[ThinkBeat]

A first step to mitigate some of the risk would be to move the system to a virtualised system. This could be in each location or more centralised which would make the maintenance of the fleet of old computers easier.

Floppy can be copied to hard disks and will not have to worry about failures of mechanical parts involved in reading floppy drives.

Developing a brand new system would take quit a lot of time. As all systems du if they need extreme uptime. Starting that effort now is ok but I would guess it would be take at leas a couple of years. Significant work would have to understand in detail what the current system does and does not do, and then map out what a system should do.

[DocTomoe]

We IT folks tend to quickly propose solutions to systems whose complexities we do not completely understand. That's fine when it is about serving ads or managing book orders. It's not ok when the stakes are high.

Virtualization just adds another layer of complexity to an already fragile system which literally thousands of human lives depend on every day. Adding more complexity is not a neutral act here, but neglectful manslaughter waiting to happen. Aviation is a low-tech, never-touch-a-running-system, risk-averse environment for a reason.

Floppies were useful because you could easily take them and take them to another, secondary, sometimes air gapped backup system. Replacing this functionality means replicating not just the data transfer, but also the safety architecture - which includes physical isolation and manual fallback paths. To recreate, the best chance would probably be something like storing the relevant info on thumb drives - but then you have whole new family of attack vectors by hostile forces (anyone still remember Stuxnet), which floppies did not have in that form?

And then there's the pesky aspect of international interoperability. One country alone cannot just storm forward. We are looking at decades of upgrades and alignments here. And that process already is underway. But proposing a radical change without acknowledging the full scope of what that entails - from certification cycles to human factors to geopolitical coordination - is not progress, it’s hubris.

[ThinkBeat]

It is true, it introduces new risks. and would have to be tested well and attempt to be certified. I am not saying its something you "just do overnight".

Still it removes more risks than it introdcues. (IF it tests out ok)

Though they have a stack of replacement PCs ready to go and lots of floppy drives and floppy disks to quikly replace whatever may break.

Writing new code from scratch, introduces a lot more risks. but also offers a promise of something much better.

[jiggawatts]

Eeeeexcept that floppies are horrifically unreliable. I remember feeding disk number 27 out of 33 only to get a "bad sector" error an hour into a software install. I'm still salty about that one.

"It's not broken" is the cry of the bad manager that hasn't done the proper analysis, hasn't actually looked at the pros and cons, but has simply become complacent and comfortable with the devil they know.

If they're still using physical floppies, then their process is broken now, so virtualising it will almost certainly un-break it.

A simple "clarifier" for this kind of thought process that I like to use is: If you were already using the new option (virtualised legacy hardware), would you think it a good idea to convert it to using open drives with convenient dust ingress, non-existent support and supply chain, glacially slow mechanical moving parts, and hilariously antiquated crunching noises for all data access? Would you? Really? Or would you recoil in horror at the very idea?

I use the same kind of logic on people who think staying on Windows Server 2012 in <current year> is a good idea. Would you downgrade Windows Server 2025 to 2012? Why not? You think it's a great platform, apparently!

PS: I worked on a large scale DOS-era software virtualisation project where we moved ~20K users onto a Windows + Citrix platform. We eliminated about 6000 floppy drives and about a million(!) tapes, and the resulting system was so much faster and reliable than the original that people were trying to bribe the project manager to be put at the front of the migration queue.

[II2II]

> I remember feeding disk number 27 out of 33 only to get a "bad sector" error an hour into a software install.

I love this fixation on floppy disks. The article likely brought it up because it is a recognizably obsolete technology, but didn't cite why (or even if) it was a problem. I'm sorry, but a nightmarish software installation scenario doesn't cut it. It is highly unlikely that they are doing in situ software installations from floppy diskettes.

The danger in such armchair quarterbacking is that it undermines the authority of the agencies that are in charge of making decisions. If there are legitimate reasons to question their authority, by all means do so. Yet, when doing so, understand their requirements and provide evidence as to why their authority should be questioned. Also be prepared to be unsatisfied by some of their answers due to differences in perspectives.

[gladiatr72]

You make good points, but now Citrix and Microsoft have them over a barrel. Curious how such a migration looks in 2025 with Microsoft pushing everything to Azure, though and Citrix's acquisition by Vista Equity (2022).

[jiggawatts]

That migration occurred in 2007, the whole thing was replaced by a web app in 2017. Over that decade they saved many millions of dollars. Large scale disturbed legacy hardware is much more expensive than Citrix licenses!

When we started in 2007, systems still using floppies were considered ludicrously legacy and people could hardly believe me when I told them it was still in production and used for a critical system affecting millions of people.

Anyone still using floppies in 2025 has no excuse, stop trying to justify incompetence and sloth.

