll read below, my posts on the winner will be a series rather than a single post, so let me do one Also Check Out here.
Bust first, also check out: Superman
Though I have some issues with the amount of fuigetry in most of the screens, and how Lex has to call out countermoves rather than have an assistant offer next most likely countermoves; the robots in the Fortress of Solitude and the crazy-cool gestural control of his spheres by Mr. Terrific make me think that interfaces and tech will not be an afterthought in DC’s new Gunn era.
(James: reach out and I’ll send you a free copy of my book about assistants, it would have helped with that Luthor interface.)
The 2026 Best Interfaces Award goes to Star Trek: Section 31
Maybe I was out of the loop, but I don’t recall hearing a lot of buzz about this movie at the time it came out. But when I finally caught it, I was impressed with the breadth, the art direction, and some interfaces of a sort I don’t think I’ve documented before. This year I’m going to honor the winner with an old-school breakdown, interface-by-interface. In this post we’ll start with a general overview, and then move to the Mission Briefer.
Note I try to only describe just enough so the interfaces can be understood, but since this is a cloak-and-dagger spy thriller, it’s still pretty intricate.
Plot overview
In the Mirror Universe of Star Trek, the mostly-good United Federation of Planets doesn’t exist. Instead it has a morally-inverse counterpart called the Terran Empire. Philippa Georgiou became ruler of this evil empire in part by defeating and enslaving the ambitious contender San. Once ascended, she exercised cruelty and ruthlessness until sci-fi shenanigans landed her in Prime Universe (the home universe of the shows), in 2257.
[Here I bypass a lot of stuff that happened in Discovery for the sake of brevity.]
Eventually she takes up an alias as “Madame Veronique du Franc”, proprietor of the pleasure space station Baraam, outside Federation territory. Section 31—essentially the Federation’s black ops—sends a team to blackmail Georgiou to help them intercept a superweapon, which happens to be en route to Baraam in the hands of a shady middleman named Dada Noe.
The team consists of their superstrong “augment” leader Alok Sahar, a mech-suit guy named Zeph, a seductress named Melle, a shape-shifting genius called Quasi, a buttoned-up Federation overseer named Garrett, and Fuzz, a microscopic Nanokin who pilots a teeny tiny spaceship and most often inhabits a black market Vulcan-looking android body.
Using some technologies called Phase Pods, Georgiou successfully separates the superweapon from Noe only to have it intercepted by a masked person also wearing a phase pod. Masked person kills Melle and escapes with the superweapon, but on the way Georgiou learns it is the Godsend, a quadrant-vaporizing weapon she had commissioned when she was Terran Emperor.
Georgiou convinces Sahar to form a partnership to recapture the weapon. They beam to his spaceship above a desolate planet where they interrogate Noe. They learn Noe is from the Mirror Universe, where he administered a facility that housed the Godsend. He hatched a plan to sell it and with the money escape to Prime Universe to retire in peace. His portal is an unknown but routinely opening rift between worlds. He tells them he is scheduled to meet his anonymous buyer when the rift next appears in four hours. He expects that if he does not deliver the weapon to his buyer—and the Terran Empire learns that the Godsend is gone—they will trace it to the rift, surge through, and conquer Prime. At that moment a massive explosion rips through the ship. The computer automatically beams the crew to the surface, but Noe dies in fiery debris. (Narratively convenient, but consider that the ship’s computer knew enough to beam our protagonists to safe, non-fiery-debris places, raising the possibility that it chose to murder Noe.)
Sahar says the explosion was sabotage by someone working with the still-unknown Godsend thief. One of them is a mole! Accusations fly, but Garrett focuses them on finding a derelict garbage scow she knows about, as a means to continue their mission. The team splits. Georgiou, Quasi, and Fuzz search for the scow. Sahar, Garette, and Zeph work to repair an antenna in an old Section 31 safehouse so they can warn the Federation of the impending danger.
Team Scow repairs the ship. We see Fuzz behaving a little strangely.
Meanwhile Zeph skips out on Team Antenna. While Sahar and Garrett search for him, the antenna gets activated, some message sent, and then the antenna is destroyed. The whole team rejoins and begins a search for Zeph. They find him dead. When they recover the video from his mech suit, they see something was controlling his suit and made him kill himself. Georgiou reasons that the mole must be Fuzz, who left his Vulcanbot on autopilot while he flew to Zeph to hook in and control him to commit the crimes and fly back to his bot. Thusly busted, Fuzz takes remote control of Zeph’s suit (grossly with Zeph’s corpse still in it) and the two try to escape on a float. The rest of the crew pursues in a second float, and there’s a vehicle combat sequence. Fuzz tells Georgiou that he’s been working with San. Then San beams Fuzz up to his ship. San speeds toward the rift to tell the Empire everything and begin the invasion. The remaining team gets the scow running and gives chase.
They catch up near the rift and the scow tries to delay its entry into the rift via tractor beam. Sahar and Georgiou beam to San’s ship to learn that San has initiated the Godsend. San fights Georgiou. Sahar fights Vulcanbot while Fuzz escapes to watch from a safe distance. On board the scow, Garrett forges a makeshift weapon in the ship’s hold and they release it at San’s ship. It lands and explodes, giving the heroes the upper hand in their respective fights. Georgiou grabs and activates the Godsend via biometric signature. Quasi manages to beam her and Sahar back to the scow just in time, leaving San, Fuzz, Vulcanbot, and the Godsend to be destroyed in the explosion as it passes back to the Mirror Universe and seal the rift forever. (And, presumably, something about the confluence of energies neuters the Godsend so it doesn’t go on to kill quadrillions in the quadrant where the rift happened to be, because that would be multiple, multiple genocides and sully whatever victory this is.)
The movie ends with the team back on the Baraam. They meet Wisp, Fuzz’ widow, piloting a second bootleg Vulcanbot. They receive a mission briefing that has them warping the Baraam (surprise, it’s also a spaceship) towards Turkana IV.
Whew.
Star Trek: Section 31 is primarily set in the 24th-century “Lost Era” between 2324–2326. This places it roughly 66 years after Discovery (2258) and about 40 years before The Next Generation. For continuity, the designers have to find some middle ground between the glowing, 3D, multiplanar translucency of Disco and the flat, 2D, highly-graphic, vibrant oranges-and-blues palette of LCARS. I think they did a really nice job. We see circular, glowing interfaces. We see hints of the fully realized LCARS to come.
There’s also a clear delineation between Federation/Section 31 interfaces, the mining colony interfaces, the few Terran Empire interfaces we see, and those of the foreign-language garbage scow.
San’s ship interior, by the way, is high-contrast red-on-black, and very pointy, making me wonder if the production designers have read my post on the Design of Evil, because it is practically an archetype of those patterns.
Anyway, now that we have a grasp on the plot, next let’s look more closely at those interfaces.
The Fritzes award honors the best interfaces in a full-length motion picture in the past year. Interfaces play a special role in our movie-going experience, and are a craft all their own that does not otherwise receive focused recognition. Best Assistant is a special award that I’m giving for the first time.
Ok but why now? Well, in March of this year I published a new non-fiction book about the design of technology that assists people doing things (as opposed to doing stuff for them). It’s called Designing Assistant Technology: AI That Makes People Smarter. In the book I lay out a framework for categorizing assistant interactions, and describe the risks and mitigations of having an assistant in the mix. I daresay it’s not only valuable for design, but for scriptwriters and futurists as well. If that intrigues you, look for a discount code near the end of this article.
Anyway, it gave me the idea to select the movie with the best examples of Assistants.
The 2026 Award for Best Assistants: M3gan 2.0
I know, I’m as surprised as you are.
The first movie, while smarter than I expected, seemed to be a horror flick that was using AI as set dressing. It did get a shout-out in the Fritzes 2024 for best HUD, but as I recall, its unbounded atomic optimization was just another way to frame it as a ruthless, efficient killer. But this second one seems to take the theme more seriously, and the scriptwriters did their homework.
