This blog post is intended to notify all Windows Hardware program partners that Microsoft has taken a strategic initiative to clean up legacy drivers published on Windows Update to reduce security and compatibility risks. The rationale behind this initiative is to ensure that we have the optimal set of drivers on Windows Update that cater to a variety of hardware devices across the windows ecosystem, while making sure that Microsoft Windows security posture is not compromised. This initiative involves periodic cleanup of drivers from Windows Update, thereby resulting in some drivers not being offered to any systems in the ecosystem.

↫ Microsoft’s Hardware Dev Center

The general gist is that Microsoft is going to remove all drivers from Windows Update for which newer versions exist – or, to put it in a different way, only the latest versions of a driver are going to remain available on Windows Update. It’s effectively a clean-up of Windows Update, and the only way older versions of drivers will remain available on Windows Update is if the manufacturer in question can make a “business justification” to keep them around.

Some of this may sound surprising, since many people assume Windows Update only offers the latest versions of drivers – annoyingly so, sometimes – but this isn’t the case. Corporations with fleets of devices can actually determine exactly which drivers get sent to their devices, including opting for older versions in case newer versions have regressions or otherwise cause issues. Sometimes you just don’t have a choice.

According to Adam Demasi, the creator and maintainer of the amazing Legacy Update service, Microsoft hasn’t deleted a single driver or update from Windows Update since 2001 (save for problematic updates). This results in a truly massive collection of updates and drivers, and that’s causing real problems for Microsoft.

Windows Update has a pretty cool system of describing whether an update is necessary to be installed on the current system, or if it is already installed. It also builds a relationship graph between updates, to indicate when they have been replaced by a newer update that includes all changes from the previous update. That system is also its downfall, causing the Windows Update service to be incredibly slow in checking for updates, possibly never completing the check at all. This issue also applies to WSUS, which despite being based on the very robust SQL Server, struggles with the number of drivers Microsoft hosts on Windows Update. As of April, we know that Windows Update hosts 1,799,339 drivers, and this creates a 138 GB database that requires almost 16 days to synchronise down from the main servers. The WSUS server is brought to its knees, with frequent timeouts while it furiously tries to complete database queries. (The PC used is a Ryzen 5700G with 32 GB of 3600 MHz RAM and 500 GB of NVMe, running Windows Server 2025 and SQL Server 2022.)

↫ Adam Demasi

From this, it’s easy to understand why Microsoft would want to perform some housekeeping, followed by a new set of rules around only keeping the latest versions of drivers around in Windows Update. Demasi also notes that these plans by Microsoft won’t affect drivers for old devices, since they will still be served their “newest” driver version, and it won’t affect Legacy Update either.

The GPU in your computer is about 10 to 100 times more powerful than the CPU, depending on workload. For real-time graphics rendering and machine learning, you are enjoying that power, and doing those workloads on a CPU is not viable. Why aren’t we exploiting that power for other workloads? What prevents a GPU from being a more general purpose computer?

↫ Raph Levien

Fascinating thoughts on parallel computation, including some mentions of earlier projects like Intel’s Larabee or the Connection Machine with 64k processors the ’80s, as well as a defense of the PlayStation 3’s Cell architecture.

The world isn’t short of classic Macintosh emulators, but one more certainly cannot hurt.

Snow emulates classic (Motorola 680×0-based) Macintosh computers. It features a graphical user interface to operate the emulated machine and provides extensive debugging capabilities. The aim of this project is to emulate the Macintosh on a hardware-level as much as possible, as opposed to emulators that patch the ROM or intercept system calls.

It currently emulates the Macintosh 128K, Macintosh 512K, Macintosh Plus, Macintosh SE, Macintosh Classic and Macintosh II.

↫ Snow’s homepage

Snow is written in Rust and open source under the MIT license.

It’s been nearly a year since a faulty CrowdStrike update took down 8.5 million Windows-based machines around the world, and Microsoft wants to ensure such a problem never happens again. After holding a summit with security vendors last year, Microsoft is poised to release a private preview of Windows changes that will move antivirus (AV) and endpoint detection and response (EDR) apps out of the Windows kernel.

