Brain Upload

Once Johnny has installed his motion detector on the door, the brain upload can begin.

3. Building it

Johnny starts by opening his briefcase and removing various components, which he connects together into the complete upload system. Some of the parts are disguised, and the whole sequence is similar to an assassin in a thriller film assembling a gun out of harmless looking pieces.

jm-6-uploader-kit-a

It looks strange today to see a computer system with so many external devices connected by cables. We’ve become accustomed to one piece computing devices with integrated functionality, and keyboards, mice, cameras, printers, and headphones that connect wirelessly.

Cables and other connections are not always considered as interfaces, but “all parts of a thing which enable its use” is the definition according to Chris. In the early to mid 1990s most computer user were well aware of the potential for confusion and frustration in such interfaces. A personal computer could have connections to monitor, keyboard, mouse, modem, CD drive, and joystick – and every single device would use a different type of cable. USB, while not perfect, is one of the greatest ever improvements in user interfaces.

Why not go wireless? Wireless devices remove the need for a physical connection, but this means that anyone, not just you, could potentially connect. So instead of worrying about whether we have the right kind of cable, we now worry about the right kind of Bluetooth pairing and WiFi encryption password scheme. Mobile wireless devices also need their own batteries, which have to be charged. So wireless may seem visually cleaner, but comes with its own set of problems.

As of early 2016 we have two new standards, Lightning and USB-C, that are orientation-independent (only fifty years after audio cables), high bandwidth, and able to transmit power to peripherals as well. Perhaps by 2021 cables will have made a comeback as the usual way to connect devices.

2. Explaining it

Johnny explains the process to the scientists. He needs them to begin the upload by pushing a button, helpfully labelled “start”, on the gadget that resembles an optical disk drive. There’s a big red button as well, which is not explained but would make an excellent “cancel” button.

jm-6-uploader-kit-b

It would be simpler if Johnny just did this himself. But we will shortly discover that the upload process is apparently very painful. If Johnny had his hands near the system, he might involuntarily push another button or disturb a cable. So for them, having a single, easily differentiated button to press minimizes their chance of messing it up.

1. Making codes

He also sticks a small black disk on the hotel room’s silver remote control. The small disk is evidently is a wireless controller or camera of some kind. The scientists must watch the upload progress counter, and as it approaches the end, use this modified remote to grab three frames from the TV display, which will become the “access code” for the data. (More on this below.)

jm-6-uploader-kit-x

None of the buttons on this remote have markings or labels, but neither Johnny nor the scientist who will be using it are bothered. Perhaps this hotel chain tries to please every possible guest by not favouring any particular language? But even in that case, I’d expect there to be some kind of symbols on the buttons and a multilingual manual to explain the meaning of each. Maybe Johnny spends so much time in hotel suites that he has memorised the button layout?

Short of a mind reading remote that can translate any button press into “what the user intended”, I have to admit this is a terrible interface.

(There is a label on the black disk, but I have no idea what it means or even which script that is. Anyone?)

0. Go go go

Johnny plugs in his implant, puts on a headset with more cables, and bites down on a mouthguard. He’s ready.

jm-6-uploader-kit-d

The scientist pushes the start button and the upload begins. Johnny sees the data stream in his headset as a flood of graphics and text.

jm-7-uploading-a

Why does he need the headset when there is a direct cable connection to the implant? The movie doesn’t make it explicit. It could be related to the images used as the access code. (More on this below.) Perhaps the images need to be processed by the recipient’s own optic nerve system for more reliable storage?

Still, in the spirit of apologetics we should try to find a better explanation than “an opportunity for 1995 cutting edge computer generated graphics.” Perhaps it is a very flashy progress indicator? Older computer systems had blinking lights on disk drives to indicate activity, copied on some of today’s USB sticks. Current-day file upload or download GUIs have progress bars. As processing and graphics capabilities increase, it will be possible for software to display thumbnails or previews of the actual data being transferred without slowing down.