In Part II of the book, I build on the see-think-do loop (that is core to interaction design) to identify the Five Universal Assists. These are the universal, exhaustive set of categories by which technology can assist users: Perceive, Know, Plan, Perform, and Reflect. And to my surprise, when you look close, there are examples of all five of the universal assists in M3gan 2.0, more than any other film in 2025.
Note: M3gan jumps bodies many times over the course of the movie, so you’ll see her described many times with the same name, but with vastly different appearances in the screen shots.
Early in the film, Cady discovers that the source code of Better Bionix is being hacked. When everyone comes over to see what’s on her screen, Tess says, “Oh, Jesus. She’s right. There’s stray commands all over the source code.” The screen we see doesn’t ask them (or us) to try and detect which out of the dozens on screen are suspect. Those lines are colored red to contrast greatly with the screen-green, and in case you were colorblind, they’re indented as well.
You might think that that M3gan’s alerting Gemma of the FBI home invasion to be an example of perceive, but she was sleeping when the alert comes. In that context, M3gan’s acting more as an agent. (More on that below.)
In act 2, Gemma asks M3gan to increase audio of two conversants at a noisy party, and that might as well be a canonical example. (And the first time she does it, M3gan substitutes audio in a very snarky way, reminding the audience that in a super-AI-mediated world, you cannot implicitly trust the media it controls, reminding us about over-reliance, another theme from Part III of the book.)
Know
In this assist, the tech helps users understand the meaning of what they’ve perceived, either in shallow ways such as names and categories, or very deeply.
HUDs have this built into the trope, and there are plenty of HUDs throughout.
But also, when beginning their joint hunt for AMELIA, M3gan explains that every battery Altwave (the villain corporation in the film) makes has a remote-controllable kill switch, explaining the meaning of what Gemma sees in the file.
When infiltrating Altwave, M3gan(toy) explains why AMELIA is there as well: She seeks to control Altwave’s cloud servers, which are half of North America. That control enables AMELIA to disable the economy, threatening “societal collapse in 10 to 12 working days”.
Plan
In this assist, it helps users plan their course of action, tactically or strategically.
When M3gan comes out of hiding and presents a deal to Gemma, she explains that she’s run a thousand simulations and if they don’t team up, more people die than if they didn’t. M3gan asks, “Who is the real killer in that situation?” Not having much of a choice, Gemma agrees.
A key part of the planning assist is helping users know what the best course of action is.
Perform
In this assist, the tech helps users perform some task.
One of the first scenes in the film has Tess and Cole demonstrating an exosuit. In their pitch they explain to the potential investor that its purpose is to help laborers avoid fatigue while performing physical tasks. To demonstrate, Cole lifts huge concrete blocks without showing any signs of exertion.
A few beats later, slimy Elon-Musk stand-in demonstrates how his neural chip helps him stand though he is ordinarily bound to his wheelchair.
In the climax, M3gan stows away on a neural chip forcibly implanted on Gemma. When Gemma dons an exosuit the AI helps her defeat many goons in hand to hand combat. It’s arguably acting as an agent here, since Gemma isn’t trying to build those skills. (Similarly when Gemma gets knocked unconscious, M3gan controls the exosuit to animate her body anyway, something we also see in Section 31, but more on this example in a later post.)
Reflect
In this assist—the most abstract of them—the technology helps users reflect on things to turn experience into knowledge, or to question their goals and future tactics.
There’s a lot less of this here, just like there is in the real world. But, we see some of it.When Cady asks M3gan(half-formed) how she can feel anything, M3gan replies, “Can you explain why you feel things?” It’s rhetorical in context, but exactly the sort of thing that a reflection assistant might ask.
When Gemma is spiraling about her parenting in the basement, M3gan(souped up) takes a moment to share counterexamples. “I saw you wake up every day at 4:00 A.M., staring at the ceiling contemplating what the future holds for her…I watched you make homemade lunches with fresh-baked sourdough…I watched you help her with her homework, even though it always ended in a fight…Gemma, it’s not a failure to feel guilt or that you’re not enough. It’s part of the job.” It’s not the best fit for the definition of this assist I give in the book, but it’s the closest thing in the movie and the closest thing in my survey of the year’s films.
Also agents
There are also many examples where M3gan(AI) acts as an agent on their behalf, but that was my last book, so I’ll skip getting into those examples. But as you watch the movie, keep an eye out for additional shouts out to the paperclip thought experiment (a metaphor for the threat of instrumental convergence), allusions to the Xerox WorkCentre scanner bug, and of course super AI as an existential threat. The whole plot can be seen as an example of Bostrom’s a priori argument that multiple super AIs are the most stable scenario. All this is why I say that the writers seemed to have done their homework..
I’m a lot less fond about the guy wanting to regulate/eliminate AI is painted the bad guy, but having positioned M3gan as sentient and the antihero of the film, I’m not sure what else they could do. But I wish it didn’t valorize AI as equivalent to humans despite all of that. We have enough LeMoinian panic about large language models as it is.
Anyway, congratulations to M3gan 2.0 for showing so many examples of assistants throughout. If you’re interested in getting the book, you can get 20% off if you purchase from Rosenfeldmedia.com and use the code “scifi26” during checkout. Use this power only for good.
And let me know in comments if you think of other examples of assistants across the year.
The Fritzes award honors the best interfaces in a full-length motion picture in the past year. Interfaces play a special role in our movie-going experience, and are a craft all their own that does not otherwise receive focused recognition.
Today we’ll be covering Best Believable. These movies’ interfaces adhere to solid computer-human-interaction principles and believable interactions. They engage us in the story world by being convincing.
The 2026 Award goes to: The Running Man
This second adaptation of Stephen King’s novel knocks it out of the park for the plot-central interfaces: The runner cuff and R-Cam box, the hideous sousveillance phone app for “fans”, the service design of the “free-v” show, and the in-home snitch interfaces. They lean towards narrative (missing a few things real-world counterparts would need), but all help articulate this dystopian world and the circumstances that drive the action. Moreover, I feel quite certain not making good real-world models of these horrible things is the right thing to do, especially given *gestures vaguely at the kakistocracy*.
On top of that it also has lots of awesome everyday interfaces, and it takes a level of commitment on the part of the filmmakers to go that deep in the worldbuilding. There’s a videophone interface with shades of Blade Runner. There’s a mailbox that signals its readiness and lifts off immediately after receiving a letter. (Though I would have flipped those red and green colors, so red meant “don’t put mail in here” and green meant “ready to receive”, but my invitation was lost in the mail.) The fare interfaces in the taxi. The self-driving interface of the citizen car. The piloting interfaces aboard the network plane. It’s all uncluttered, straightforward, and believable. Really well done, really well presented, and that’s hard to do in intense-action movies.
Also check out: War of the Worlds (2025)
It got universally panned. Fair enough, neither ubiquitous government surveillance nor the current DHS bears valorization. (Also the virus-but-its-digital twist was already done), but I am impressed that this take on the classic Wells story is told almost entirely through interfaces, and each of them is detailed and mostly-realistic. The editing around the interface can be dizzying, and I wondered why William Radford had to do so much digital hunting at the beginning when an assistant should have been guiding his attention. But it’s impressive to bring that tale to life mostly through this unsung medium.
Also check out: Companion
With soft echoes of the interfaces in Westworld (2016), the interfaces in Companion control android and gynoid companions. (Yes, that term is deliberately coy.) They are clean and simple, which underscores the robots’ horror that they are under that much control by their owners.
