Hacker News Reader

Comment by dpark | Hacker News Reader

Air traffic (and ground traffic) control are not simple problems. La Guardia has 350k aircraft operations (takeoffs and landings) every year. 1000/day. Peak traffic is almost certainly more than 1 plane every minute. Runways are always in use and the idea that some simple software will solve all the safety problems is not grounded in reality.

View on HN

This isn’t hypothetical, this system just exists in other countries. Digital systems can confirm flight instruction from ATC with zero radio communication.

I’m not saying we couldn’t move more into automation. What I’m saying is that doing so will not solve all of our air/ground control problems. We still have human pilots and humans driving vehicles on the ground. Switching from humans directing landings to machines might improve some things but will not solve for all (and probably not most) risks.

Literally the crash here was caused by a fire truck entering the runway.

The ATC told them to enter the runway because they were confused or distracted due t overwork.

No one here or anywhere is saying automation would solve or be able to handle everything that human operators handle, that's an argument you invented that no one is making.

People are saying automation could handle a significant portion of the routine things allowing humans to handle the more complex/finicky issues.

Even if automation could handle 10% of the most common situations it would be a huge boon. In reality its probably closer to 50%.

> The ATC told them to enter the runway because they were confused or distracted due t overwork.

Very possibly. It will be interesting what comes from the investigation.

> No one here or anywhere is saying automation would solve or be able to handle everything that human operators handle, that's an argument you invented that no one is making.

I’m asking if it would have solved even the current situation. The truck presumably saw the red light, and was asking to cross. Would traffic control have said no if more had been automated and if so, what automation would fix this? Unless we are supposing the truck would be autonomously driven and refuse to proceed when planes are landing, in which case, maybe, though that’s not really ATC automation anymore.

an automated system that could check if a plane is about to land on a runway and show some kind of alert or red light is hardly a stretch of the imagination

That’s such a great idea that it already exists and is deployed at La Guardia.

https://www.faa.gov/air_traffic/technology/rwsl

Thank you for providing your aviation knowledge to this discussion. What a classic example of tech people thinking that because they're smart, every other industry must be dumb and they can just jump in and fix it.

I also do not like this persistent tone of “everyone else is stupid; software would easily fix it” that pops up so often. Not all problems are easy to fix with some code.

To be clear, though, I don’t even have significant aviation knowledge. But this isn’t hard to learn about. That’s part of what irks me so much about this tone. It’s not just “I’m so smart” it’s “I’m so confident that you’re dumb that I don’t need to know anything about the domain you’re working in to know better than you”. Someone could ask ChatGPT why airports don’t have stoplights to stop traffic from crossing the runway and it would reveal the existence of this system.

> "I don’t need to know anything about the domain you’re working in to know better than you"

This frustates me to no end. Is it just an example of the Dunning–Kruger effect?

Something like that. It feels a bit different because it’s less about overestimating one’s knowledge/ability and more about underestimating the complexity of domains outside one’s expertise. But yeah. Very similar.

Me too, but I don’t like referring to Dunning-Kruger ever for multiple reasons. There are perfectly good labels like cockiness, arrogance, ignorance, presumptuousness, and wrongheaded. ;)

There are many issues with DK, and the paper’s widely misunderstood. For one, the primary figure demonstrates a positive correlation between confidence and competence, so according to DK’s own paper, high confidence is not an indicator of incompetence, contrary to popular belief. The paper also measured things in a very funny way (by having participants rank themselves against other people of unknown skill), and it measured only very simple things (like basic grammar, and ability to get a joke), and it only polled Cornell undergrads (no truly incompetent people), and there were a tiny number of participants receiving extra credit (might exclude the As and Fs in the class). Many smart people have come to the conclusion that DK is a statistical artifact of the way they did their experiment, not a real cognitive bias. Some smart people have pointed out that DK is probably popular because it’s really tempting to believe - we like the idea of arrogant people getting justice. The paper also primes the reader, telling them what to believe even though the title isn’t truly supported by the data. It’s an interesting read that I think would not pass today’s publication criteria.

Anyway, sorry, slash rant.

Yes, in fact I had considered adding your same thought to my initial comment. It's not impossible that a smart tech person might be able to improve the existing systems. The problem is the arrogance of not even checking what existing systems there might be, as if obviously they'd be too backwards to have any.

