Hacker News new | past | comments | ask | show | jobs | submit login
What is good and bad about the F-35 cockpit (hushkit.net)
149 points by edward 34 days ago | hide | past | favorite | 146 comments

You can learn quite a lot about aviation UI experience from civilian airliners' instruments. I did work a little bit with Boeing-777X ones.

It is not great, actually. PFD (primary flight displays) are cluttered and information-noisy. HUD is a much better tool for flying experience. Fortunately for civilian pilots, this is becoming more common now.

FMS (flight management system) has all the usability of IBM mainframes from 1960s and about the same performance.

The tasks that pilot have to do before the flight are rather simple: you have to input weather conditions, aircraft load and waypoints for autopilot. But with even the most modern FMS it is a tedious and frustrating process, you have non-intuitive control flow and non-qwerty keyboard. Also all text-based, non-graphical interface.

The better parts of aicraft UI are EICAS/ECAM (engine information and alerts) - they are both useful and intuitive to understand, with emphasis on graphical indication.

A lot of hard to use bits are not from any technology limitations - modern aircraft displays are rather capable, but from the decades of industry legacy and expensive certifications required for any change.

Ask any pilot what the most complex part is about switching aircrafts and getting a new type rating. It's never the way the plane handles, is always either remembering the mandatory systems knowledge or how to use the FMS.

I flew along on a small private jet recently, their system was 10x easier to understand and more capable than the typical airliner... So it is possible, probably even at a lower cost.

This seems to be an almost universal problem at old, large engineering companies. Somewhere along the way (probably as they slowly ossify into bureaucracies) all passion for innovation slowly withers.

I don't know what the answer is but attitudes generally flow down from the top, so I'd imagine a good start would be to somehow re-install creative technical people in the top-end leadership roles.

You're not wrong, but there's also an element of justifiable conservatism. Aviation killed a lot of people before the associated industry and community mostly figured out how to make sure they didn't. Trying something innovative sometimes introduces a safety issue that hadn't even been considered.

It's a consequence of development times. Large, complex projects take a lot of calendar time. It's generally more important to keep things from going awry than excelling.

Consequently, you look around and see nothing but conservative people putting in their time. If you're someone more driven, you get tired and leave.

I work in that industry, and I think it's very harsh to say that people have lost passion for innovation. Most engineers I've been around care about making a good product, and get excited about using better technology or finding a better design.

However there's a constant care about making things that are certifiable, and make business sense (otherwise it's just not going to happen). It seems to me that it's difficult for people outside this type of industry to appreciate how certification impacts the decisions.

We strive to make systems that are reliable, easy to use and maintain, and efficient; but they have to make their way to the customer, through the authorities and business constraints.

Bureaucratic inefficiency isn't the problem in this case. The people who make these systems aren't optimizing for UX, they're optimizing for safety and reliability.

It's the same reason banks are still running things like their 30yo COBOL code. The cost of replacing it isn't the problem. The problem is, if it's been running reliably for 30 years, then what justification could you possibly come up with for accepting the risk of replacing it?

> a good start would be to somehow re-install creative technical people in the top-end leadership roles.

Hell no. In aeronautics, you want safety-conscious, technically conservative, quality paranoid engineers. You need that because you don't want to retrain pilots on the last release every Patch Tuesday. A lot of the design that goes into the cockpit is to minimize training costs and that is a major procurement driver (remember the 737 MAX?)

Cockpit user interface tour of one of the smallest private jets, the Phenom 300:


I recently watched a few videos of flights in the Cirrus Visionjet, literally the smallest and lowest cost single engine private jet aircraft on the market. Its electronic cockpit looks very straightforward and logical in design.

The Vision Jet uses the Garmin G3000 which is an upgraded version of the G1000 which is quickly becoming the standard avionics for general aviation (new) airplanes. Incidentally the Phenom 300 uses the G3000 as well.

If you can fly a G1000 in a Cirrus piston or Cessna, the G3000 is an easy transition.

The avionics and FMS of a typical airliner is positively COBOL by comparison. To get type rated in modern GA jets is pretty easy because the avionics are generally what you are already used to. However, the G1000/G3000 aren’t part 25 certified, however there is now a G5000 that is Part 25 legal, so I expect there to be some considerable improvements trickling up into the bigger jets. (Part 25 is basically “transport category” aircraft.)

Systems in small aircraft are a lot simpler than in airliners, by nature.

Jets are actually in many ways simpler than a piston engine.

Mechanically simpler or not, I finished a Garmin install last week on a piston-engine Lancair. I would absolutely call the avionics and engine indicating systems intuitive and easy to use, even without being a pilot.

While it's no G1000 or G3000, the G500 TXi, GTN 750 Xi, and GI 275 are all very straightforward, with engine indications being displayed on the G500 Xi in an easy to navigate package.

At least as far what I've gathered reading airliner crash reports go (these obviously are going to be skewed) one of the challenges seems to be just being up to date / adjusting from one airliner to the next as far as procedure, and recovery processes for a given aircraft.

Being behind the aircraft as far as interpreting the situation/ indicators and not knowing what recovery steps to take seem to be a reoccurring problem, at least as far as catastrophic type situations.

As the scenario likelihood decreases, the probability of correctly identifying the associated recovery procedure probably also tends to decrease -- particularly for cases where recovery is not super clear and for cases where the scenario itself is memory intensive to describe or baroque -- "a occurred before b instead of b before a as normally expected and sensor d has condition c..."

I wonder if simulated error scenarios ever occur during flight -- could be a good way to expose pilots to rarely used recovery procedures within realistic day to day contexts ...

I think the tough part is trying to account for the human unpredictability in human errors.

You can have all the warnings and etc, but how the human processes and interprets and responds is just not always predictable. If they misread a situation, the time to recover/ self correct isn't often very long.