My hackles are raised from “Intelligence” being a single slider. Intelligence is much more complicated than that, and this notion that it’s a single scalar variable has done a lot of damage over time. Even if they’d had a little expando control, it would have pointed at the idea that we’re looking at a simplification. Also I wish they’d provided a live preview of the eye color, because even with its intended use—of an owner controlling their companion’s eye color—this control has them glancing up to see the effect and then back down again to adjust, which is not a satisfying feedback loop. I use this very control as an example of a “plan” assistant in my new book. Hey, all of Hollywood: Buy it!
Our next 3D file browsing system is from the 1994 film Disclosure. Thanks to site reader Patrick H Lauke for the suggestion.
Like Jurassic Park, Disclosure is based on a Michael Crichton novel, although this time without any dinosaurs. (Would-be scriptwriters should compare the relative success of these two films when planning a study program.) The plot of the film is corporate infighting within Digicom, manufacturer of high tech CD-ROM drives—it was the 1990s—and also virtual reality systems. Tom Sanders, executive in charge of the CD-ROM production line, is being set up to take the blame for manufacturing failures that are really the fault of cost-cutting measures by rival executive Meredith Johnson.
The Corridor: Hardware Interface
The virtual reality system is introduced at about 40 minutes, using the narrative device of a product demonstration within the company to explain to the attendees what it does. The scene is nicely done, conveying all the important points we need to know in two minutes. (To be clear, some of the images used here come from a later scene in the film, but it’s the same system in both.)
The process of entangling yourself with the necessary hardware and software is quite distinct from interacting with the VR itself, so let’s discuss these separately, starting with the physical interface.
Tom wearing VR headset and one glove, being scanned. Disclosure (1994)
In Disclosure the virtual reality user wears a headset and one glove, all connected by cables to the computer system. Like most virtual reality systems, the headset is responsible for visual display, audio, and head movement tracking; the glove for hand movement and gesture tracking.
There are two “laser scanners” on the walls. These are the planar blue lights, which scan the user’s body at startup. After that they track body motion, although since the user still has to wear a glove, the scanners presumably just track approximate body movement and orientation without fine detail.
Lastly, the user stands on a concave hexagonal plate covered in embedded white balls, which allows the user to “walk” on the spot.
Closeup of user standing on curved surface of white balls. Disclosure (1994)
Searching for Evidence
The scene we’re most interested in takes place later in the film, the evening before a vital presentation which will determine Tom’s future. He needs to search the company computer files for evidence against Meredith, but discovers that his normal account has been blocked from access. He knows though that the virtual reality demonstrator is on display in a nearby hotel suite, and also knows about the demonstrator having unlimited access. He sneaks into the hotel suite to use The Corridor. Tom is under a certain amount of time pressure because a couple of company VIPs and their guests are downstairs in the hotel and might return at any time.
The first step for Tom is to launch the virtual reality system. This is done from an Indy workstation, using the regular Unix command line.
The command line to start the virtual reality system. Disclosure (1994)
Next he moves over to the VR space itself. He puts on the glove but not the headset, presses a key on the keyboard (of the VR computer, not the workstation), and stands still for a moment while he is scanned from top to bottom.
Real world Tom, wearing one VR glove, waits while the scanners map his body. Disclosure (1994)
On the left is the Indy workstation used to start the VR system. In the middle is the external monitor which will, in a moment, show the third person view of the VR user as seen earlier during the product demonstration.
Now that Tom has been scanned into the system, he puts on the headset and enters the virtual space.
The Corridor: Virtual Interface
“The Corridor,” as you’ve no doubt guessed, is a three dimensional file browsing program. It is so named because the user will walk down a corridor in a virtual building, the walls lined with “file cabinets” containing the actual computer files.
Three important aspects of The Corridor were mentioned during the product demonstration earlier in the film. They’ll help structure our tour of this interface, so let’s review them now, as they all come up in our discussion of the interfaces.
There is a voice-activated help system, which will summon a virtual “Angel” assistant.
Since the computers themselves are part of a multi-user network with shared storage, there can be more than one user “inside” The Corridor at a time. Users who do not have access to the virtual reality system will appear as wireframe body shapes with a 2D photo where the head should be.
There are no access controls and so the virtual reality user, despite being a guest or demo account, has unlimited access to all the company files. This is spectacularly bad design, but necessary for the plot.
With those bits of system exposition complete, now we can switch to Tom’s own first person view of the virtual reality environment.
Virtual world Tom watches his hands rezzing up, right hand with glove. Disclosure (1994)
There isn’t a real background yet, just abstract streaks. The avatar hands are rezzing up, and note that the right hand wearing the glove has a different appearance to the left. This mimics the real world, so eases the transition for the user.
Overlaid on the virtual reality view is a Digicom label at the bottom and four corner brackets which are never explained, although they do resemble those used in cameras to indicate the preferred viewing area.
To the left is a small axis indicator, the three green lines labeled X, Y, and Z. These show up in many 3D applications because, silly though it sounds, it is easy in a 3D computer environment to lose track of directions or even which way is up. A common fix for the user being unable to see anything is just to turn 180 degrees around.
We then switch to a third person view of Tom’s avatar in the virtual world.
Tom is fully rezzed up, within cloud of visual static. Disclosure (1994)
This is an almost photographic-quality image. To remind the viewers that this is in the virtual world rather than real, the avatar follows the visual convention described in chapter 4 of Make It So for volumetric projections, with scan lines and occasional flickers. An interesting choice is that the avatar also wears a “headset”, but it is translucent so we can see the face.
Now that he’s in the virtual reality, Tom has one more action needed to enter The Corridor. He pushes a big button floating before him in space.
Tom presses one button on a floating control panel. Disclosure (1994)
This seems unnecessary, but we can assume that in the future of this platform, there will be more programs to choose from.
The Corridor rezzes up, the streaks assembling into wireframe components which then slide together as the surfaces are shaded. Tom doesn’t have to wait for the process to complete before he starts walking, which suggests that this is a Level Of Detail (LOD) implementation where parts of the building are not rendered in detail until the user is close enough for it to be worth doing.
Tom enters The Corridor. Nearby floor and walls are fully rendered, the more distant section is not complete. Disclosure (1994)
The architecture is classical, rendered with the slightly artificial-looking computer shading that is common in 3D computer environments because it needs much less computation than trying for full photorealism.
Instead of a corridor this is an entire multistory building. It is large and empty, and as Tom is walking bits of architecture reshape themselves, rather like the interior of Hogwarts in Harry Potter.
Although there are paintings on some of the walls, there aren’t any signs, labels, or even room numbers. Tom has to wander around looking for the files, at one point nearly “falling” off the edge of the floor down an internal air well. Finally he steps into one archway room entrance and file cabinets appear in the walls.
Tom enters a room full of cabinets. Disclosure (1994)
Unlike the classical architecture around him, these cabinets are very modern looking with glowing blue light lines. Tom has found what he is looking for, so now begins to manipulate files rather than browsing.
Virtual Filing Cabinets
The four nearest cabinets according to the titles above are
Communications
Operations
System Control
Research Data.
There are ten file drawers in each. The drawers are unmarked, but labels only appear when the user looks directly at it, so Tom has to move his head to centre each drawer in turn to find the one he wants.
Tom looks at one particular drawer to make the title appear. Disclosure (1994)
The fourth drawer Tom looks at is labeled “Malaysia”. He touches it with the gloved hand and it slides out from the wall.
Tom withdraws his hand as the drawer slides open. Disclosure (1994)
Inside are five “folders” which, again, are opened by touching. The folder slides up, and then three sheets, each looking like a printed document, slide up and fan out.
Axis indicator on left, pointing down. One document sliding up from a folder. Disclosure (1994)
Note the tilted axis indicator at the left. The Y axis, representing a line extending upwards from the top of Tom’s head, is now leaning towards the horizontal because Tom is looking down at the file drawer. In the shot below, both the folder and then the individual documents are moving up so Tom’s gaze is now back to more or less level.
