Remember, this is the government we're talking about. That system was likely being worked on from 1980s to 1990s before itactually went in to implementation phase in 1990s.
Kind of makes sense for the time given hardware was evolving so rapidly after Apple and MS brought PCs to the masses; a solution deployed in 1980 was out of date by 1981.
Workers of the day were used to paper first routines; PCs on every desk were not common until the 90s when home computers, like phones in 2026, hit a fast enough for most tasks in an office constant upgrades are not necessary.
I feel like people under a certain age don't really understand this.
While we see improvements in hardware and software today, it's NOTHING like the late 80's and 90's. Our first home pc was out of date in less than a year (thank you for that recommendation, Tiger Direct). As a child, I couldn't run games that came out more than about 10-12 months after we upgraded our machine. It really was stunning to be a part of.
And now my home pc is about 4 years old, and replaced a machine that was 9. I plan to run this one for at least another 6 or 7 years unless something amazing happens that nobody sees coming.
If you didn't upgrade often, it could happen that you bought a new machine that was 10x as fast as the old one. It was just incredible how fast everything was after such an upgrade. A 486-66 to a Celeron-416 (common overclock at the time) in my case. The 486-66 was a year or two old for me - I had fished board and CPU out of the trash somewhere because before that I had a 486-33 bought in '94.
>The FAA cited an October 13 event where a fiber optic cable connecting Houston George Bush Intercontinental Airport to the FAA’s network was cut outside an airport gate, leading to the loss of data services between FAA facilities and requiring a ground stop.
Do they seriously not have a backup connection? I would have expected them to have a backup stelite or maybe even a 4G/5G mobile data system in place.
If I can use my phone as a Hotspot when Comcast is doing some service so I can work, I don't understand what idiot designed an airport without backup connections.
Useful aphorism in general, not terribly applicable in this case. Even if it had been tested regularly, it was cut around the same time the primary failed. Regular testing doesn't stop construction crews from cutting a line.
A more useful comment might have been to say something about the depth of the backup, having only one backup connection not being enough or something.
If you are curious: one of the most frequent problems that disables fire sprinklers is maintenance forgetting to turn the water back on. No one notices until it is too late.
No one notices the thousands of times the redundant links works just fine. We only notice when the primary has an outage and a construction crew cuts a fiber line at the same time.
> I don't understand what idiot designed an airport without backup connections.
I have literally spraypainted the pavement and wall - "One fibre goes here, and down on the corner there, the other fibre goes there" - just to ensure a good separation of routes. I want the pinch point to be within the rack that the routers sit in.
In my carreer I have seen plenty of system integrations working on ftp'ing request response documents to file folders on "SaaS" systems while the SV boys were considering SOAP obsolete and overnight Cobol and RPG mainframes document sincs while they were waxing about microservises and noSQL.
I was in a tower last year. They were sequencing aircraft with little loops of paper placed onto a wooden dowel. It can and should be better. Doesn't need to be whiz-bang, but it should not be paper slips and wooden sticks. (and floppy disks)
Paper flight progress strips [1] have famously resisted replacement, even though they keep trying [2].
Every 5 years or so, a different publication will write an in-depth story about how difficult they are to replace, because nothing works nearly as well, and the controllers go back to using paper every time a new system is trialed. I think I've even read about them in a cognitive psychology textbook at uni.
This is nearly the exact reason most 'smart home' systems fail eventually. They become very brittle after the hardware gets 'old'. Where as the dumb tech from the 1950s still 'just works' and costs 1/20th of the smart switch.
In that system you can tell what is going on by looking at in a glance and works with no power. Sometimes simple is a great answer.
However, I can think of one reason the 'no power' argument fails. They do a lot by radio. That needs power. Also you would need to retrain everyone. Oh and you probably would get locked into one company that wants to control the whole software stack and sell tons of support.
I have a friend who is a real estate agent for the wealthy. He laughs at the people who automate their homes. The automation inevitably fails, the vendor has disappeared, and the owner is faced with ripping all the wiring out of the walls and replacing it. Usually it gets replaced with conventional wiring.