Granted, they've obviously made great progress to some extent, as safety as far as airliners go is pretty incredible these days.

are private jets under different regulations (rules)?

It depends on the weight, seat count and type of operation. Very generally speaking small jets with 9 seats or less and not being too heavy are CS-23 certified while larger (transport category) aircraft are CS-25.

So the small jet in my example indeed has less regulations than a larger airliner. And would probably not be allowed to be used for commercial air transport, only charter flights.

Yes. Both for pilots and the airplane.

There is also less concern about training costs.

You say information noisy, but some might WANT it to be information-noisy so that they don't have to press buttons to find what information they need.

I'm not saying they do (I'm a software developer Jim, not a pilot), just something to keep in mind. I'm rebuilding a webapp myself with more modern design, and in an early demo I've already gotten a remark about information density.

> Also all text-based, non-graphical interface.

I would consider this a positive. Graphics do not imply improved readability, the information is dense and you have no choice but to present it as concisely as possible. You also need a dim cockpit for low light operations.

What could be a win is integration with the QRG to make access to information easier during a abnormal operating condition. You spend precious seconds locating and flipping through paper books to identify mitigation procedures.

To elaborate on my point. Arcade flight and space simulators have probably the optimal possible UI for "aviate/navigate/communicate" activities.

Sort of like Nikon cameras or adobe photoshop. Their primary customer is folks who already own their stuff, so the UI is going to stay obscure.

It’s infuriating trying to use the touch screen in my wife’s minivan to change the radio station. I can’t imagine the frustration trying to use a touch screen to fly a fighter jet.

More buttons please!

Lots of general aviation cockpits are moving to touch screens too. For example Garmin just got the G3X Touch approved for certified aircraft: https://buy.garmin.com/en-US/US/p/682215/pn/G3X-TCERT-13

The G3000 integrated flight deck is pretty much the default in new business jets too.

In my opinion the touch screens are a lot easier to use than the old G1000 screens except for when you are getting tossed around in turbulence. Fortunately, there are still physical buttons for all of the "safety critical" or "time critical" inputs such as configuring the autopilot, changing radio frequencies, or adjusting the barometric pressure for the altimeter.

Source: I am a private pilot and have used the G1000, G300, and G3X touch.

I've always wanted to build a big rig of buttons to surround my pc monitor that I could program for various functions I do every day.

When working on Bioshock I had a strip of buttons that could be programmed to send keystrokes like it was a keyboard. I was working in Unreal Engine on the Xbox 360. I could plug these keys it into the dev kit, and with one click could enter all kinds of obscure console commands I could never remember. Was great.

Update: was one of these. https://xkeys.com/

My company builds and works with industrial automation, custom CNC machines, and industrial robots. I've observed a shift from purely button-and-neon operated panels from early PLC or relay logic machines, to multifunction keys on the human-machine interface display (HMI) before touchscreens, to a fitful few years when people thought it was a good idea to build machines with VB6 and various serial to digital IO adapters, to purely touchscreen-driven machines with one legally-mandated physical emergency stop button (about when I arrived in the industry, the other machines were mostly historic beasts I've occasionally been charged with maintaining), and back towards multifunction keys. There are usually a few dedicated buttons and indicators for common operations (reset, cycle start, control power on, feed hold, feed rate override, etc), that still makes sense most of the time.

I really think multifunction keys are the best of both worlds. As the author of this piece describes, the multifunction display with 20 keys around the outside (or, for CNCs, 10 keys across the bottom of the monitor and 10 off to the right, in a 1-4-4-1 spacing so you can feel exactly which button you're hovering over while you're staring unblinking at a cutter chewing through 5-figure assemblies) is a good compromise. It takes some serious concentrated planning to design a set of keys that are intuitive (top to scroll up, bottom to scroll down, one dedicated for enter, two for context-specific operations, etc), but it gives you the freedom to design relatively shallow but featureful menu systems that you can memorize and get tactile feedback to operate with confidence.

Can anyone recommend reading about UX design for these kinds of combined digital/physical interfaces.

There is a Military Specification MIL-STD-1472H http://everyspec.com/MIL-STD/MIL-STD-1400-1499/MIL-STD-1472H...

I have a StreamDeck[1] that I use for various random things. There's an open source driver for it too that you can drive from Python[2], letting you run Python code when a button is pressed. You can do some cool stuff with it (eg I've used mine to make HTTP requests on button press)

[1] https://www.elgato.com/en/gaming/stream-deck

[2] https://github.com/abcminiuser/python-elgato-streamdeck

Ah, yes - writing our comments in parallel. :) The SD is great. Thanks for the link to Python code, I'll check it out.

Haha, yes, indeed! The python library is quite easy to use and lets you go beyond what the official application allows (plus, I use it on Linux).

I actually looked at those XKeys ones first, but IIRC they were a little too expensive for what I wanted (yes, the StreamDeck isn't that cheap either, but its easier to use out of the box)

StreamDeck is great for non-streaming tasks. I'm very happy with it.

XKeys looks nice!

I've got a StreamDeck XL mounted under my monitor, in a button box; its capabilities as a macro machine are vastly underrated, although a few people are catching on that it's not just for live content creation.

It's absolutely amazing when doing remote work to be able to fire up OBS Studio, the meeting join, microphone control etc in a single click - and I have that just on one page, along with other meeting controls such as camera scenes. Another for VS Studio, there's integration with that which is quite useful to send commands directly. I have all my most of ten work websites on buttons of their own.

As mentioned, I also have button boxes however I'm still working out what the best way of sending commands would be; Joy2Key seem to mess up alt keys sometimes, making it sticky. When it worked it was nice, imagine adjusting text size just twisting a rotary for example.