Close up of three “pages” within a virtual document. Disclosure (1994)
At this point the film cuts away from Tom. Rival executive Meredith, having been foiled in her first attempt at discrediting Tom, has decided to cover her tracks by deleting all the incriminating files. Meredith enters her office and logs on to her Indy workstation. She is using a Command Line Interface (CLI) shell, not the standard SGI Unix shell but a custom Digicom program that also has a graphical menu. (Since it isn’t three dimensional it isn’t interesting enough to show here.)
Tom uses the gloved hand to push the sheets one by one to the side after scanning the content.
Tom scrolling through the pages of one folder by swiping with two fingers. Disclosure (1994)
Quick note: This is harder than it looks in virtual reality. In a 2D GUI moving the mouse over an interface element is obvious. In three dimensions the user also has to move their hand forwards or backwards to get their hand (or finger) in the right place, and unless there is some kind of haptic feedback it isn’t obvious to the user that they’ve made contact.
Tom now receives a nasty surprise.
The shot below shows Tom’s photorealistic avatar at the left, standing in front of the open file cabinet. The green shape on the right is the avatar of Meredith who is logged in to a regular workstation. Without the laser scanners and cameras her avatar is a generic wireframe female humanoid with a face photograph stuck on top. This is excellent design, making The Corridor usable across a range of different hardware capabilities.
Tom sees the Meredith avatar appear. Disclosure (1994)
Why does The Corridor system place her avatar here? A multiuser computer system, or even just a networked file server, obviously has to know who is logged on. Unix systems in general and command line shells also track which directory the user is “in”, the current working directory. Meredith is using her CLI interface to delete files in a particular directory so The Corridor can position her avatar in the corresponding virtual reality location. Or rather, the avatar glides into position rather than suddenly popping into existence: Tom is only surprised because the documents blocked his virtual view.
Quick note: While this is plausible, there are technical complications. Command line users often open more than one shell at a time in different directories. In such a case, what would The Corridor do? Duplicate the wireframe avatar in each location? In the real world we can’t be in more than one place at a time, would doing so contradict the virtual reality metaphor?
There is an asymmetry here in that Tom knows Meredith is “in the system” but not vice versa. Meredith could in theory use CLI commands to find out who else is logged on and whether anyone was running The Corridor, but she would need to actively seek out that information and has no reason to do so. It didn’t occur to Tom either, but he doesn’t need to think about it, the virtual reality environment conveys more information about the system by default.
We briefly cut away to Meredith confirming her CLI delete command. Tom sees this as the file drawer lid emitting beams of light which rotate down. These beams first erase the floating sheets, then the folders in the drawer. The drawer itself now has a red “DELETED” label and slides back into the wall.
Tom watches Meredith deleting the files in an open drawer. Disclosure (1994)
Tom steps further into the room. The same red labels appear on the other file drawers even though they are currently closed.
Tom watches Meredith deleting other, unopened, drawers. Disclosure (1994)
Talking to an Angel
Tom now switches to using the system voice interface, saying “Angel I need help” to bring up the virtual reality assistant. Like everything else we’ve seen in this VR system the “angel” rezzes up from a point cloud, although much more quickly than the architecture: people who need help tend to be more impatient and less interested in pausing to admire special effects.
The voice assistant as it appears within VR. Disclosure (1994)
Just in case the user is now looking in the wrong direction the angel also announces “Help is here” in a very natural sounding voice.
The angel is rendered with white robe, halo, harp, and rapidly beating wings. This is horribly clichéd, but a help system needs to be reassuring in appearance as well as function. An angel appearing as a winged flying serpent or wheel of fire would be more original and authentic (yes, really: Biblically Accurate Angels) but users fleeing in terror would seriously impact the customer satisfaction scores.
Now Tom has a short but interesting conversation with the angel, beginning with a question:
Tom
Is there any way to stop these files from being deleted?
Angel
I’m sorry, you are not level five.
Tom
Angel, you’re supposed to protect the files!
Angel
Access control is restricted to level five.
Tom has made the mistake, as described in chapter 9 Anthropomorphism of the book, of ascribing more agency to this software program than it actually has. He thinks he is engaged in a conversational interface (chapter 6 Sonic Interfaces) with a fully autonomous system, which should therefore be interested in and care about the wellbeing of the entire system. Which it doesn’t, because this is just a limited-command voice interface to a guide.
Even though this is obviously scripted, rather than a genuine error I think this raises an interesting question for real world interface designers: do users expect that an interface with higher visual quality/fidelity will be more realistic in other aspects as well? If a voice interface assistant has a simple polyhedron with no attempt at photorealism (say, like Bit in Tron) or with zoomorphism (say, like the search bear in Until the End of the World) will users adjust their expectations for speech recognition downwards? I’m not aware of any research that might answer this question. Readers?
Despite Tom’s frustration, the angel has given an excellent answer – for a guide. A very simple help program would have recited the command(s) that could be used to protect files against deletion. Which would have frustrated Tom even more when he tried to use one and got some kind of permission denied error. This program has checked whether the user can actually use commands before responding.
This does contradict the earlier VR demonstration where we were told that the user had unlimited access. I would explain this as being “unlimited read access, not write”, but the presenter didn’t think it worthwhile to go into such detail for the mostly non-technical audience.
Tom is now aware that he is under even more time pressure as the Meredith avatar is still moving around the room. Realising his mistake, he uses the voice interface as a query language.
“Show me all communications with Malaysia.” “Telephone or video?” “Video.”
This brings up a more conventional looking GUI window because not everything in virtual reality needs to be three-dimensional. It’s always tempting for a 3D programmer to re-implement everything, but it’s also possible to embed 2D GUI applications into a virtual world.
Tom looks at a conventional 2D display of file icons inside VR. Disclosure (1994)
The window shows a thumbnail icon for each recorded video conference call. This isn’t very helpful, so Tom again decides that a voice query will be much faster than looking at each one in turn.
“Show me, uh, the last transmission involving Meredith.”
There’s a short 2D transition effect swapping the thumbnail icon display for the video call itself, which starts playing at just the right point for plot purposes.
Tom watches a previously recorded video call made by Meredith (right). Disclosure (1994)
While Tom is watching and listening, Meredith is still typing commands. The camera orbits around behind the video conference call window so we can see the Meredith avatar approach, which also shows us that this window is slightly three dimensional, the content floating a short distance in front of the frame. The film then cuts away briefly to show Meredith confirming her “kill all” command. The video conference recordings are deleted, including the one Tom is watching.
Tom is informed that Meredith (seen here in the background as a wireframe avatar) is deleting the video call. Disclosure (1994)
This is also the moment when the downstairs VIPs return to the hotel suite, so the scene ends with Tom managing to sneak out without being detected.
Virtual reality has saved the day for Tom. The documents and video conference calls have been deleted by Meredith, but he knows that they once existed and has a colleague retrieve the files he needs from the backup tapes. (Which is good writing: the majority of companies shown in film and TV never seem to have backups for files, no matter how vital.) Meredith doesn’t know that he knows, so he has the upper hand to expose her plot.
Analysis
How believable is the interface?
I won’t spend much time on the hardware, since our focus is on file browsing in three dimensions. From top to bottom, the virtual reality system starts as believable and becomes less so.
Hardware
The headset and glove look like real VR equipment, believable in 1994 and still so today. Having only one glove is unusual, and makes impossible some of the common gesture actions described in chapter 5 of Make It So, which require both hands.
The “laser scanners” that create the 3D geometry and texture maps for the 3D avatar and perform real time body tracking would more likely be cameras, but that would not sound as cool.
And lastly the walking platform apparently requires our user to stand on large marbles or ball bearings and stay balanced while wearing a headset. Uh…maybe…no. Apologetics fails me. To me it looks like it would be uncomfortable to walk on, almost like deterrent paving.