When my home was built in 1999, I poured over all the home automation systems. I opted for conventional wiring. But I did wire every room (myself) with cat5 cable and RG6 coax, in a star configuration. It has served me very well, no problems at all.
I don't use the RG6 anymore, as my TV service comes over the intertoobs now. The cat5 isn't supposed to work with gigabit equipment, but it works fine. Wifi has improved by leaps and bounds, but I still rely on hardwired ethernet, as it is stupid reliable.
> I did wire every room (myself) with cat5 cable and RG6 coax
I think the best idea is to install conduit as the house is built, so that outdated wires can be replaced with newer wires by just tying the end of the old wire to the end of the new wire and pulling.
Just be sure to inspect the house as it is being built, because I think one user on HN wrote that conduit was required by law, but when he went to swap the wires, he couldn't pull them - he discovered there was only pipe in the visible terminations, not the whole route.
Conduit is indeed a better idea, but it would have been too much work for me. I figured the bases were covered well enough with 4 cables to each destination - 2 cat5 and 2 rg6. The duality was in case one of the cables failed, or got a nail driven through it.
Those things actually work great, not everything has to be modernized for modernization sake... that's how we get cars with touchscreens instead of buttons and knobs.
Better in what way? They have avoided replacement specifically because they work extremely well and none of the alternatives have been demonstrably better.
a mechanical (wood, paper , etc) state machine is very resilient and easy to read even from across the room. No electricity demands, no lighting needed. Contains it's own mutual-exclusion (only one person can modify it at any time).
the digital version would require 10+ endpoints with 1000x points of failure.
So my mother worked at a bank when she was in college. This is the 70s. Later in life she went back to work, in the 90s, at a bank. When the computers went down and they had to close the books on paper and she was the only one that knew how to do it.
There is something to be said for knowing how to deal with critical systems when the automation breaks.
My grandpa used to be a manual typesetter in the newspaper industry and retired in the 80s.
A few decades later, similar story, he came back to work part time, when desktop publishing was all the rage.
Welp, of course the computers went down.
So they fired up the old hot metal typesetting machines and he manually set all 100,000 words of that Newspaper by hand, and then they went to print using 19th century steam powered print lines.
Except, that story is total bullshit that I made up. Because "the computer went down" is something we can mitigate much easier than doing everything by hand.
Sometimes the old technology actually is objectively worse, inefficient, and obsolete. Like floppy disks, or manually having humans close the books at a bank.
That is really not the worst thing about ATC. Far from it.
I have heard experienced pilots lamenting that we're still using voice on a scratchy radio, with garbled audio and people accidentally talking over other transmissions both being constant problems, for essentially all vital comms.
While I have no problem with keeping that as a fallback, there's zero reason why we shouldn't have, at least for airliners that have two pilots, a text-based communication system. We can still convert the text to speech in the cockpit to keep awareness the same, but we could play overlapping transmissions one after the other. And think of the other advantages. People wouldn't get callsigns wrong, could be alerted specially when their callsign was being talked to...
It's complicated because we will need to cooperate with all countries, but it's not like we don't have ICAO to work within.
We already have some systems in place for this, specifically stuff like ACARS and CPDLC. I suspect the issue is in congested terminal airspace, though; if you have many aicraft talking to an approach controller it is not uncommon to have multiple transmissoins at the same time.
However, approach phases are exactly the same periods when the pilots probably shouldn't be looking at their FMCs and texting. Voice communication is faster, and so is acknowledgement of voice commands, while allowing the pilots to concentrate on more critical controls. For example, in many airports it is possible to get departure clearances and such via text, but take-off clearance and taxi instructions will still be voice over radio.
I think anything we could roll out today would be equally superior whether audio or text mode. Both ought to be available. The important part is instead of enduring the potato audio quality and the zero managing of conflicts, we could have crystal clear audio and, with some kind of a CDMA/packet switched thing, two transmissions sent at the same time could be buffered and play sequentially, and also without too much work, the message “for you” or comms for an emergency aircraft would be prioritized.