[DocTomoe]

> I remember feeding disk number 27 out of 33 only to get a "bad sector" error an hour into a software install. I'm still salty about that one.

That's why mission-critical systems have several sets of floppy disks, and disk-multiplication stations.

> Would you? Really? Or would you recoil in horror at the very idea?

Depends. If the old system is certified and has all error modes defined, while the other new system is a black box with exciting new ways to screw up, I'd go old system ten out of ten times. Which incidentally is why NASA uses ancient chips when they build new robotic drones.

> I worked on a large scale DOS-era software virtualisation project where we moved ~20K users onto a Windows + Citrix platform.

Respectfully: How many lives would you have extinguished had your new system failed? How many failure modes did you encounter during your virtualisation project? How many external systems - which also relied on a very specific way of doing things and would have murdered people if talked to wrongly did you interface with?

No need to answer. We have all had such projects. We know things break before, during, and after the switchover. Only in some environments, systems absolutely cannot break, ever. Aviation is not your average 'let's get us a new mail server' migration project.

[jiggawatts]

> How many lives would you have extinguished

I’ve worked on life & death Citrix modernisation projects several times: the local equivalent of the 911 emergency phone call centre and then computers on wheels used for during paediatric surgeries. A help line for suicidal children too.

People conflate the usecase with the technology, assuming that “important thing” must have some mystical properties that requires legacy or some other “special flavor” of IT architecture. They’re wrong.

The best example of this flawed thinking was some person arguing with me about the computer upgrade that F-22 fighters are receiving this year… to the same level of performance as a first-gen Apple Watch!

Of course, that costs an absurd amount of money and is already delayed.

“But it’s a stealth fighter!” people will argue until they’re blue in the face.

Sure. Yes. But that’s a property of the outside surface, not the computer inside.

Other modern fighters, including stealth fighters, have hilariously better computers for a fraction of the cost. The F-22 procurement process was corrupted and some vendor is doing the minimum, twenty years late, at ten times the price. That’s what happened. Everything else is a “story”. A fiction. A cover of the ass type.

Same thing here. There’s a contract for providing IT services to the FAA. It’s a bad contract. That’s what happened. That’s all. There is no mystical or magical capability provided by floppies that can’t be better served by, for example, USB thumb drives.

[trollbridge]

Are you aware the hardware in an F-22 has to be hardened against an EMP?

[jiggawatts]

At the risk of replying to someone with "troll" in their username...

Yes, but the entire point is that other fighter planes have identical requirements but don't have comically out-dated avionics.

Don't guess. Don't make up stories. Don't carry water for incompetent people that are protecting their own backsides.

Compare. Look at what other, more competent people have done, and use that as your benchmark.

That's always the key with these things. You don't have to be an expert. You don't have to have secret knowledge. Other people do. Just look at what they've achieved (or haven't), and compare against that.

When people come up with excuses, you don't have to believe them, even if they're experts in an esoteric, specialist field such as "nuclear-war-resistant stealth fighter design". Even if they're some sort of "authority", so are other people that designed their own stealth fighters.

This is a very useful life skill. Use it!

As a random example, there's countless arguments from "experts" such as economists, politicians, and industry thought leaders about how unaffordable universal government-funded health-care would be in the United States. Meanwhile, dozens of similar countries have done it for decades! Just look at the success of other countries and then dismiss the excuses you hear back at home out of hand because now you know: they're just excuses, not reasons.

[HPsquared]

I wonder if anyone makes a virtual floppy drive that replicates the performance characteristics. I.e. to avoid a faster virtual drive uncovering dormant race conditions. Something like a developer assuming "I have enough time to do this processing before the disc makes another rotation" etc.

[guax]

I think any of the "off the shelf" gotek emulators should suffice for this. They're made for people to keep playing games on old hardware. I would assume copy protection and other shenanigans would be the creme de la creme of abusing the hardware.

This is to get rid of the media only. You'll still be using the original compute hardware. But it would be an interesting step.

I feel that most of the desire to upgrade is cultural and not technical. People love to talk about the floppies being used while its just a small part of the equation. Cost and risk of creating a new system with the same reliability expectations is hard when the incumbent has decades of iteration. For systems that do not require more performance or energy efficiency the accounting on upgrading looks very different.

[kube-system]

https://www.amazon.com/Gotek-SFR1M44-U100-1-44MB-Floppy-Emul...

[shakna]

Probably MartyPC [0]. It replicates most of the physical aspects of devices.

> In May 2023, MartyPC became the first PC emulator capable of emulating every effect in the PC demo Area 5150.

[0] https://martypc.net/

[hulitu]

> A first step to mitigate some of the risk would be to move the system to a virtualised system.

... running Windows 11. Flight delayed because Windows is updating.