Software
The Corridor, unlike the 3D file browser used in Jurassic Park, is a special effect created for the film. It was a mostly-plausible, near future system in 1994, except for the photorealistic avatar. Usually this site doesn’t discuss historical context (the “new criticism” stance), but I think in this case it helps to explain how this interface would have appeared to audiences almost two decades ago.
I’ll start with the 3D graphics of the virtual building. My initial impression was that The Corridor could have been created as an interactive program in 1994, but that was my memory compressing the decade. During the 1990s 3D computer graphics, both interactive and CGI, improved at a phenomenal rate. The virtual building would not have been interactive in 1994, was possible on the most powerful systems six years later in 2000, and looks rather old-fashioned compared to what the game consoles of the 21st C can achieve.
For the voice interface I made the opposite mistake. Voice interfaces on phones and home computing appliances have become common in the second decade of the 21st C, but in reality are much older. Apple Macintosh computers in 1994 had text-to-speech synthesis with natural sounding voices and limited vocabulary voice command recognition. (And without needing an Internet connection!) So the voice interface in the scene is believable.
The multi-user aspects of The Corridor were possible in 1994. The wireframe avatars for users not in virtual reality are unflattering or perhaps creepy, but not technically difficult. As a first iteration of a prototype system it’s a good attempt to span a range of hardware capabilities.
The virtual reality avatar, though, is not believable for the 1990s and would be difficult today. Photographs of the body, made during the startup scan, could be used as a texture map for the VR avatar. But live video of the face would be much more difficult, especially when the face is partly obscured by a headset.
How well does the interface inform the narrative of the story?
The virtual reality system in itself is useful to the overall narrative because it makes the Digicom company seem high tech. Even in 1994 CD-ROM drives weren’t very interesting.
The Corridor is essential to the tension of the scene where Tom uses it to find the files, because otherwise the scene would be much shorter and really boring. If we ignore the virtual reality these are the interface actions:
Tom reads an email.
Meredith deletes the folder containing those emails.
Tom finds a folder full of recorded video calls.
Tom watches one recorded video call.
Meredith deletes the folder containing the video calls.
Imagine how this would have looked if both were using a conventional 2D GUI, such as the Macintosh Finder or MS Windows Explorer. Double click, press and drag, double click…done.
The Corridor slows down Tom’s actions and makes them far more visible and understandable. Thanks to the virtual reality avatar we don’t have to watch an actor push a mouse around. We see him moving and swiping, be surprised and react; and the voice interface adds extra emotion and some useful exposition. It also helps with the plot, giving Tom awareness of what Meredith is doing without having to actively spy on her, or look at some kind of logs or recordings later on.
Meredith, though, can’t use the VR system because then she’d be aware of Tom as well. Using a conventional workstation visually distinguishes and separates Meredith from Tom in the scene.
So overall, though the “action” is pretty mundane, it’s crucial to the plot, and the VR interface helps make this interesting and more engaging.
How well does the interface equip the character to achieve their goals?
As described in the film itself, The Corridor is a prototype for demonstrating virtual reality. As a file browser it’s awful, but since Tom has lost all his normal privileges this is the only system available, and he does manage to eventually find the files he needs.
At the start of the scene, Tom spends quite some time wandering around a vast multi-storey building without a map, room numbers, or even coordinates overlaid on his virtual view. Which seems rather pointless because all the files are in one room anyway. As previously discussed for Johnny Mnemonic, walking or flying everywhere in your file system seems like a good idea at first, but often becomes tedious over time. Many actual and some fictional 3D worlds give users the ability to teleport directly to any desired location.
Then the file drawers in each cabinet have no labels either, so Tom has to look carefully at each one in turn. There is so much more the interface could be doing to help him with his task, and even help the users of the VR demo learn and explore its technology as well.
Contrast this with Meredith, who uses her command line interface and 2D GUI to go through files like a chainsaw.
Tom becomes much more efficient with the voice interface. Which is just as well, because if he hadn’t, Meredith would have deleted the video conference recordings while he was still staring at virtual filing cabinets. However neither the voice interface nor the corresponding file display need three dimensional graphics.
There is hope for version 2.0 of The Corridor, even restricting ourselves to 1994 capabilities. The first and most obvious is to copy 2D GUI file browsers, or the 3D file browser from Jurassic Park, and show the corresponding text name next to each graphical file or folder object. The voice interface is so good that it should be turned on by default without requiring the angel. And finally add some kind of map overlay with a you are here moving dot, like the maps that players in 3D games such as Doom could display with a keystroke.
Film making challenge: VR on screen
Virtual reality (or augmented reality systems such as Hololens) provide a better viewing experience for 3D graphics by creating the illusion of real three dimensional space rather than a 2D monitor. But it is always a first person view and unlike conventional 2D monitors nobody else can usually see what the VR user is seeing without a deliberate mirroring/debugging display. This is an important difference from other advanced or speculative technologies that film makers might choose to include. Showing a character wielding a laser pistol instead of a revolver or driving a hover car instead of a wheeled car hardly changes how to stage a scene, but VR does.
So, how can we show virtual reality in film?
There’s the first-person view corresponding to what the virtual reality user is seeing themselves. (Well, half of what they see since it’s not stereographic, but it’s cinema VR, so close enough.) This is like watching a screencast of someone else playing a first person computer game, the original active experience of the user becoming passive viewing by the audience. Most people can imagine themselves in the driving seat of a car and thus make sense of the turns and changes of speed in a first person car chase, but the film audience probably won’t be familiar with the VR system depicted and will therefore have trouble understanding what is happening. There’s also the problem that viewing someone else’s first-person view, shifting and changing in response to their movements rather than your own, can make people disoriented or nauseated.
A third-person view is better for showing the audience the character and the context in which they act. But not the diegetic real-world third-person view, which would be the character wearing a geeky headset and poking at invisible objects. As seen in Disclosure, the third person view should be within the virtual reality.
But in doing that, now there is a new problem: the avatar in virtual reality representing the real character. If the avatar is too simple the audience may not identify it with the real world character and it will be difficult to show body language and emotion. More realistic CGI avatars are increasingly expensive and risk falling into the Uncanny Valley. Since these films are science fiction rather than factual, the easy solution is to declare that virtual reality has achieved the goal of being entirely photorealistic and just film real actors and sets. Adding the occasional ripple or blur to the real world footage to remind the audience that it’s meant to be virtual reality, again as seen in Disclosure, is relatively cheap and quick. So, solving all these problems results in the cinematic trope we can call Extradiegetic Avatars, which are third-person, highly-lifelike “renderings” of characters, with a telltale Hologram Projection Imperfection for audience readability, that may or may not be possible within the world of the film itself.
Our first example is a single scene from Jurassic Park, set entirely in the control room of Isla Nublar. Apologies in advance for repeating some material already covered by the book and website, but it is necessary to focus on the aspects that are of interest to this study.
Drs. Sattler and Grant enter the control room along with Lex and Tim. Jurassic Park (1993)
The eponymous Jurassic Park is heavily automated, with the entire park designed to be controlled from the computer systems in this room. Villainous computer system designer Nedry took advantage of this to shut down systems across the entire park, releasing all the dinosaurs, to cover his industrial espionage. Most of the park staff had already been evacuated due to a storm warning, and the small team of core technical staff who remained have, by this point in the film, all been killed by dinosaurs. (Including Nedry—who, had he been given time for extrospection, would probably have rethought those aspects of his plan concerning the release of carnivorous dinosaurs.)
Four of the survivors have gathered in the control room after managing to restore the power, but must still restart the various computer systems. They have discovered that the computer control extends down to door locks, which are consequently not working and have suddenly become the number one priority due to the velociraptors trying to break in.