The HP-300 computer had buttons on the right side of the screen.


This is why the macbook touchbar is so awesome. I know it got a lot of hate when it first launched, but I use it all the time and have with all sorts of customizations.

The touchbar doesn't have physical buttons.

Physical buttons are critical. That's the key difference between an MFD and a touchscreen. The touchbar is a one-dimensional touchscreen, with most of its downsides.

This isn't to say that displays on the buttons themselves are bad. But we pay way too little attention to the ergonomics of fine motor control and tactile/muscle memory. If you look at the article linked in this post, the key point being made is how hard it is to hunt for a featureless touchscreen button while your wrist is in the air being jostled around by turbulence. You don't have to be a pilot to appreciate interfaces that go easy on your fine motor skills though. Any action that requires extended use of fine motor skill (like holding your wrist in the air while hitting a tiny spot) will be at minimum tiring (both mentally and physically), and maybe impossible for children and the elderly.

An MFD with a wrist rest, uneven button spacing/sizing/tactile patterns, rotary controls in the corners, and maybe button LEDs to color-code functions in addition to on-screen labels, would be superior to the touchbar. Actually, the numpad on "full-size" keyboards is tantalizingly close to that in some ways - it would be great if someone took a full-size keyboard and replaced the numpad with something that sort of looks like it but is functionally an MFD.

By the way, all keyboards already have tactile pattern aids - the little bumps on the F and J keys that let your fingers quickly know if they're correctly positioned on the home row. Everyone knows that, but it's just a glimpse of what's possible if we invest more in tactile interfaces.

I don't think people hate the touchbar per se, but the fact that they replaced a genuine row of keys to fit it in.

If it had been in addition to a normal keyboard layout it would have been far, far better - in my opinion, speaking as a PC Streamdeck user.

I wouldn’t be surprised to hear that it was originally designed to live in the bottom bezel of the screen, but that got cut due to aesthetic concerns. Seems like a more natural functional fit.

Sounds interesting! Could you give some examples on how you use it? E.g. how did you create the shortcuts, how many do you use, how do you display them or roughly what tasks are you accomplishing?

> I've always wanted to build a big rig of buttons to surround my pc monitor that I could program for various functions I do every day.

I think it's called a keyboard ;)

> I can’t imagine the frustration trying to use a touch screen to fly a fighter jet.

A war plane is something you have to operate while it's burning, or while you're bleeding on it, or while you can't see properly because it's full of smoke or someone just blinded you with a laser. The adoption of touch screens in this sort of cockpit seems misguided. Particularly for anything related to controlling comms or navigation.

That’s really not how we design modern jet fighters. Air combat involves a huge number of tradeoffs and ejecting is now the correct response to a wide range of issues. For example, the F-35 so engine so engine failure is likely to result in a lost aircraft.

It’s basically been decided that we are going to spend silly money keeping a small number of absolutely cutting edge aircraft flying rather than thousands if not tens of thousands of of likely more efficient but less capable possibly drone aircraft.

PS: To be clear it’s possible their making the correct choice. I personally doubt it, but I don’t have access to the kind of classified documents to justify things in one way or another. An effective labor weapon for example might render vastly cheaper drone fleets ineffective.

No, the post above is exactly correct.

Ejecting from an aircraft is an absolute last resort, and you stand a good chance of dying or being badly injured if you have to do it, and even if you survive the ejection, the parachute ride, and the landing, it's a better than even chance you're going to be captured and beaten, tortured, or killed on the ground. That's if you land on land of course -- if you land in the water you might just drown or never be found.

> it’s better than even chance you’re going to be captured and beaten, tortured, or killed on the ground.

That’s simply not the reality of modern fighter jets. The vast majority of time is going to be spent flying over either open ocean or friendly territory. That’s been true even if you’re only looking at recent combat missions.

Anyway, single engine aircraft are always going to be at higher risk for mechanical failure, yet that’s the chosen design. Just look through this list: https://en.wikipedia.org/wiki/List_of_accidents_and_incident.... In commercial aviation 4 crashes per year worldwide not that great, in military aviation 4 crashes a month worldwide is a very good month.

It's exactly the reality of modern fighter jets.

You're describing training flights, not combat missions. It is very common for aircraft to fly missions that are hundreds of miles long, most of which is over enemy territory. I used to sit in CDC on a CVN and watch the missions every day during Operation Southern Watch, which wasn't even really war -- but those pilots were flying over enemy territory for many hours every day.

Same thing in Afghanistan and Iraq, every flight was essentially over enemy territory for its entire duration -- surveillance, strike, and transport. You really don't want to bail out of an aircraft in the tribal areas if there's any way to get it home.

Modern aircraft are extremely reliable, they don't break often. Modern ejection seats are extremely reliable, they don't fail often. And yet, everything I posted above is exactly correct.

The essentially over enemy territory and wasn’t really a war are important qualifications. Compare the odds of a downed pilot avoiding capture in a WWII bombing mission over Germany with the majority of missions over Afghanistan and it’s clear they where contested territory. Still high risk, but also a very different situation.

Close Air Support is definitely a major role for the F-35, but you have troupes there so it’s contested territory. Patrolling no fly zones is currently much better suited for drone aircraft, but might become important again.

So yes exceptions exist, but you really don’t want to run air to air combat missions out of air force bases anywhere near contested airspace. The risk vs rewards very much favor longer flights with mid air refueling.

You can't patrol a no fly zone with a drone. Drones don't do air to air combat, the only real use of drones at this time is ground surveillance.

That may change some day in the future, but that's the reality now.

And the F35 can do CAS, sort of, but that's not really its role. We've got a whole host of fourth gen aircraft to do CAS as CAS is nearly always done in an environment where we have full air superiority.