Unfortunately there is an argument against this, which is that the obvious upload progress indicator is a numeric display counting gigabytes down to zero, and it makes a fast chirping sound as a sonic indicator as well. The counter shows the data flowing at gigabytes per second, the entire upload lasting about a minute. There’s also the problem that it’s not Johnny who is interested in knowing whether the upload is scientific data rather than, say, a video collection; but the scientists, and they can’t see it.

jm-7-uploading-b

As the counter drops below one hundred, the scientist points the remote with black disk at the TV display, currently showing a cartoon, and presses the middle button. The image from the TV appears overlaid on the data stream to Johnny. This is a little odd, because Johnny assured the scientists that he wouldn’t know what the access codes were himself. Maybe these brief flashes are not enough time for him to remember these particular images among the gigabytes of visual content. But the way they’re shown to us, I’ll bet you can remember them when they come up again later in the plot.

jm-7-uploading-d

Two more images are grabbed before the counter stops. When the upload finishes, the three images are printed out. (In the original film this is shown upside down, so I have rotated the image.)

jm-7-uploading-f-rotated

Tagged

So what are the images for? The script isn’t clear. I suggest that the images are being used as the equivalents of very large random numbers for whatever cryptography scheme protects the data against unauthorised access. Some current day systems use the timing of key presses and mouse movements as a source of randomness because humans simply can’t move their fingers with microsecond precision. Here, the human element makes it impossible to predict exactly which frame is chosen.

Humans also find images much easier to recognise than hundred digit numbers. Anyone who has seen the printout will be able to say whether a particular image is part of the access code or not with a high degree of confidence. In computer systems today, Secure Shell, or ssh, is a widely used encrypted terminal program for secure access to servers. Recent versions of ssh have a ‘randomart’ capability which shows a small ASCII icon generated from the current cryptographic key to everyone who logs on. If this ASCII icon appears different, this alerts everyone that the server key has been changed.

There’s one potential usability problem with the whole “pick three random images” mechanism. The last frame was grabbed when the counter was very close to zero. What would have happened if he had been too slow and missed altogether? Wouldn’t it be more reliable to have the upload system automatically grab the images rather than rely on a human? Chris suggests that maybe it secretly did grab three images that could have used without human input, but privileged the human input since it was more reliably random.

Quick aside: You may be asking, if images would be so wonderful, why aren’t we using them in this way already? It’s because our current security systems need not just very large random numbers, but very large random numbers with particular mathematical properties such as being prime. But let’s cut Johnny Mnemonic some slack,  saying that by 2021 we may have new algorithms.

OK, back to the plot.

-1. Sharing the codes

The access codes are to be faxed from Beijing to Newark, although this gets interrupted by the Yakuza intruders. This is yet another device with unmarked buttons.

jm-7-uploading-g

This device makes the same beeps and screeches as a 1990s analog fax machine. Since we’ll later learn that all the fax messages and phone calls are stored digitally in cyberspace, this must be a skeuomorphism, the old familiar audio tones now serving just as progress indicators.

As with other audio output, the tones allow the user to know that the transmission is proceeding and when it ends without having to pay full attention to the device. On the other hand, there is potential for confusion here as the digital upload is (presumably) much faster. Most current day computer systems could upload three photos, even in high resolution, well before the sequence of tones would complete. Users would most likely wait longer than actually necessary before moving on to their next task.

-2. Washing up

During the upload Johnny clenches his fists and bites his mouthguard. When the upload finishes, he retreats to the bathroom in considerable pain. At one point blood flows from his nose, and he swipes his hand over the tap to wash it down the drain. The bathroom announces that the water temperature is 17 degrees. We’ll come back to this later.

jm-8-bathroom-tap

As Make It So emphasises in the chapter on brain interfaces, there is nothing in our current knowledge to suggest that writing or reading memories to or from a human brain would be painful. On the other hand, we know that information in the brain isthe shape of the neurons in the brain. Who knows what side effects will happen as those neurons are disconnected and reconnected as they need to be? We don’t know, so can’t really say whether it would hurt or not.

-3. Escaping the Yakuza

As mentioned in a prior post, while he is in the bathroom, the motion detector Johnny installed on the hotel door isn’t very effective and the Yakuza break in, kill everyone else, and acquire the second of the three access code images. Johnny escapes with the first image and flies to Newark, North America. 

Ford Explorer Status

image00

One computer in the control room is dedicated to showing the status of the Jeeps out on tour, and where they currently are on the island.

Next to the vehicle outline, we see the words “Vehicle Type: Ford Explorer” (thank you, product placement) along with “EXP” 4–7.  EXP 4 & 5 look unselected, but have green dots next to them, while EXP 6 & 7 look selected with red dots next to them.  No characters interact with this screen. Mr. Arnold does tap on it with a pen (to make a point though, not to interact with it).

On the right hand side of the screen also see a top-down view of the car with the electric track shown underneath, and little red arrows pointing forward.  Below the graphic are the words “13 mph”.  The most visible and obvious indicator on the screen is the headlights.  A large “Headlights On” indicator is at the top of the screen, with highlighted cones coming out of the Jeep where the headlights are on the car.

Jumbled Hierarchy

It is very difficult to tell from this page what the most important systems on the tour are.  The most space and visual weight is given to the car itself and its headlights, but we never see any data on the actual car itself. Did Hammond’s deal with Ford include constant advertising to his handful of tour monitors?

When Drs. Grant and Sattler leave the jeep to walk out into the park, we only see two doors open; but all four doors show as open on the main projector status display in the control room.

Probably an editing error, but with Nedry programming things, who can say?
They can remote control the steering wheel and pedals, but not the locks?Probably an editing error, but with Nedry programming things, who can say?
image01
Probably an editing error, but with Nedry programming things, who can say?

At best, the system is attempting to display a binary indicator (doors open/doors closed) based on limited data.  At worst, the system is unreliable and can’t be trusted to deliver even basic status information correctly.

We also see several buttons that are labeled like they should be active, such as “Hold”, “Quit”, “New”, and “Next”.  What could this mean?

  • Erroneous labels indicating action where there is none
  • Disabled buttons because they aren’t appropriate for where the Jeeps are right now
  • Something that could be active in the future when more coding is done.

Ideally, these are indicating normal actions that aren’t available right now.  The control team would still need to be trained on why they’re disabled and when they’re active, which puts an unnecessary burden on their memory. That information should be apparent in any interface on which lives depend.

Missing Information

Many systems appear to be missing from this display, or are indicated in a way that is too cryptic to easily identify:

  • Self driving features?
  • Basic diagnostic info, like oil temp and tire pressure?
  • What event is playing in the Jeep?

It could also be improved with information that the system can surely collect.  Trend information would be the most useful:

  • How efficiently is the Jeep moving?
  • Is it breaking down?
  • What kind of baseline is the tour establishing?
  • Has it been accelerating or decelerating? At what rate?

While you’re an unethical capitalist, there is even information that could be used to track the effectiveness of the park, by tracking the affective states of the passengers in the car:

  • Heartrate
  • Motion within the car (direction, proximity to windows)
  • Breath rate
  • Skin temperature
  • Conversation (and valence): words, pace, and pitch

At their least offensive, this data would be anonymously aggregated and analyzed by location to understand where the experience can be improved. Where is the tour the most boring? Most exciting? Where are the passengers most likely to view dinosaurs (and should have their attentions keyed)? You could also use the information in real time to know when there is likely a problem that needs the attentions of a remote operator.

Finally, if you have cameras on the vehicles, you have another data collection channel such that the system could compare views of the road and paddocks, compare to prior images, and know when there are plants that need trimming away from the rail, or when deformations appear in the walls and need attention from maintenance teams. Heck, you could turn those sensors into an upsell opportunity. Charge a few extra bucks, and friends and family back home can go on the live tour with you, or you could sell a 360° video back to the riders as a souvenir. Dinobucks to be made, here, people.

The Map

One of the few pieces that is straight-forward here is the map.  We see a dot for where the tour is, the number and type of vehicles on the tour, and location information.  For a single tour, this would probably work well.

It would be a nightmare with more than one tour out in the field at a time.

With more than a handful of vehicles out on the island with the current information density, the display would quickly be overwhelmed with numbers overlapping each other and changing constantly making it impossible to read.  The large vehicles too might overlap more important information, like the terrain or possible problem areas on the tracks.

Color contrast is an issue here too: green on green would be very difficult to see for anyone without perfect color vision.  If the color was the only thing to change on a change in status from good to bad, the color green is also an issue because of how many other colors it conflicts with.  Accessibility would be improved by choosing a color like blue, or adding an outline to the dot.

_Map
scifiinterfaces comp.

Better would be to have a basic icon for each tour on the map, and a basic color/label combination next to the icon to show a “nominal” status.  The icon could also be an indicator of how many vehicles were in the tour.

If something happened, the icon status could change to a different color, and the label would change to the new status. Icons would make that information more glanceable. Additional info would flow onto the screen for just that tour.

This would provide a clear indicator of which tour on the map to pay attention to, and what was broken about it.

Overall

I would hope that this wasn’t a screen that I would have to look at day-in, day out.  The entire future control crew of Jurassic Park was lucky they weren’t forced to deal with this. But, with a few tweaks to the map and a complete reorganization of the information on the Jeep Status screen, it might become usable.