When I was 10 years old, way back in 1980, I got my first 5.25" diskette from the salesman of a COMPUTER AGE shop. He got tired of watching me type in the same old BASIC listings after school and spared me the damage of it by giving me the diskette, on the provision that I let other customers use the demo APPLE II's.
Sitting here now in 2026 I have a box containing almost 3000 3.5" FHD floppy disks.
Each about 1.544 Megabytes, unless you cut the notch.
I literally dumpster dived this rescource after being invited in to do so at a local electronics shop closure, some years ago now, on the proviso that I did not leave empty handed.
There are about 12 devices in my living room that can use these things, but I'll never use more than 12.
Some day, I'll see the last one go. That'll be weird.
Same. Over and over again. Every time, learning new things. And then trying to port things from one system to another is how I became a disciplined cross-platform systems software developer. And then, when C was kind of new, I bought a compiler instead of drugs and booze, like my peers at the time. Oh, what a world that has wrought!
I have heard of exploits that can root a machine from inserting a USB device (in fact, this has been used by attackers to conspicuously leave thumb drives in the parking lots of organizations they want to attack).
I have never heard of a sploit that roots your machine from inserting a floppy.
true, there were viruses, and they came over floppy. but USB has a hidden firmware layer. I remember one on Apple a few years ago they said is completely un-patchable because it affected the firmware on the USB chip itself. I dont remember what happened after that, though.
You're going to run that on a computer from the late 1980s or early 1990s? And how do you get the data onto those computers when they're not connected to a network, or the same network, as your other systems?
Problem is that kind of thing is usually made by and for hobbyists. So it would require a ton of certifications to prove that it’s ok for life or death stuff.
I'm more interested with the data integrity protections before the data got onto the floppy, and why those protocols (e.g. signature verification) couldn't be completed via email at the endpoint.
> “We’re moving away from outdated, manual processes and replacing technology that once relied on floppy disks with a modern system that allows controllers to instantly share critical flight alerts, restricted airspace schedules and pilot reports,"
So they make their systems run on the internet now?
86 comments
[ 3.3 ms ] story [ 13.1 ms ] threadAnother article with some more details: https://www.pcmag.com/news/the-faa-is-still-using-floppy-dis...
I know they are saying IDS-4 is a "1980s" system, but it appears IDS-4 was actually developed in the early 1990s; it uses MS-DOS and 3.5 inch floppies: https://sysatl.com/systems-atlanta-inc-awarded-faa-ids4-sust...
Workers of the day were used to paper first routines; PCs on every desk were not common until the 90s when home computers, like phones in 2026, hit a fast enough for most tasks in an office constant upgrades are not necessary.
There was a lot of ducks to align back then.
While we see improvements in hardware and software today, it's NOTHING like the late 80's and 90's. Our first home pc was out of date in less than a year (thank you for that recommendation, Tiger Direct). As a child, I couldn't run games that came out more than about 10-12 months after we upgraded our machine. It really was stunning to be a part of.
And now my home pc is about 4 years old, and replaced a machine that was 9. I plan to run this one for at least another 6 or 7 years unless something amazing happens that nobody sees coming.
Do they seriously not have a backup connection? I would have expected them to have a backup stelite or maybe even a 4G/5G mobile data system in place.
If I can use my phone as a Hotspot when Comcast is doing some service so I can work, I don't understand what idiot designed an airport without backup connections.
That was the backup connection. The primary failed, and concurrently the backup had been severed by a construction crew.
A more useful comment might have been to say something about the depth of the backup, having only one backup connection not being enough or something.
No one notices the thousands of times the redundant links works just fine. We only notice when the primary has an outage and a construction crew cuts a fiber line at the same time.
I have literally spraypainted the pavement and wall - "One fibre goes here, and down on the corner there, the other fibre goes there" - just to ensure a good separation of routes. I want the pinch point to be within the rack that the routers sit in.