Adding sensors to a drone isn’t a major issue. https://www.defenseworld.net/news/28773/GA_ASI_Integrates_Lo...

That said, you’re right, F-35’s have already been delivered and represent actual capacity not hypothetical alternatives.

However, my personal experience was mostly on the R&D side of the DoD which colors my thinking about this stuff.

Is this due to some fundamental change in warfare or just due to the fact that modern wars are typically a large country fighting a comparatively tiny and unequipped militia? Like, would it still be true if China and the U.S. actually went toe-to-toe?

Both and the relative rarity of armed conflict.

High aircraft speeds combined with mid air refueling give nearly arbitrary ranges, add in and the risk associated with air defenses makes long distance attacks more common. You don’t want your air bases within easy striking distance of the enemy.

Several other factors are relevant for example prop aircraft can handle rough runways, but jets need a clean surface as they can suck debris into engines. So, even in the event of a US / China war, it’s not clear how long both sides would be capable of having effective air forces anywhere near each other.

> That’s simply not the reality of modern fighter jets.

Even Iraq managed to capture US and allied pilots. I think I'll forego the rest; you're reality and actual reality are a bit too divergent.


> it’s better than even chance you’re going to be captured and beaten, tortured, or killed on the ground

Is demonstrably false. War is still an extremely unpleasant business, but the term acceptable losses applies even in peacetime.

Every mission training or otherwise is rolling the dice that the aircraft will be lost. Yet, they still send pilots up knowing those risks in preparation for war. Aircraft designs really do involve similar tradeoffs and injured, dead, or captured pilots really are part of the equation.

If you eject over enemy territory, you are very likely to be captured and tortured or/and killed.

That's a fact. And nothing else you've posted argues against it.

Of course people take risks, mitigate risks, and weigh those risks against the benefits from the missions.

That doesn't change anything or falsify the statement I made.

That’s really not what I was arguing about, if over the 50+ year lifespan of an aircraft it spends less than 10% of total flight time over enemy territory and ejecting over enemy territory has a 99% odds of capture.

Then overall odds of ejecting resulting in capture is way less than 50%. Further, if mechanical failure is expected to bring down say 15% of aircraft, then the odds of mechanical failure over enemy territory is low vs actually being shot down.

That said, actual odds are a little different as mechanical failure isn’t evenly distributed, and hours spent over disputed territory also risk capture etc. Still it’s clear that a low percentage of ejecting pilots are expected to be captured.

A pilot who ejects over enemy territory is almost certain to be captured. Obviously, a pilot who ejects on a training flight over Tonopah isn't going to get captured, and thus my statement about a pilot being captured would not apply to such a training flight where there is zero risk of capture. The numbers you've made up mean nothing, even if they were correct, which they aren't. Nearly no aircraft have a 50 year lifespan, and any aircraft shot down over enemy territory is going to end up with the pilot almost certainly either killed or captured, just as I said.

You're adding in all kinds of things that have no bearing on the discussion and don't mean anything in the context of the discussion. And the numbers you made up are silly.

> Nearly no aircraft have a 50 year lifespan.

Individual fighters don’t, but widely adopted designs from the 70’s and after stuck around. The F-16 family is currently expected to be in service for 75 years though it’s only reached 42 years at this point. F-15’s have done 45 years at this point and “We intend to maintain the F-15C fleet through the 2020s.” Suggesting 55+ with a fair amount of uncertainty. F-14 was introduced in 1974 and ironically enough is still in use by Iran. Hell the F-4 entered service in 1961, and is still used by Japan.

As to what you said, nearly certainly killed or captured isn’t it’s better than even chance. You can quibble about meaningless numbers chosen to convey meaning not specifics, but as your making a different statement clearly you agree a small percentage of total elections are expected to result in capture. Death and serious injury is a much greater risk, but again overall risks for the lifespan of the design is what’s important, the US military will sacrifice far more lives than the a most few hundred F-35 pilots we are talking about. As demonstrated by the decision to double down on manned aircraft in the first place rather than try and make a air to air combat capable drone.

You're just moving goalposts all over the place.

The F14 is not in service anymore. Iran has probably less than 10 that are actually flyable, and they've been so heavily modified as to be mostly not recognizable as F14s anymore. The only other actual airframes that are anywhere near a 50 year lifespan are the B-52s, and they've been totally redesigned and rebuilt multiple times over those years.

There are no F15 airframes over 45 years, there are no F4 airframes over 45 years. The F15 production line is still open, there are 7 F4s still in service and they are a heavily upgraded version that wasn't produced in 1984 -- and those 7 leave service next month. You're confusing first date of production with the actual production dates of the aircraft still flying in order to make up a point that isn't even accurate as you've stated it.

Anyway, you just keep making up stuff so I'm going to leave this discussion here for now.

I mentioned that in my post, Individual fighters don’t

However, again for any future readers it’s irrelevant how long an individual aircraft is in service from a design perspective. What’s relevant is the what people using that design use it for, and this how long. Bullets for example are single use. However, individual designs can be used for very long periods of time as long as their serial useful, specific aircraft designs for the same criteria where you can see some of them manufactured for 20+ years.

As the other point was dropped I will take this as you conceding the argument. ;-)

No, I'm not "conceding" anything. Your goalpost moving straw man arguments from a position of total ignorance of the subject aren't worth addressing anymore.

Good day.

> Even Iraq managed to capture US and allied pilots

Uh-Huh. Thirty (repeat 3-0) years ago.

The situation can escalate very quickly in a fighter jet so that leaving the last resort to late could be fatal. Encouraging ejection may be a good idea when the pilots natural inclination is always to stay with the plane.