↫ Tom Warren at The Verge

After the CrowdStrike incident, one of the first things Microsoft hinted as was moving antivirus and EDR applications out of the kernel, building an entirely new framework for these applications instead. The company has been working together with several large security vendors on these new frameworks and APIs, and it’s now finally ready to show off this new work to the outside world. Instead of designing the new frameworks and APIs in-house and just dumping them on the security vendors, Microsoft requested the security vendors send them detailed documentation on how they want the new frameworks and APIs to work.

This first preview of the new implementation will be private, and will allow security vendors to request changes and additional features. Microsoft states it will take a few iterations before it’s ready for general availability, and on top of that, security software is only the first focus of this new effort. It turns out Microsoft wants to move more stuff out of the kernel, with anti-cheat software – more accurately described as rootkits, like Riot’s Vanguard – being an obvious next target.

Perhaps this effort could have some beneficial side effects for gaming on Linux, which you should be doing anyway if you want better performance, because Windows games seem to perform better on Linux than they do on Windows.

For a while now I’ve been wondering if Microsoft would blink when it comes to Windows 10’s rapidly approaching end of support date. Only a few weeks ago, Microsoft at the very least twitched by extending support for Microsoft Office on Windows 10, which should’ve been an indication of what was to come. Today, Microsoft actually blinked: regular consumers wishing to keep using Windows 10 after support ends in October will now be able to sign up for an additional year of security updates.

Microsoft is making this possible by allowing Windows 10 users to sign up for the Windows 10 Extended Security Update program for one year of extended updates, for free. This program is normally only available to paying enterprise customers, and this marks the first time the company is letting regular consumers make use of it. The “for free” requires some serious caveats, though, as depending on how you look at it, it’s not free at all. You options are to either pay around $30, pay 1000 Microsoft points, or to sign up for the Windows Backup application to synchronise your settings to Microsoft’s computers (the “cloud”).

This last option is technically free, but not only does the free tier include just 5GB of online storage, it also makes use of OneDrive, so if you’re using OneDrive to store your documents and other files you may need to pay for additional storage. On top of that, anything that requires the use of OneDrive is simply not “free”, and only allows Microsoft to further get its claws in you. If Sartre was alive today, Huis clos would’ve declared “L’enfer, c’est OneDrive” instead.

Regardless, it’s the stay of execution many Windows 10 users have been waiting for, even if it isn’t entirely perfect. Sure, choosing between an unmaintained Windows 10, Windows 11, and using OneDrive is about as pleasant as shoving shards of glass underneath your fingernails, and I have a feeling quite a few people are about to find out.

IceWM, the venerable X11 window manager, has released version 3.8.0, with some small tweaks, bug fixes, and changes. The theme menu now loads faster, the processor and memory graphs use less processor cycles, among other small tidbits.

Asterinas is a new Linux-ABI-compatible kernel project written in Rust, based on what the authors call a “framekernel architecture”. The project overlaps somewhat with the goals of the Rust for Linux project, but approaches the problem space from a different direction by trying to get the best from both monolithic and microkernel designs.

↫ Ronja Koistinen at LWN.net

Ronja Koistinen has done an outstanding job diving into this new operating system kernel and approach to kernel architecture, including its intended focus and goals. Head on over to the source and read it over there.

For a while, X terminals were a reasonably popular way to give people comparatively inexpensive X desktops. These X terminals relied on X’s network transparency so that only the X server had to run on the X terminal itself, with all of your terminal windows and other programs running on a server somewhere and just displaying on the X terminal. For a long time, using a big server and a lab full of X terminals was significantly cheaper than setting up a lab full of actual workstations (until inexpensive and capable PCs showed up). Given that X started with network transparency and X terminals are so obvious, you might be surprised to find out that X didn’t start with them.

↫ Chris Siebenmann

I did indeed assume X terminals were part of the ecosystem from day one, but it makes sense that it took a while, and that they didn’t enter the scene until X had established itself as the standard windowing system in the UNIX world. I’ve been trying to get my hands on specifically the last HP X terminal, but they’re hard to find and often very expensive. I’d love to get a taste of a proper networked X environment on real UNIX, in the way people actually used to use it professionally.

