It's not only that it was text based, but that it failed in all sorts of ways.
I had a particularly bad afternoon in early 2000s when I tried to install Linux on various machines (4 or 5, each one with different hardware) and failed with all of them. Of course, I wasn't as savvy as I am now, but it's been years since I've seen an issue with a Linux install; although I might have to download and compile a driver for one thing or another, the install itself doesn't fail.
Same with me - late 90s. Celeron 366Mz, 64 MB RAM, 4.3 GB HDD. Don't remember the video card and the sound card. Used to stay up all night to try and fix X11. And once you got it running - victory !!!
It's also the kind of project where LLM license-washing is going to be the most dangerous for the project itself. Projects like Wine and ReactOS have to be extremely careful about what kind of code to accept, because taking code from one of the countless Windows source leaks is just a recipe for legal disaster.
Looks like a lot of improvements on long-standing hard-to-find issues. Kudos to the team and thank you for providing an alternative OS that keeps getting better
One of my favorite projects, maybe someday it can be a true daily driver, I wonder if that's the goal for 1.0 is to claim that they've gotten it stable enough to run as though it were Windows 7 or something to that effect.
TIL that the GUI of ReactOS isn't based on React. All these years spent baffled that who would like to build an OS with JS, thanks to me only reading the headlines lol
I do a lot of industrial automation and it is the one thing that keeps me boat-anchored to Windows. Every little programmable doodad like signal conditioners and temperature transmitters have their own USB programming cables and their own, mostly crappy windows config software. PLC software is another abomination, hopelessly Windows anchored, bloated GUI monsters, some built on top of Visual Studio (Beckhoff TwinCAT.)
I have donated to ReactOS in the past in the hopes of one day freeing myself from this hell.
That is only recently become an option you could try, and we don't know long term how well it will work. Remember laws may change (and there are some legal issues around LLMs that are not clear at all)
Because that requires breaking the almost assured stipulation in a software license agreement that says "You will not try to RE this software", which is fine to ignore for home gamers, but not something you generally want to mess around with in a business setting where the risks are much higher.
The thing about this type of thinking is that it often requires risking millions of dollars and a major distraction from your primary goal in order to find out if it's true or not.
As would be running critical software certified for Windows on ReactOS as envisioned by GP. So, either we talk about an enterprise setting, in which GP already fucked up, or a more personal setting, in which both whaat GP and I said are reasonable
Also, the one thing that's important above all else in industrial automation is reliability. If your LLM doesn't know about one tiny little detail that ends up causing your product to fail, your customers start losing money and you can no longer point the finger at your PLC manufacturer. Make bad gambles like that and you won't be in business for long.
The reason industrial PLCs cost so much is because of their reliability. They need to be working flawlessly 50 years in the future. Trying to vibe code your way into a PLC compiler feels like it could blow up in your face after a little while.
That is also ignoring the safety PLCs with safety instructions that are certified for use with ISO 13849. Can you imagine getting inspected because your safety subsystem cause injury and you explain how you don't use the officially licensed software to program it.
You completely stripped what I said from its context. GP explicitely said they donate to ReactOS in the hopes of one day being able to run that industrial software on it instead of on Windows. In this context, running the software on ReactOS instead of official Windows would already fail your hypothetical inspections. This is why I suggested trying the LLM route.
If I had to explain a failure in production I would much rather say "due to availability of old hardware, we ran this on a different version of Windows/inside a VM/inside a Windows compatibility layer/Wine/ReactOS" than "we vibecoded the entire driver stack from scratch, go ask Claude why the plant blew up lol".
Running the same software in a different environment is unlikely to cause issues. But with vibe coding, you're throwing away the code the manufacturer certified as working, and the best possible outcome is it bricks your hardware in dev before things escalate to a post-mortem.
You can forward USB devices into a VM. About 15 years ago I was using a Windows VM on Linux to program PIC18F??? devices with the custom (JTAG based?) programmer provided by Microchip. The upside is that you can preserve and share this windows environment for other coworkers, and have separate VMs for different projects (perhaps one you can update, and one you can't ...)
thataccount | 15 hours ago
forinti | 14 hours ago
I had a particularly bad afternoon in early 2000s when I tried to install Linux on various machines (4 or 5, each one with different hardware) and failed with all of them. Of course, I wasn't as savvy as I am now, but it's been years since I've seen an issue with a Linux install; although I might have to download and compile a driver for one thing or another, the install itself doesn't fail.
thataccount | 14 hours ago
unlimit | an hour ago
prmoustache | 11 hours ago
Unrecognized/unsupported hardware is another thing but I always could finish the install and at least reach the console login prompt.
I've never used bleeding edge equipment though, euther second hand or brand new but proven tech.
tiahura | 15 hours ago
actionfromafar | 15 hours ago
aussieguy1234 | 14 hours ago
Now we have an infinite loop
binary132 | 13 hours ago
Koshkin | 7 hours ago
TonyTrapp | 13 hours ago
tiahura | 12 hours ago
montroser | 15 hours ago
mysterydip | 15 hours ago
Aldipower | 15 hours ago
MisterTea | 14 hours ago
Lizard5260 | 14 hours ago
tech234a | 12 hours ago
Koshkin | 7 hours ago
> must be running a 64-bit x86
giancarlostoro | 14 hours ago
guitarlimeo | 14 hours ago
kijashdkujdfhas | 14 hours ago
lucasoshiro | 11 hours ago
If you an OS based in JS, maybe the closer we had was NodeOS [1]
[1] https://node-os.com
Koshkin | 7 hours ago
MisterTea | 14 hours ago
I have donated to ReactOS in the past in the hopes of one day freeing myself from this hell.
Almondsetat | 14 hours ago
bluGill | 14 hours ago
wpm | 13 hours ago
Almondsetat | 13 hours ago
jmcqk6 | 13 hours ago
Almondsetat | 13 hours ago
estebarb | 13 hours ago
ktm5j | 9 hours ago
kvuj | 13 hours ago
That is also ignoring the safety PLCs with safety instructions that are certified for use with ISO 13849. Can you imagine getting inspected because your safety subsystem cause injury and you explain how you don't use the officially licensed software to program it.
For personal project, that could be fun however.
Almondsetat | 13 hours ago
nvme0n1p1 | 9 hours ago
Running the same software in a different environment is unlikely to cause issues. But with vibe coding, you're throwing away the code the manufacturer certified as working, and the best possible outcome is it bricks your hardware in dev before things escalate to a post-mortem.
ploxiln | 13 hours ago
Koshkin | 7 hours ago
Hasn't the point been for the OS to be compatible with the existing Windows drivers?
Telaneo | 2 hours ago
prvt | 22 minutes ago