> It’s basically been decided that we are going to spend silly money keeping a small number of absolutely cutting edge aircraft flying rather than thousands if not tens of thousands of of likely more efficient but less capable possibly drone aircraft.

What is the basis for saying that? From what I understand, the U.S. military is investing very heavily in unmanned aerial vehicles, including having F-35's control large numbers of them.

I meant in terms of opportunity costs. Spending hundreds of billions on F-35’s is money they can’t spend on something else.

It's the drone boogeyman again. They are cheap because I can buy them in a brick and mortar store. Surely the military can just buy tens of thousands of them.

>It’s basically been decided that we are going to spend silly money keeping a small number of absolutely cutting edge aircraft flying rather than thousands if not tens of thousands of of likely more efficient but less capable possibly drone aircraft.

Well, you're missing the fact that there will be thousands of F-35s on earth (500 already built). Meanwhile there are less drones like MQ-9 Reaper (200 built) which are basically primarily to bully some ground troops than F-35s despite them only costing $40 million a pop. After all they are cheap. Because you see, the brick and mortar store sells them, so they must be cheap even when they are not.

>than thousands if not tens of thousands of of likely more efficient but less capable possibly drone aircraft.

In a war less capability might mean the total loss of all aircraft. You only need a few F-35s and some anti air support on the ground and the drones will be shot down. Yes, you might be able to execute a saturation attack but once you lost your entire fleet how are you going to respond to the counterattack? Oh right, with your own 5th generation fighter that matches the F-35.

Honestly, the biggest benefit of drones is that you can equip your infantry with loitering munitions and thereby reduce the need for conventional CAS provided by a conventional air force. This isn't a stupid lets build ten million drones swarm buzzword bingo to scare some people who barely know anything about the capability of military weapons. That's nonsensical. It's like plastering machine guns everywhere because "cheap". As if militaries only care about stockpiling weapons with no logic behind mission objectives.

Yep, that 40 million really is cheap. An apples to apples comparison puts 200 MQ-9 for 8 billion vs 600 F-35’s for 400 billion. So about 40 Million vs 667 Million.

But that’s just the start, the MQ-9 costs about $3,600 per flight hour vs ~35,000 per hour for the F-35. Though that slightly favors the F-35 because training an F-35A pilot costs a little over 10 million dollars.

That said, the MQ-9 is designed for a very different role so per aircraft numbers are deceptive. F-35’s have a much higher payload, higher top speed, stealth etc.

> Meanwhile there are less drones like MQ-9 Reaper (200 built) [...] costing $40 million a pop.

Prices like this for military equipment are deceptive. They're not $40mil a pop. They're the product of an $8 billion project. The build price is probably less than $100k... or at least it would be if a military contractor wasn't absolutely creaming it.

Glancing at the specs range 1,850 km, top speed: 482 km/h

I don't think there are aircraft of any kind doing anything like that for $100k

Something like a TBM 910 has comparable specs as a light aircraft and cost something like $4m.

There are lots of buttons and switches on the throttle and stick. Most modern western fighters are designed for HOTAS (hands on throttle and stick) operation, where the pilot never takes their hands off the throttle and stick during combat.

Not all touch screens are created equal. The one in my Tesla is as good as a phone or Ipad

> Something that has been lost in all glass cockpits is the tactile feel of pressing buttons and knowing you got a response – I found you could enter Lat/Longs by feel whilst looking out the window. This is something you definitely can’t do ‘on the glass’ on current jets.

I also really miss this from the dumbphone era. I used to be able to type SMS texts by feel using a T9 keypad while walking down the street, and only occasionally have to check the screen to see what I wrote. Smartphone QWERTY makes this impossible.

Same for cars interfaces. My 17-years old BMW has 0 big touchscreens, for everything there is a manual knob/button. I've found out I don't need more. I tweak most of stuff I need like AC and radio while driving without losing contact with the road even for a split of second.

I know from my experience that I took my sight off the road, there could be an atomic blast in front of me and I would hardly notice it. Somebody slamming brakes in front of me might be peripherally noticed but without context could cause a stupid and dangerous reaction.

When Tesla started basically putting tablets instead of knobs some folks cheered. I couldn't be more sad for this trend. Clearly worse direction for those like me.

Or even worse: little touchscreens that resemble a physical knob/button, even with haptic feedback!

All physical buttons/switches in my 20 year old Mitsubishi still work. I have a feeling the newfangled, fancy LCD-switches don't last that long, and will be very expensive to fix when they break.

> Smartphone QWERTY makes this impossible

I can absolutely type in my iPhone without looking at it. I just typed this comment with zero visual contact with my phone and no edits.

(EDIT: this is a weirdly divisive comment.)

Good for you! That’s an awesome skill. The point is, tactile feedback is essential for doing that to most of us, not that 0 people can do it.

> not that 0 people can do it

Everyone in my (much) younger sister’s age group can do this. They grew up on touch devices. You and I didn’t.

Just a note that one group’s personal experiences shouldn’t be generalised to the point of describing others’ as “impossible.”

I agree that we should be careful with personal anecdotes, that's why I did not comment earlier.

But I can say that personally I was also able to do it without looking to some extent (perhaps a whole sentence with very low issues to correct), and I'm slowly losing that ability (which was accentuated with a change of hardware, but I was also feeling the loss of accuracy beforehand).

Now, this gets weird. I used to be able to do this on my iPhone 7. Now on the iPhone X I am making typos constantly...

Same. I think those of us who learned to touch type on post-T9 devices find the smaller phones’ keyboards more familiar. Contrast that with twentysomethings I know, and the disparity disappears.

M testing this out now, let’s see if I can dk it

Edit: the answer is no

I would agree with you from a personal perspective but my partner frequently does this on her Android. I personally find it creepy of her looking at me while obviously doing something else.