As a sidenote, Siebenmann is doing such an excellent job with these stories about UNIX, X11, and related matters. He’s like the Raymond Chen of the UNIX world.

Over the past few months, YouTube has been trying another round of anti-adblock measures. Currently the anti-adblock stuff is being A/B tested, and one of my accounts is in the experimental group. I wrote a filter that partially avoids one of the anti-adblock measures, fake buffering, on uBlock Origin (and Brave browser, since it uses the same filter rules). (It’s already in the default filter lists, you don’t need to manually add the filter.)

One thing that people have ran into is “fake buffering”, where videos will take a while to load due to a lot of buffering, but only at the very start of the video (there’s no mid-video fake buffering). As I’ll explain, the fake buffering is 80% of the length of the ads you would’ve seen, so even with fake buffering you’re still saving time using an adblocker.

↫ iter.ca

The battle between YouTube on one side, and users wanting a non-shitty experience without paying for YouTube Premium on the other, is unlikely to end any time soon. Your computer, your rules, so I’m on the side of the people wanting to block ads on YouTube – the same applies to OSNews if you don’t want to pay for our ad-free version – but I’m still intrigued to find out just how far Google is willing to go.

I sometimes see YouTube with ads at other people’s homes. It’s a nightmare.

It’s 2025, and we’re going to talk about BeOS, AtheOS, Cosmoe, and OpenBeOS, all in one news item, right here, right now, on OSNews.

In the very early 2000s, Cosmoe was a unique project that started out as a merger of the AtheOS userland with the Linux kernel. AtheOS, in turn, was one of the quintessential hobby operating systems of the golden age of the advanced hobby operating systems, the early 2000s. AtheOS would eventually be abandoned in 2002, but would be forked into Syllable and continue development until it, too, was eventually abandoned in 2012.

Cosmoe was the brainchild of Bill Hayden, and originally consisted of the AtheOS userland running on top of the Linux kernel, in order to address the lack of supported hardware a custom operating system kernel inevitably has to deal with. Not long after the start of Cosmoe, AtheOS was abandoned, as mentioned above, but a new project had entered the scene: OpenBeOS, now known as Haiku. Hayden switched gears, and instead started porting the parts that made up OpenBeOS to run on the Linux kernel.

This project progressed nicely, but in 2007 Cosmoe came to a halt (ironically, our last item about Cosmoe is “Cosmoe is back“) as Hayden had no more free time left to work on it, being a father of five, and so he decided to put the project on hold indefinitely. That is, until last year, when everything changed.

In mid-2024, my 3rd son Joshua, not even born when I started this project but who is now in college studying to be a programmer himself, had some questions about operating systems. I decided to dust off Cosmoe and see if I could get it running again, to show him what I had worked on. At first it would only compile and run on extremely old 32-bit versions of Mandrake Linux from 2007. But I had caught the bug again. Not only had I forgotten how fun Cosmoe was to program, but the intervening 17 years of progress made by OpenBeOS (now Haiku) made the certain aspects of this revival come at lightning speed. Day by day, week by week, I got it running on newer versions of Linux, and re-synchronized it with ever-more-recent releases of Haiku. After about 2 months of late-night effort, I had a version of Cosmoe that was 64-bit compatible, ran on multiple modern Linux releases, and was almost completely up-to-date with the latest Haiku source changes.

↫ Cosmoe’s history page

We’re halfway through 2025 now, and Cosmoe now exists as two separate, but related projects. There’s Cosmoe Classic, which is the updated and modernised incarnation of Cosmoe’s original concept: Haiku’s userland running on top of the Linux kernel. In its current form, it runs inside an SDL window on your Linux desktop, as there’s no native video driver. Cosmoe Classic, however, is not what Hayden is focusing on.

Instead, Hayden is focusing on the new Cosmoe, which takes the same idea – the Haiku userland running on a Linux kernel – but implements it in a completely different way:

Cosmoe is a C++ class library that allows developers to build rich, native Linux apps with the easy-to-use BeOS API. This library is a light-weight, serverless version of Cosmoe Classic which targets the Wayland compositor on Linux.