Our interface user is Lex, a teenage visitor, being given an advance tour of the park before its official opening. The others are Dr Grant, paleontologist; Dr Sattler, paleobotanist; and Lex’s younger brother Tim, dinosaur enthusiast. As a self -described computer hacker Lex is easily the best person qualified to work with the computers as everyone else in the room only has expertise in subjects more than sixty-six million years old.
Lex sitting before the computer and looking at the /usr directory in the 3D file browser. Jurassic Park (1993)
The computers were all rebooted when the power came back on but the programs that control Jurassic Park did not automatically restart. Dr. Sattler spent a moment in front of the computer with Lex, but all she seemed to do is deactivate the screen saver. It’s up to Lex to find and start whatever program runs the security systems for the control room.
Backworlding aside: Unix-savvy viewers might be wondering why these control programs, since they are critical to the park functionality, don’t automatically start when the computer is rebooted. I hazard that perhaps normally they would, but Nedry turned this off to ensure that no-one could undo his sabotage before he got back. The file system of the computer is rendered as a tree, with directory names (/usr in the image above) shown as text labels, the contents of each directory shown as LEGO-like blocks, and lines linking directories to subdirectories.
The park directory, and two levels of subdirectories in the distance. Jurassic Park (1993)
Most of the information is drawn on a flat two-dimensional plane. The third dimension is used to present information about the number of, and perhaps sizes, of the files in each directory. Note in the image above that the different directories below the foremost park block have different sized heights and areas.
Rendering this plane in perspective, rather than as a conventional 2D window, means that areas closest to the viewpoint can be seen in detail, but there is still some information given about the directories further away. In the image above, the subdirectory of park on the right is clearly smaller than the others, even though we can’t make out the actual name, and also has a number of larger subdirectories.
Up close we can see that each file can have its own icon on top, presumably representing the type of file.
Individual blue files within one directory, and subdirectories beyond. Jurassic Park (1993)
The viewpoint stays at a constant height above the ground plane. Moving around is done with the mouse, using it as a game-style directional controller when the mouse button is held down rather than as an absolute pointing device. It is almost “walking” rather than “flying” but there is a slight banking effect when Lex changes direction.
Closeup of Lex’s hand on the mouse, pressing the left mouse button. Jurassic Park (1993)
Here Lex has browsed through the hierarchy and discovered a promising file. She selects it, but we don’t see how, and a spotlight or sunbeam indicates the selection.
The “Visitors Center” icon highlighted by a beam from above. Jurassic Park (1993)
This is the last of the 3D interactions. The 3D file browser is just a file browser, not an entire operating system or virtual environment, so opening a file or program will open a new interface.
Tagged: 3D, 3D rendering, blue, cathode ray tube, color, comparison, constant movement, control room, cyan, desk, direct manipulation, disambiguation, finger press, flight control, flying, green, icon, interaction design, light, lighting, map, missing information, motion cue, navigating, pink, point to select, projection rays, selection, sense making, stress, up is more
When Lex runs this program (again, we don’t see how) it is in fact the security system controller for the visitor centre, including the control room. This has a conventional 2D GUI interface and she immediately switches everything on.
The 2D GUI. Security window in green on left, boot progress screen in blue on right. Jurassic Park (1993)
Success! Well, it would be if the control room did not also have very large windows which are apparently not velociraptor-proof. But the subsequent events, and interfaces, are not our concern.
Analysis
This isn’t a report card, since those are given to complete films or properties, not individual interfaces. But we can ask the same questions.
How believable is the interface?
In this case, very believable. The 3D file browser seen in the film is a real program that was shipped with the computers used in the film. It was created by the manufacturer Silicon Graphics as a demonstration of 3D capabilities, not as an effect just for this film.
How well does the interface inform the narrative of the story?
It supports the narrative, but isn’t essential — there’s plenty of drama and tension due to the velociraptors at the door, and the scene would probably still work if the camera only showed Lex, not the interface. The major contribution of using the 3D file browser is to keep the technology of Jurassic Park seemingly a little more advanced than normal for the time. Apart from dinosaurs, both the book and the film try not to introduce obviously science fictional elements. A 2D file browser (they did exist for Unix computers at the time, including the SGI computers shown in the film) would have been recognisable but boring. The 3D file browser looks advanced while still being understandable.
How well does the interface equip the characters to achieve their goals?
The most interesting question, to which the answer is that it works very well. One problem, visible in the film, is that because the labels are rendered on the 2D ground plane, users have to navigate close to a file or a folder to read its name. Rotating the names to vertical and to always face the user (“billboarding”) would have made them recognisable from further away.
Both at the time of the film and today some computer people will argue that Lex can’t be a real computer hacker because she doesn’t use the command line interface. Graphical user interfaces are considered demeaning. I disagree. Lex is in a situation familiar to many system administrators, having to restore computer functionality after an unexpected power loss. (Although the velociraptors at the door are a little more hostile than your typical user demanding to know when the system will be back up.) Earlier in the film we saw Ray Arnold, one of the technical staff, trying to restore the system and he was using the command line interface.
Ray Arnold sitting before SGI computer, typing into blue command line window. Jurassic Park (1993)
So why does Lex use the 3D file browser? Because, unlike Ray Arnold, she doesn’t know which programs to run. Rebooting the computers is not enough. The various programs that control Jurassic Park are all custom pieces of software developed by Nedry, and nothing we’ve seen indicates that he would have been considerate enough to write a user guide or reference manual or even descriptive file names. Everyone who might have known which programs do what is either dead or off the island.
Lex needs an interface that lets her quickly search through hundreds or even thousands of files without being able to specify precise search criteria. For a problem involving recognition, “you’ll know it when you see it”, a graphical user interface is superior to a command line.
Film making challenge: diegetic computers
Writing for SciFiInterfaces can be quite educational. Chris asked me to write about the “diegetic” aspects of rendering 3D graphics in film, and I agreed to do so without actually knowing what that meant. Fortunately for me it isn’t complicated. Diegetic images or sounds belong to what we see in the scene itself, for instance characters and their dialog or hearing the music a violinist who is on-screen is playing; while non-diegetic are those that are clearly artefacts of watching a film, such as subtitles, voice overs, or the creepy violin music that is playing as a character explores a haunted house—we don’t imagine there is some violinist in there with them.
So, SciFiinterfaces.com focuses on the diegetic computer interfaces used by characters within the film or TV show itself. We’ve just been discussing the 3D file browser in Jurassic Park. Which, since it was a real interactive program, just meant pointing a camera at the actor and the computer screen, right?
It’s not that easy. Our human eyes and brain do an enormous amount of interpolation and interpretation of what we actually see. There’s the persistence of vision effect that allows us to watch a film in a cinema and see it as fluid motion, even though for a significant percentage of the time we’re actually looking at a blank wall while the projector shutter is closed. Cameras, whether film or digital, take discrete snapshots and are not so easily fooled, leading to various odd effects. One example that’s been known since the early days of filmmaking is that at certain speeds spoked wheels can appear to be rotating far more slowly than expected, or even to be rotating backwards.
Jurassic Park was made in the days when television sets and computer monitors used Cathode Ray Tube (CRT) technology. A CRT cannot display an entire frame at once, instead starting at the top left and drawing pixels line by line (“scan lines”) to the bottom. Just as the top line of pixels fades out, the new frame begins. At 50 or 60 frames a second we see only continuous moving images thanks to our persistence of vision; but a camera, usually running at 24 frames a second, will capture a dark line moving slowly down the screen and the images themselves will flicker. This was a common sight in TV news reports and sometimes in films of the time, when computer monitors were in the background. Here’s a shot from the 1995 film The Net where the new frames have been half-drawn:
View from above of computer expo. The two stacked monitors center right are not genlocked, showing crawl lines. The Net (1995)
One technique that avoids this is to film the computer interface in isolation and composite the graphics into the footage afterwards. This is very easy in the 21st century with all digital processing but Jurassic Park was made in the days of optical compositing, which is more expensive and limits the number of images that can be combined before losing picture quality.
So to shoot CRT monitors with their graphics live, the camera shutter opening must be synchronised to the start of each frame. In TV studios and film sets this is done with genlocking, connecting all the monitors and cameras via cables to a single electronic timing signal. This was apparently the technique used in Jurassic Park, with impressive results. In one control room scene the camera pans across at least eight different monitors, and none of them are flickering.
The origin story here is that I wanted to review Hackers, a film I enjoy and Chris describes as “awesome/ly bad”. However, Hackers isn’t science fiction. Well, I could argue that it is set in an alternate reality where computer hackers are all physically attractive with fashionable tastes in music and clothing, but that isn’t enough. The film was set firmly in the present day (of 1995) and while the possibilities of computer hacking may be exaggerated for dramatic purposes, all the computers themselves are quite conventional for the time. (And therefore appear quaint and outdated today.)
With the glorious exception of the three dimensional file storage system on the “Gibson” mainframe. This fantastic combination of hardware and software was clearly science fiction then, and remains so today. While one futuristic element is not enough to justify a full review of Hackers, it did start us thinking. The film Jurassic Park also has a 3D file system navigator, which wasn’t covered in depth by either the book or the online review. And when Chris reached out to the website readers, they provided more examples of 3D file systems being added to otherwise mundane computer systems.
So what we have here is a review of a particular interface trope rather than an entire film or TV show: the three dimensional file browsing and navigation interface.
Scope
This review is specifically limited to interfaces that are recognisably file systems, where people search for and manipulate individual documents or programs. Cyberspace, a 3D representation of the Internet or World Wide Web, is too broad a topic, and better covered in individual reviews such as those for Ghost in the Shell and Johnny Mnemonic.
I also originally intended to only include non-science fiction films and shows but Jurassic Park is an exception. Jurassic Park has been reviewed, both in the book and on the website, but the 3D file system was a comparatively minor element. It is included here as a well known example for comparison.
The SciFiInterfaces readership also provided examples of research papers for 3D file system browsing and navigation—rather more numerous than actual production systems, even today. These will inform the reviews but not be discussed individually.
Because we are reviewing a topic, not a particular film or TV show, the usual plot summaries will be shortened to just those aspects that involve the 3D file system. As a bonus, we can also compare and contrast the different interfaces and how they are used. The worlds of Ghost in the Shell and Johnny Mnemonic are so different that it would be unfair to judge individual interfaces against each other, but for this review we are considering 3D file systems that have been grafted onto otherwise contemporary computer systems, and used by unaugmented human beings to perform tasks familiar to most computer users.
Sources
Having decided on our topic and scope, the properties for review are three films and one episode of a TV show.
Jurassic Park, 1993
“I know this!” and Jurassic Park is so well known that I assume that you do too. We will look specifically at the 3D file system that is used by Lex in the control room to reactivate the park systems.
Disclosure, 1994
This film about corporate infighting includes a virtual reality system, complete with headset, glove, laser trackers, and walking surface, which is used solely to look for particular files.
Hackers, 1995
As mentioned in the introduction this film revolves around the hacking of a Gibson mainframe, which has a file system that is both physically three dimensional and represented on computer screens as a 3D browser.
All three of these films date from the 1990s, which seems to have been the high point for 3D file systems, both fictional and in real world research.
Community, season 6 episode 2, “Lawnmower Maintenance and Postnatal Care” (2016)
In this 21st century example, the Dean of a community college buys an elaborate virtual reality system. He spends some of his time within VR looking for, and then deleting, a particular file.
Clockwise from top left: Jurassic Park (1993), Disclosure (1994), Hackers (1995), Community (2016)
And one that almost made it
File browsing in two dimensions is so well established in general-purpose computer interfaces that the metaphor can be used in other contexts. In the first Iron Man film, at around 52 minutes, Tony Stark is designing his new suit with an advanced 3D CAD system that uses volumetric projection and full body gesture recognition and tracking. When Tony wants to delete a part (the head) from the design, he picks it up and throws it into a trashcan.
Tony deletes a component from the design by dropping it into a trashcan. Iron Man (2008)
I’m familiar with a number of 2D and 3D design and drawing applications and in all of them a deleted part quietly vanishes. There’s no more need for visual effects than when we press the delete key while typing.
In the film, though, it is not Tony who needs to know that the part has been deleted, but the audience. We’ve already seen the design rendering moved from one computer to another with an arm swing, so if the head disappeared in response to a gesture, that could have been interpreted as it being emailed or otherwise sent somewhere else. The trashcan image and action makes it clear to the audience what is happening.
So, that’s the set of examples we’ll be using across this series of posts. But before we get into the fiction, in the next post we need to talk about how this same thing is handled in the real world.
Whatever it is, it ain’t going to construct, observe, or repair itself. In addition to protection and provision, suits must facilitate the reason the wearer has dared to go out into space in the first place.
One of the most basic tasks of extravehicular activity (EVA) is controlling where the wearer is positioned in space. The survey shows several types of mechanisms for this. First, if your EVA never needs you to leave the surface of the spaceship, you can go with mountaineering gear or sticky feet. (Or sticky hands.) We can think of maneuvering through space as similar to piloting a craft, but the outputs and interfaces have to be made wearable, like wearable control panels. We might also expect to see some tunnel in the sky displays to help with navigation. We’d also want to see some AI safeguard features, to return the spacewalker to safety when things go awry. (Narrator: We don’t.)
Mountaineering gear
In Stowaway (2021) astronauts undertake unplanned EVAs with carabiners and gear akin to mountaineers use. This makes some sense, though even this equipment needs to be modified for use by astronauts’ thick gloves.
Stowaway (2021) Drs Kim and Levinson prepare to scale to the propellant tank.
Sticky feet (and hands)
Though it’s not extravehicular, I have to give a shout out to 2001: A Space Odyssey (1969), where we see a flight attendant manage their position in the microgravity with special shoes that adhere to the floor. It’s a lovely example of a competent Hand Wave. We don’t need to know how it works because it says, right there, “Grip shoes.” Done. Though props to the actress Heather Downham, who had to make up a funny walk to illustrate that it still isn’t like walking on earth.
With magnetic boots, seen in Destination Moon, the wearer simply walks around and manages the slight awkwardness of having to pull a foot up with extra force, and have it snap back down on its own.
Destination Moon (1950): Jim Barnes tests his magnetic boots.
Battlestar Galactica added magnetic handgrips to augment the control provided by magnetized boots. With them, Sergeant Mathias is able to crawl around the outside of an enemy vessel, inspecting it. While crawling, she holds grip bars mounted to circles that contain the magnets. A mechanism for turning the magnet off is not seen, but like these portable electric grabbers, it could be as simple as a thumb button.
Battlestar Galactica (2008, Season 4, Episode 5, “The Road Less Traveled”): Sergeant Erin Mathias inspects the damaged cylon raider for tracking devices.
Iron Man also had his Mark 50 suit form stabilizing suction cups before cutting a hole in the hull of the Q-Ship.
Avengers: Infinity War (2018)
In the electromagnetic version of boots, seen in Star Trek: First Contact, the wearer turns the magnets on with a control strapped to their thigh. Once on, the magnetization seems to be sensitive to the wearer’s walk, automatically lessening when the boot is lifted off. This gives the wearer something of a natural gait. The magnetism can be turned off again to be able to make microgravity maneuvers, such as dramatically leaping away from Borg minions.
Star Trek: First Contact (1996)
Star Trek: Discovery also included this technology, but with what appears to be a gestural activation and a cool glowing red dots on the sides and back of the heel. The back of each heel has a stack of red lights that count down to when they turn off, as, I guess, a warning to anyone around them that they’re about to be “air” borne.
Star Trek: Discovery (2017–) S01E01, “The Vulcan Hello” Ensign Burnham takes a step out onto the hull before lifting off toward the mysterious beacon.