I think I will never really get the hang of the mobile keyboard(though I tend to use touch swiping), because I just consider it inferior to any physical normal sized keyboard.

> At present I am pressing the wrong part of the screen about 20% of the time in flight due to either mis-identification, or more commonly by my finger getting jostled around in turbulence or under G.

That seems... bad, but also totally unsurprising.

I'm surprised there isn't more hand rests used in these applications, similar to how the Dragon capsule has those finger "shelves" for the touch panels. Would make accurate touch input at least slightly easier

Hand rests are only half the problem--even with a totally steady finger, it's tough to consistently find a small area on a touchscreen (and that's even ignoring the level of customization the F-35 allows) without looking at it. The author wasn't joking when he talked about inputting coordinates without looking at the control panel, and I expect it to take some massive advancements in haptics before we can do anything similar with touchscreens.

Yeah, this seems like a fairly easy workaround. A possible limitation is the cramped cockpit may literally not have enough room, your knees might hit the rests when working the rudder.

Maybe if it were a flip-down thing where you could rest your hand when needed, but also bounce it out of the way when its not needed.

Some general aviation planes have this, I've only used one during a short test flight in a simulator but it seemed to work quite well.

The author adds

> think how much easier it is to type on a smartphone with your thumbs versus trying to stab at a virtual keyboard on a large tablet with just your index finger.

My experience typing on a touch screen on bumpy roads or even a seemingly smooth highway at normal speed (as passenger of course) is that I miss a lot of keys, even if I grab my phone with two hands and I type with thumbs, which are a little too wide. Luckily I learned to swype and that works much better because of the constant contact with the screen. Fingers can't jump away. Still, getting shifted characters right is a problem. A physical keyboard is much better in that environment.


This is way more sparse and streamlined than I thought. Wonder how big the manual for flight simulator game would be. Some video of UI in action on simulator.


Are there any design documentations for UX/UI for military hardware? Is there a name for this style. Is it actually as functional as it looks? I wonder how these designers feel about dealing with military powerpoints.

As an intern 25+ years ago I had an assignment to draw up mocks for a new UH-1N interface based on pilot feedback. There were a lot of rules (I think all under the umbrella of "MIL-SPEC"), but I think the one that sunk in the most was that the design had to work in monochrome. Depending on shades and colors breaks the design for the color-blind but also ruins interactions in bad lighting conditions, wearing helmets, etc.

It was a bigger, bolder look than was common on desktop computers, and I've always personally loved the style. There was a brief rise of similar rules in the civilian world as things moved mobile and we had to work on small screens with big fingers, but the direction now is to try to remove anything that might hint of interaction from the screen entirely. (Not a fan of this myself!)

I've been in civilian work pretty much ever since so I have no idea how things have changed since then. Edit: on second glance I think a lot of the same principles are still in effect. While colorful I'm not sure that anything depends on reading the color, and seen from the pilot's chair it probably isn't all that crowded.

tangentially, your point about meaningful shades/colors breaking the design for non-color-blind people situationally is a great example of the need to always keep accessibility in mind

There are many design guidelines and studies for A&D (aerospace and defense); everything from which typeface to use, to how to display quantities and directions. I am not away of any comprehensive style guide or all-encompassing name for it.

> Wonder how big the manual for flight simulator game would be. One manual for the F-14B is about 1k pages, so I would assume this would clock in significantly more than that due to the added capabilities (in information at least, possibly not pages due to the F-14 being pre-HOTAS and things like that http://server.3rd-wing.net/public/Ked/natops%20F14B.pdf

Looking at the picture, that's surprisingly sparse.

Wow, how did the fact that the G-forces will make it difficult to use voice commands or to interact with a touch screen not kill these F35 cockpit features a long time ago?

To be really specific for those not aware, pilots need to do special breathing routines when under high g-forces.

Deep, rapid breaths. They need to suck oxygen into their body as quickly and efficiently as possible to maintain consciousness, because the g-forces forces blood away from their heads and towards their legs and feet.

Here's a "classic" video of a pilot successfully evading an unbelievable number of SAMs over Iraq. The rapid breathing might make you think he's panicking. Nope, he's got it 100% under control and is following training perfectly. Though I'm sure he certainly needed a drink after making it home.


There's also sort of a method of flexing their thighs that they learn. Restricts blood flow to the legs, so there's more blood for the rest of the body. That's a bit of an athletic endeavor in and of itself. Try flexing your thighs... now hold them that way for 5, 10, 20 minutes. Yikes.

Anyway, how the hell could you even bark out voice commands while doing that sort of breathing!?!?

These voice commands are usually just for changing radio settings and other non-essential functionality. Stuff that would be just as difficult to change with tactile switches in high G. The Eurofighter has been using a similar system since the 90's

First, I'm curious if they use a throat mike or something bone conducting.

Second, given how shitty voice interfaces are, I'm shocked they would use them at all. I assume the military funds a lot of EEG research- that would be ideal here.

According to the article the author knows of no pilots that use the voice interface.

Given how finicky they are I can't say I'm surprised. Maybe with modern pseudo-AI systems they could be reliable enough to depend on, but none of that is in military hardware specced out 20 years ago.

If you used the nato phonetic alphabet maybe, ie shortcodes for voice???

It would probably help, but there is still the enormous noise floor to contend with. Jet Fighters are unbelievably loud. I mean that literally. If you've never watched one take off you would think I'm exaggerating the noise level. It's hard to describe because most people have never experienced something that loud.

Not sure how much I'd trust my brain for this kind of thing... Thinking "I better not press that eject button!" might be problematic

Ejection isn't on the screen it's the big yellow loop between their legs.


>I'm curious if they use a throat mike or something bone conducting.