↫ Cosmoe’s GitLab page

What Cosmoe on Wayland (to differentiate it from Cosmoe Classic) allows you to do is run BeOS/Haiku applications on Linux, provided you are running Wayland. The project is in an alpha state, but once compiled, it comes with a few BeOS/Haiku sample applications you can run right on your Wayland-based Linux desktop. Hayden states that about 95% of the BeOS API is implemented in Cosmoe, with the TODO file giving an idea of what tasks need to be done to improve compatibility and implement other improvements.

The return of Cosmoe is certainly not something I saw coming, but I’m incredibly excited. I’m not entirely sure about the usefulness of running Haiku applications on Wayland on Linux, but who the hell cares – this is an awesome project, with a ton of cherished history behind it that gives me butterflies in my stomach. It’s absolutely beautiful to see a project like this come back to life in 2025.

Cosmoe is back. Again.

rou2exOS is a 64-bit DOS-like operating system (OS). The system is mainly written in Rust, but some portion of x86 assembly is used as well (inline + freestanding code for the stage2 kernel loading).

↫ Blog post about rou2exOS at blog.vxn.dev

It can do basic VGA operations, comes with a very barebones networking stack, realtime clock support, a FAT12 driver, and a few more tidbits. It’s a rewrite of the previous iteration of the hobby operating system.

Welcome news coming out of Jolla, the company that develops Sailfish OS. Up until now, if you bought their Jolla C2 smartphone, you had to pay a yearly subscription fee in order to get updates (with the first year included in the purchase price). Today they’ve announced their dropping this construction, and they now guarantee five years of free updates.

We’re happy to announce that from now onwards long-term Sailfish OS updates are included free-of-charge to all Jolla C2 devices for a minimum of 5 years. This applies also to everybody who have already purchased the Jolla C2.

↫ Announcement at the Jolla forums

People don’t like subscriptions, and I wouldn’t be surprised if Jolla was simply running into a lot of resistance to this subscription model from potential customers. Nobody likes subscriptions, and I think that counts doubly so for the kinds of people interested in buying a phone like the C2 with Sailfish OS.

Kian Bradley was downloading something using BitTorrent, and noticed that quite a few trackers were dead.

Most of the trackers were totally dead. Either the hosts were down or the domains weren’t being used.

That got me thinking. What if I picked up one of these dead domains? How many clients would try to connect?

Kian Bradley

It turns out the answer is a lot.

Accessibility is something that doesn’t get nearly enough attention, especially considering because not only will we need accessibility features eventually as we grow older, but also because a lot of accessibility features are just helpful even if you don’t technically need them. Given these facts, it’s a shame that accessibility is usually an afterthought, doubly so on open source desktops, a problem we recently talked about.

But what if you don’t just need to use a few applications as, say, a blind person, but also actually program as a blind person? Acidic Light, accessibility engineer at KDE e.V., has published a blog post about how screen readers actually work, and what it’s like to program while blind, and the conclusions are not exactly great.

I truly feel that, based on my experience with KDE and my experience actually delving into the weeds with AccessKit in a custom UI system, that accessibility programming just isn’t accessible. Unless you happen to already understand the way each platform works, trying to find resources on how to actually let a screen reader know your UI exists is just painful. It’s going to involve reading code other people have already written. It’s going to involve hours, if not days, if not weeks of research and painful debugging. You likely won’t be able to ask many people for help, because they’ll know as much as you do.

↫ Acidic Light

If the people who know most what is needed to make a program accessible have so many problems actually making programs accessible, because the tooling, documentation, and institutional knowledge just isn’t there, what hope do other programmers have to make their code accessible? If a blind programmer can’t scratch their own itch, so to speak, we’re never going to reach a point where accessibility becomes a given.

I’m very happy awareness of accessibility is growing, but I feel like this isn’t the first time we’ve seen an increase in accessibility awareness only for it to eventually fizzle out without meaningful improvements for those that need it the most. I really hope it sticks this time.