Quick “gotcha” aside: neither Destination Moon nor Star Trek: First Contact bothers to explain how characters are meant to be able to kneel while wearing magnetized boots. Yet this very thing happens in both films.
Destination Moon (1950): Kneeling on the surface of the spaceship.
Star Trek: First Contact (1996): Worf rises from operating the maglock to defend himself.
Controlled Propellant
If your extravehicular task has you leaving the surface of the ship and moving around space, you likely need a controlled propellant. This is seen only a few times in the survey.
In the film Mission to Mars, the manned mobility unit, or MMU, seen in the film is based loosely on NASA’s MMU. A nice thing about the device is that unlike the other controlled propellant interfaces, we can actually see some of the interaction and not just the effect. The interfaces are subtly different in that the Mission to Mars spacewalkers travel forward and backward by angling the handgrips forward and backward rather than with a joystick on an armrest. This seems like a closer mapping, but also seems more prone to error by accidental touching or bumping into something.
Mission to Mars (2000): Woody inspects the craft for the leak.
The plus side is an interface that is much more cinegenic, where the audience is more clearly able to see the cause and effect of the spacewalker’s interactions with the device.
Mission to Mars (2000): The crew performs a daring EVA.
If you have propellent in a Moh’s 4 or 5 film, you might need to acknowledge that propellant is a limited resource. Over the course of the same (heartbreaking) scene shown above, we see an interface where one spacewalker monitors his fuel, and another where a spacewalker realizes that she has traveled as far as she can with her MMU and still return to safety.
Mission to Mars (2000): Woody sees that he’s out of fuel.
Mission to Mars (2000): Terry sees that she cannot continue her rescue of Woody.
For those wondering, Michael Burnham’s flight to the mysterious signal in that pilot uses propellant, but is managed and monitored by controllers on Discovery, so it makes sense that we don’t see any maneuvering interfaces for her. We could dive in and review the interfaces the bridge crew uses (and try to map that onto a spacesuit), but we only get snippets of these screens and see no controls.
Star Trek: Discovery (2017–) S01E01, “The Vulcan Hello”
Iron Man’s suits employ some Phlebotinum propellant that lasts for ever, can fit inside his tailored suit, and are powerful enough to achieve escape velocity.
Avengers: Infinity War (2018)
All-in-all, though sci-fi seems to understand the need for characters to move around in spacesuits, very little attention is given to the interfaces that enable it. The Mission to Mars MMU is the only one with explicit attention paid to it, and that’s quite derived from NASA models. It’s an opportunity for film makers should the needs of the plot allow, to give this topic some attention.
Spacesuits must support the biological functioning of the astronaut. There are probably damned fine psychological reasons to not show astronauts their own biometric data while on stressful extravehicular missions, but there is the issue of comfort. Even if temperature, pressure, humidity, and oxygen levels are kept within safe ranges by automatic features of the suit, there is still a need for comfort and control inside of that range. If the suit is to be warn a long time, there must be some accommodation for food, water, urination, and defecation. Additionally, the medical and psychological status of the wearer should be monitored to warn of stress states and emergencies.
Unfortunately, the survey doesn’t reveal any interfaces being used to control temperature, pressure, or oxygen levels. There are some for low oxygen level warnings and testing conditions outside the suit, but these are more outputs than interfaces where interactions take place.
Mission to Mars (2000), McConnell verifies that the team can remove their helmets.
There are also no nods to toilet necessities, though in fairness Hollywood eschews this topic a lot.
The one example of sustenance seen in the survey appears in Sunshine, we see Captain Kaneda take a sip from his drinking tube while performing a dangerous repair of the solar shields. This is the only food or drink seen in the survey, and it is a simple mechanical interface, held in place by material strength in such a way that he needs only to tilt his head to take a drink.
Sunshine (2007): Truman asks spacewalker Shelby to calm down since his vital signs are stressed.
Similarly, in Sunshine, when Capa and Kaneda perform EVA to repair broken solar shields, Cassie tells Capa to relax because he is using up too much oxygen. We see a brief view of her bank of screens that include his biometrics.
Sunshine (2007): Cassie warns Capa that he’s using a lot of oxygen.
Remote monitoring of people in spacesuits is common enough to be a trope, but has been discussed already in the Medical chapter in Make It So, for more on biometrics in sci-fi.
Crowe’s medical monitor in Aliens (1986).
There are some non-interface biological signals for observers. In the movie Alien, as the landing party investigates the xenomorph eggs, we can see that the suit outgases something like steam—slower than exhalations, but regular. Though not presented as such, the suit certainly confirms for any onlooker that the wearer is breathing and the suit functioning.
Alien (1979): The away team’s suit exhales as they walk the alien wreckage. Look for the white plume in the left image and a dark plume on the right.
Given that sci-fi technology glows, it is no surprise to see that lots and lots of spacesuits have glowing bits on the exterior. Though nothing yet in the survey tells us what these lights might be for, it stands to reason that one purpose might be as a simple and immediate line-of-sight status indicator. When things are glowing steadily, it means the life support functions are working smoothly. A blinking red alert on the surface of a spacesuit could draw attention to the individual with the problem, and make finding them easier.
The Fifth Element (1997): Mondoshawan spacesuits have tiny lights along the exterior.
Emergency deployment
One nifty thing that sci-fi can do (but we can’t yet in the real world) is deploy biology-protecting tech at the touch of a button. We see this in the Marvel Cinematic Universe with Starlord’s helmet.
Guardians of the Galaxy (2012) Peter Starlord puts his Helmet on Gamora in space outside the Kyln.
If such tech was available, you’d imagine that it would have some smart sensors to know when it must automatically deploy (sudden loss of oxygen or dangerous impurities in the air), but we don’t see it. But given this speculative tech, one can imagine it working for a whole spacesuit and not just a helmet. It might speed up scenes like this.
What do we see in the real world?
Are there real-world controls that sci-fi is missing? Let’s turn to NASA’s space suits to compare.
The Primary Life-Support System (PLSS) is the complex spacesuit subsystem that provides the life support to the astronaut, and biomedical telemetry back to control. Its main components are the closed-loop oxygen-ventilation system for cycling and recycling oxygen, the moisture (sweat and breath) removal system, and the feedwater system for cooling.
The only “biology” controls that the spacewalker has for these systems are a few on the Display and Control Module (DCM) on the front of the suit. They are the cooling control valve, the oxygen actuator slider, and the fan switch. Only the first is explicitly to control comfort. Other systems, such as pressure, are designed to maintain ideal conditions automatically. Other controls are used for contingency systems for when the automatic systems fail.
Hey, isn’t the text on this thing backwards? Yes, because astronauts can’t look down from inside their helmets, and must view these controls via a wrist mirror. More on this later.
The suit is insulated thoroughly enough that the astronaut’s own body heats the interior, even in complete shade. Because the astronaut’s body constantly adds heat, the suit must be cooled. To do this, the suit cycles water through a Liquid Cooling and Ventilation Garment, which has a fine network of tubes held closely to the astronaut’s skin. Water flows through these tubes and past a sublimator that cools the water with exposure to space. The astronaut can increase or decrease the speed of this flow and thereby the amount to which his body is cooled, by the cooling control valve, a recessed radial valve with fixed positions between 0 (the hottest) and 10 (the coolest), located on the front of the Display Control Module.
The spacewalker does not have EVA access to her biometric data. Sensors measure oxygen consumption and electrocardiograph data and broadcast it to the Mission Control surgeon, who monitors it on her behalf. So whatever the reason is, if it’s good enough for NASA, it’s good enough for the movies.
Back to sci-fi
So, we do see temperature and pressure controls on suits in the real world, which underscores their absence in sci-fi. But, if there hasn’t been any narrative or plot reason for such things to appear in a story, we should not expect them.