No, and no. Just good, old-fashioned microphones and speakers. If you're lucky, you have a set of "communication ear plugs" wired up to your helmet so the speakers in your earcups aren't trying to compete with engine noise and your ear plugs.

EEG requires highly stable signals. G forces would make any EEG measurements extremely unreliable.

I only know some standard "anti-G" actions involve breathing really hard while keeping your lungs inflated. I guess that makes speaking more difficult?

I play a lot of Digital Combat Simulator, which models a lot of these planes. At first I was interested in shooting things, then I was interested in the flying, but in the end I learned I was most interested in the computer systems and how these games are mostly virtual simulations of old computers. It's fun to see what changed over the years and between cultures. The MiGs are completely different than their Western counterparts. How you navigate from point to point is often completely different. It's also interesting how durable the UX of MFDs were in cockpit design. When I went for my GA license, it was funny to see how many newer models of planes had moved to "glass" cockpits. It some ways, it felt like putting wifi on your fridge. Most of the GA planes folks fly were made in the 70s.

Agreed! If the information in this article is even slightly interesting to you, then you owe it to yourself to spend some time with Digital Combat Simulator. In there you can get first-hand experience with the Harrier and the F/A 18 discussed in the article. You will spend a lot of time with the cockpits and start to develop your own affinity for things in the cockpit, and develop your own ideas about what you like and what you don't.

Do a Google Images search for "DCS FA18 cockpit" or "DCS Harrier cockpit" and you'll see that the real life photos are pretty much indistinguishable from the screenshots, if you take the word "DCS" out of the search term. Every switch and gauge are faithfully simulated, and you can use the real-life operating manuals to operate these simulators.

There's a lot of overlap with a lot of the things I love about computer systems. It's fun to learn, and to operate systems, and to feel that feeling of mastery as you gain confidence in making the machines do your bidding. Those types of feelings are much the same (and different) between operating computers and operating aircraft systems. To a large extent these modern planes feels a lot like operating a big flying computer, at least in the simulators. This is also true of things like the Cessna G1000 glass cockpit in Flight Sim 2020.

> DCS FA18 cockpit

One thing that's particularly impressive about the F/A-18 is that the Spec available in DCS (give or take a few features) is almost the same as the first batch from the 80s UI wise.

In 1983, serving alongside the almost completely analog/60s F-14, the F/A-18 had 3 pretty modern MFDs. Definitely more advanced than most Star Wars cockpits, for example.

"Star wars cockpits" is an interesting comparison.

I wonder at what point the Star Wars people will thematically have to update their cockpits/other tech because it will otherwise date.

Kind of how CyberPunk has dated because they missed out on smartphones...

I'm flabbergasted that MFDs haven't made more inroads in automotive and computer input device applications. We even have decent HUDs in cars now, but no MFDs. I'm convinced a good MFD would be far superior to the typical touchscreen crap - if Tesla had a MFD and a HUD instead of the giant TV screen, I would have bought one by now.

I think it's because the MFD model is designed for trained pilots who have to take a test to use them, i.e. I'd love one, but I can just imagine the complaints about the different pages and things like that.

Also, I'd love one for home automation too. There's only so much data you need to display or enter so the reduced latency of a dumb screen going straight into the back of a internet connected SBC (or even microcontroller) could be really nice to have in the kitchen, rather than having to fiddle around with either just a phone or what would end up being a tablet bolted to a wall.

It was like that in mid 2000s Lexus as well. I think it's a great system... less fingerprints to clean off the panel.

The head-unit in my Ford Focus (2004-ish?) had an MFD. Parallax was a bit of an issue (which buttons were for which function looked different from passenger and driver seat). I rate it as just "okay." The one place it was clearly better than touchscreen was for selecting the radio-station from a list of "favorites." Each station was always on the same button, so switching was simple.

Note that this was for controlling the audio system only; this predates integrated "infotainment" center stacks.

He kinda buried the lede:

> Having bashed the interface, the way this jet displays information to you is incredible. The sheer amount of situational awareness I gain from this aircraft and its displays is like nothing I’ve experienced before. The off-boresight helmet is much more accurate than legacy JHMCS systems and I find it clearer to read (although I still want a wide-angle HUD for flight and fight-critical data!). About the only thing missing from the whole cockpit is the lack of ‘feel’.

The prevailing narrative on HN and elsewhere (that seems to be a US DoD disinfo campaign) is that the JSF is an overpriced, overcomplicated piece of junk. Then there are these murmurs (largely from actual operators) leaking the fact that the aircraft is actually "incredible."

I upgraded my car's stereo and it went from a knob for volume control to two small buttons. Hate it.

It was difficult enough to find stereo with dedicated volume buttons. The rest of the functionality is via touch. Not a fan. I've never driven a BWM, but I believe they have a knob beside the shifter that allows you to go through the menus.. that feels like a much better system.

They do, and also a physical volume knob that you can find and turn without looking at it.

When I had a rental Renault that had everything on a touchscreen I thought "cool, looks modern!" when I looked at it stationary. And then totally hated it after 5 minutes of driving because it is so much worse to use on the road than the BMW system with actual buttons and the controller you can use blindly.

Renault cars have audio controls behind the right hand side of the steering wheel, alongside the touchscreen. It works fine and you keep your hands on the wheel.

Not sure this one had, it was a very basic clio.

With another user interface decision that baffled me... The button to enable cruise control was under the central armrest, and if pressed the wrong way would enable a speed limit instead of cruise. Never seen another brand put cruise control under your right elbow instead of somewhere on the wheel.

Aah the French and their quirky sense of design

My 2018 Mazda has a touchscreen, but 100% of the functionality can be done with physical buttons. It's the best interface for a car center-stack I've used on any car in the past 15 years or so.