Agreed, but I see this in every industry. And though it's certainly arrogant on some level, I think of it in a more positive light: people are generally optimistic and want to solve problems.

My grandfather had a rule at his business for 55-ish years: we welcome your ideas and suggestions, but not for the first year. You spend that time learning our processes, decisions behind them, pain points, areas that need improvement, etc. You also spend that time doing the work and hearing from your colleagues. Then you can (hopefully) make informed suggestions. That's not possible in every situation, but I like the intent.

> people are generally optimistic and want to solve problems.

This is an amazingly positive spin on the behavior.

The habit where HN commenters greenfield solutions that are slightly worse versions of the ones experts already have in place is unmatched.

I meant something in-vehicle for ground vehicles, like an extremely simple extrapolation of current velocity and the extremely predictable trajectory of a plane, instead of depending on going back and forth over radio asking a very busy fallible human, but sure

even my cheap car has geofencing and automatic braking

I've worked on avionics professionally and I haven't crashed any of my planes yet...

“These lights … turn red in response to traffic, providing direct, immediate alerts without the need for input from controllers”.

It will be interesting to see what the report says. Did the light system not function? Did they override it? Do they ignore it consistently?

> geofencing and automatic braking

I’m not at all sure I want emergency vehicles to be blocked like this. And if they can override then it’s no different. They didn’t roll onto the runway on accident.

> I've worked on avionics professionally and I haven't crashed any of my planes yet...

Is this relevant somehow?

There's unfortunately an alertness problem WRT automated systems.

If the reason you have the human there is to handle the unusual cases, you run the real risk that they just aren't paying attention at critical moments when they need to pay attention.

It's pretty similar to the problem with L3 autonomous driving.

Probably the sweet spot is automation which makes clear the current set of instructions on the airport which also red flags when a dangerous scenario is created. I believe that already exists, but it's software that was last written in 1995 or so.

Regardless, before any sort of new automation could be deployed, we need slack for the ATC to be able to adopt a new system. That's the biggest pressing problem. We could create the perfect software for ATC, but if the current air traffic controllers are all working overtime and doing a job designed for 3 people rather than one, they simply won't have the time to explore and understand that new system. It'll get in the way rather than solve a problem. More money is part of the solution here, but we also need a revamped ATC training program which can help to fill the current hole.

In an ideal world this would be like rail traffic, where the runway would be 'locked' (red signal) due to the landing plane, and the fire engine would have to explicitly request an override to cross the locked runway, and importantly, this process has to be _rare_. If it's something that's done 5000 times a day, it'll be normalized. Everyone involved should be aware of the dangers of traversing a 'locked' runway.

My understanding that this scenario is exactly what happened here.

Changing the delivery method doesn't do anything to solve the problem of a controller sending an instruction that creates a hazard.

> Digital systems can confirm flight instruction from ATC with zero radio communication.

Digital comms is available in the US:

* https://www.faa.gov/air_traffic/technology/DataComm

The issue is that the final approach and landing (and taxiing?) environments are probably too dynamic for that: in this particular situation one of the vehicles was responding to an emergency (fire).

In addition to huge planes, there is baggage transportation, passenger buses (to mid-field terminals), fuel pumpers, emergency vehicles, snow plows, deicers, and general maintenance vehicles (clear debris off runways).

What is _really_ needed is a replacement of the archaic narrowband analog FM radio. Where you can't listen and talk at the same time. There are probably at least several dozen accidents where the inability to communicate with an aircraft or a road vehicle was a contributing factor.

I would settle for a good digital system with an ability to issue emergency/priority calls to specific receivers. Oh, and full-duplex communication.

I'm practicing for a sports pilot license, and I really have problems with understanding other pilots and the ATC.

Not only that, if 2 people talk at once they can cancel each other out and neither can be heard by anyone else.

Much of aviation is still based on pre WWII tech and practices like this and people underestimate how slow and difficult it is to change. Many piston aircraft still run on leaded gas, for example, the last existing market for it in the US.

> more than 1 plane every minute

Software routinely solves database coordination problems with millions of users per second.

Yup, by having backup runways.

And no fire trucks crossing the runways.

....they need to get to fucking fire

....if they go around kilometer of the runway the fire will turn into bigger fire

Two trucks

Did it? They didn’t get there so did we get bigger fire at their target?

I imagine the training will consist of something like changing the comms protocol to say “runway lights are on, control. Truck 1 confirming cross runway 4D?” prior to crossing. Double check so to speak.

Ground vehicles consistently have radio conventions that just don't fit into the aviation world. It feels like a contributor to this accident, you can hear the controller's brain skip a couple gears trying to understand the goofy word order from the truck.

Pilots and controllers speak the same language in the same order; ground vehicles just kinda say stuff.

The aviation-ized version of your proposal would be something like this:

> tower truck 1 short of 4 at delta, red status lights

That makes sense! I imagine the word order thing is just that ground vehicles are not aviation trained. They just happen to be in the same space.

A third runway for Heathrow was formally proposed in 2007 and is projected for completion in 2040. This is an airport so overburdened people are buying and trading slots.

This isn't a Kubernetes cluster where you can add VMs in 30 seconds.

https://en.wikipedia.org/wiki/Expansion_of_Heathrow_Airport

I'm pretty sure the amount of data isn't the problem here. Maybe it's the number of corner cases? You would still want some human-in-the loop with quality UI for ATC.

There are plenty of stories of ATC helping to guide pilots back to the ground after an engine failure or after a student pilot had their instructor pass out on them or something like that.

Even if most of the work is routine, you definitely still want a human in the loop.

It's worth pointing out that plenty of pilots take off and land safely at uncontrolled airports. ATC is a throughput optimization; the finite amount of airspace can have more aircraft movements if the movements are centrally coordinated. It feels like we are nearing the breaking point of this optimization, however, and it's probably worth looking for something better (or saying no to scheduling more flights).

The FAA already does issue temporary ground stops for IFR flights when ATC capacity is saturated. This acts as a limit on airlines scheduling more flights, although the feedback loops are long and not always effective. The FAA NextGen system should improve this somewhat.

https://www.faa.gov/nextgen

True. But to avoid 1 minute unavailability per year requires 99.9999 % availability

Like any scale system, degrade the experience. Use radio if the more advanced systems are unavailable?

>> Software routinely solves database coordination problems with millions of users per second.

A naive view that confuses the map with the territory.

While in a database state you write a row and reality updates atomically....for aircraft they exist in a physical world where your model lives with lag, noise, and lossy sensors, and that world keeps moving whether your software is watching or not. Failed database transactions roll back, a landing clearance issued against stale state does not. The hard problem in ATC is not coordination logic but physical objects with momentum, human agency, and failure modes that do not respect your consistency model.

with extremely controlled conditions. There is no fog in database, nor fallible humans involved, What an ignorant response

In a digitized environment. We cannot yet simulate the real world.

That makes digitization even more important, you sold me.

No one said it was simple. You're tilting at windmills.

Literally called it “low hanging fruit”.

But context is important. "Low-hanging fruit" doesn't mean the solution is "easy" in a vacuum, it just means this specific aspect is the easiest and/or most obvious place to start attacking a problem.

Or to stick with the language of the analogy, every fruit tree has some fruit that is lower than the others. That doesn't mean all "low-hanging fruit" is within arm's reach of the ground, some fruit just doesn't require as big of a ladder as other fruit.

This comment isn't a judgment of this specific case. I don't know enough about ATC to have any confidence in my opinion on the viability of replacing humans with software.

That goal post moved so fast it made a whooshing noise as it passed

I think you're mistaken. That whooshing sound must have been my comment flying over your head.

That was my first comment in this thread, so there was no established goal to change. My sole goal was to clarify the meaning of an idiom that the comment I was replying to was misstating.

I even included a disclaimer that "This comment isn't a judgment of this specific case", so I don't know how you could have received it as such.

I disagree with you entirely. "Lowest-hanging fruit" isn't the same as "low-hanging fruit". The phrase "low-hanging fruit" does specifically mean that the solution is easy, in a vacuum - the fruits are "low", which is not relative to the other fruits or the height of the tree, but relative to the ground.

I'll just defer to Merriam-Webster[1]:

>the obvious or easy things that can be most readily done or dealt with in achieving success or making progress toward an objective

So not only can it be "obvious" rather than "easy", it is also in the context of "achieving success or making progress toward an objective". There is nothing in the definition that requires either this specific step or the overall goal to be easy.

[1] - https://www.merriam-webster.com/dictionary/low-hanging%20fru...

Hehehehe, grounded.

Reply on news.ycombinator.com