I don't get why an airport can't do similar.
Every 5 years or so, a different publication will write an in-depth story about how difficult they are to replace, because nothing works nearly as well, and the controllers go back to using paper every time a new system is trialed. I think I've even read about them in a cognitive psychology textbook at uni.
[1] https://en.wikipedia.org/wiki/Flight_progress_strip
[2] https://www.faa.gov/air_traffic/technology/tfdm/efs
In that system you can tell what is going on by looking at in a glance and works with no power. Sometimes simple is a great answer.
However, I can think of one reason the 'no power' argument fails. They do a lot by radio. That needs power. Also you would need to retrain everyone. Oh and you probably would get locked into one company that wants to control the whole software stack and sell tons of support.
When my home was built in 1999, I poured over all the home automation systems. I opted for conventional wiring. But I did wire every room (myself) with cat5 cable and RG6 coax, in a star configuration. It has served me very well, no problems at all.
I don't use the RG6 anymore, as my TV service comes over the intertoobs now. The cat5 isn't supposed to work with gigabit equipment, but it works fine. Wifi has improved by leaps and bounds, but I still rely on hardwired ethernet, as it is stupid reliable.
I think the best idea is to install conduit as the house is built, so that outdated wires can be replaced with newer wires by just tying the end of the old wire to the end of the new wire and pulling.
Just be sure to inspect the house as it is being built, because I think one user on HN wrote that conduit was required by law, but when he went to swap the wires, he couldn't pull them - he discovered there was only pipe in the visible terminations, not the whole route.
the digital version would require 10+ endpoints with 1000x points of failure.
There is something to be said for knowing how to deal with critical systems when the automation breaks.
A few decades later, similar story, he came back to work part time, when desktop publishing was all the rage.
Welp, of course the computers went down.
So they fired up the old hot metal typesetting machines and he manually set all 100,000 words of that Newspaper by hand, and then they went to print using 19th century steam powered print lines.
Except, that story is total bullshit that I made up. Because "the computer went down" is something we can mitigate much easier than doing everything by hand.
Sometimes the old technology actually is objectively worse, inefficient, and obsolete. Like floppy disks, or manually having humans close the books at a bank.
I have heard experienced pilots lamenting that we're still using voice on a scratchy radio, with garbled audio and people accidentally talking over other transmissions both being constant problems, for essentially all vital comms.
While I have no problem with keeping that as a fallback, there's zero reason why we shouldn't have, at least for airliners that have two pilots, a text-based communication system. We can still convert the text to speech in the cockpit to keep awareness the same, but we could play overlapping transmissions one after the other. And think of the other advantages. People wouldn't get callsigns wrong, could be alerted specially when their callsign was being talked to...
It's complicated because we will need to cooperate with all countries, but it's not like we don't have ICAO to work within.
However, approach phases are exactly the same periods when the pilots probably shouldn't be looking at their FMCs and texting. Voice communication is faster, and so is acknowledgement of voice commands, while allowing the pilots to concentrate on more critical controls. For example, in many airports it is possible to get departure clearances and such via text, but take-off clearance and taxi instructions will still be voice over radio.
Sitting here now in 2026 I have a box containing almost 3000 3.5" FHD floppy disks.
Each about 1.544 Megabytes, unless you cut the notch.
I literally dumpster dived this rescource after being invited in to do so at a local electronics shop closure, some years ago now, on the proviso that I did not leave empty handed.
There are about 12 devices in my living room that can use these things, but I'll never use more than 12.
Some day, I'll see the last one go. That'll be weird.
Equally likely: they'll see YOU go. Pape... uhm I mean floppies are patient.
I assume they just mean _old_ computers here. Unless they have some specialized analog air traffic computers? That would be pretty cool.
I have never heard of a sploit that roots your machine from inserting a floppy.
0. https://github.com/keirf/flashfloppy
(in which case, why replace)
So they make their systems run on the internet now?
What could possibly go wrong?