I love the BMW system. The knob is very tactile and its always clear what exactly you're doing. The control is right by your hand, and you don't have to reach for anything. I've driven the Toyotas which have a touchscreen, and it feels much more vague, and often takes me a couple taps to get the desired outcome.

> two MFDs with the classic 20 pushbuttons around the outside

MFD = multi-functional display


There's always acronyms in these sort of posts that I get to learn. And probably relevant to this:

> Something that has been lost in all glass cockpits is the tactile feel of pressing buttons and knowing you got a response

As I'm assuming they all went touch screen

This particular acronym is also used a lot in the marine industry. Marine GPS, also known as "chartplotters", are mostly called MFDs nowadays, as it's a much more descriptive of what they've become.

Touch panels on military boats are called MFDs too. They are actually surprisingly good but they come with large handles you can use to stabilize yourself while the boat is moving.

I wonder if they've considered wearable haptics. A little buzzer somewhere in their flight suit might provide pilots with 'a button was pushed' confirmations. Maybe that would be too distracting though.

I've missed notifications on my phone on long flights because I didn't feel the buzzer. I can't imagine a fighter jet is more conducive to feeling a buzzer than a commercial airline.

Honestly, I'm amazed the buttons around the edge of the screen aren't context aware buttons based off which screen a non-touchscreen display is showing.

The issue is less the confirmation that a thing happened and more the ability to locate the right thing to begin with. It's easy to count buttons by feel and be sure you have the right one, but almost impossible to locate an arbitrary part of a touchscreen without looking at it.

It would probably be a little hard to feel when you're already getting buzzed constantly by the turbine engines and airflow.

True... maybe if they could get buzzers in the fingertips of the gloves... maybe. You're probably right though, the vibrations of the aircraft would drown it all out.

> One of the biggest drawbacks is that you can’t brace your hand against anything whilst typing – think how much easier it is to type on a smartphone with your thumbs versus trying to stab at a virtual keyboard on a large tablet with just your index finger

This is a common problem for large touch screens, that I've experienced for instance in Tesla's massive displays too. the way around it: always rest part of your hand against something stable, ie in the tesla wrap your pinky around the display and then use your thumb or index finger. For me it increased accuracy from 80% to 99%. This doesn't work if you need to touch the center of the screen, all other parts work great.

Tangent, but something similar is my main beef with the Tesla Model 3 (& Y), which otherwise would be the perfect candidate for my next car. I really wish it had some physical controls.

What's more infuriating is the lack of a binnacle right in front of you. The older Model S & X have one, but supposedly the upcoming refresh is going to take it away.

At least the F-35 has a helmet mounted display even if no HUD.

I believe Musk's rationale is that when these cars become self driving, there will be no need of a binnacle or even a HUD.

I thought this was pretty interesting.

The basic problem of lack of tactile feedback is one you can test easily by trying to type on on an iphone vs. an old school blackberry while riding a roller coaster.

When you have something to brace against, it's just a lot easier to be stable. I'm surprised researchers haven't figured out a way to solve this in a better way (maybe it's a haptic screen that can be reconfigured... the way LED keyboards work)

The 20% error rate the pilot describes seems... like a really really big problem.

Looking at those cockpit photos, I suddenly remember that I sat in F18 cockpit during one of the public military show. It is small, I barely fit into it, those pilots must be slim. Cockpit also looked a bit rundown/worn off, not sure the design itself or simply wear and tear.

In conclusion, it is a meh. So unlike the image portrayed in Top Gun novie.

Simple solution: each pilot gets the military equivalent of a stream deck[1] that is loaded with their own configuration. Have a mounting option in the cockpit and Bob's your uncle.

1. https://www.elgato.com/en/gaming/stream-deck

I'm fascinated by the picture from inside the Harrier cockpit, it's very much inline with the interior I've seen from other fighter planes. It just looks grimy and worn down, and fighter planes always seem to look like that in the cockpit.

I'd hate to be a carrier pilot and the helmet fails on blue water ops, you can't land, only option is to go for a swim. And single engine too over water... hmmmm no thanks. No backup instrumentation either.

A lot of the references you use to land on carriers are outside the aircraft, and the process is becoming largely automated. Besides, we've been flying single-engine aircraft blue-water for decades--it's not THAT bad in the grand scheme of things...

Carrier landings should move eventually to JPALS (all-weather landing system) in some time in the future. It's military Autoland that works even for carriers and amphibious assault ships).

Drones will also be using it, I think.

The carrier f-35 can hover. You can just line it up by sight.

I love how immediately obvious it is that the Harrier cockpit lets you see downwards.

Also, the red button on the stick really brings home just how much flying these machine is like a flying a loaded gun.

> I love how immediately obvious it is that the Harrier cockpit lets you see downwards.

That's mostly the camera angle. The F/A-18 and F-35 cockpit shows have the camera positioned about where the pilot's head would be (looking straight ahead through the HUD) whereas the Harrier photo has the camera positioned higher looking down into the cockpit.

Yeah, it's a fantastic cockpit. For what it's worth, just pressing either of the red buttons on the stick won't operate any of the weapons systems--there are plenty of other safety interlocks in place to prevent a bomb coming off accidentally.

Surprised that this was released... should be secretive

And then comes musk with neuralink. Oh boy imagine the number of zeroes on their contracts with the military.

F-35 is a slush fund for politicians and Military Industrial Entertainment Technology sector.

$1.5 Trillion and counting.

Alright, HN. Some required listening for you :

The Harrier : https://www.fighterpilotpodcast.com/episodes/056-av-8-harrie...

The F35 : https://www.fighterpilotpodcast.com/episodes/078-f-35-lightn...

The Eurofighter (for their short discussion on the voice commands): https://www.fighterpilotpodcast.com/episodes/058-eurofighter...

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact