Key findings
- Some developers are trying to revamp the operating system design to reduce legacy issues and improve cybersecurity.
- Existing operating systems contain decades-old code, which can result in bloated, disjointed software.
- Attempts to replace common operating systems such as Windows, macOS and Android have been unsuccessful due to the inertia of existing systems.
Every operating system in use today is a direct descendant of a previous system. But software is not biology, and there is no ironclad rule that says the “genes” of previous systems must be present in the operating systems of the future. So is it time to reinvent the wheel, or is this just a fundamentally stupid idea?
This has all happened before
Of course, people had to invent operating systems at some point, so there is no infinite regress here, but in a world where some operating systems have established themselves in the market like a software Mount Rushmore, there are still some brave souls who try to do something different every now and then.
An interesting example is Google's Fuchsia operating system. Its source code repository simply appeared online one day and made its debut on the Google Nest Hub device a few years later. Fuchsia uses a new kernel called Zircon that is neither Linux nor Android related and, while inspired by Unix, does not function like Unix and is not a Unix-like operating system. There has been some speculation that Fuchsia was intended to replace Android and ChromeOS, at least on Google devices, but so far this has not happened.
Microsoft's Midori is the other big example that comes to mind. Midori was a secret project at Microsoft that was part of a plan to completely replace Windows. Not much is known about Midori, but it seems that its basic design was completely different to Windows, and that cloud technology and the ability to run programs distributed across multiple hardware nodes were built into the basic design.
Source: X (WalkingCat)
https://x.com/_h0x0d_/status/1819724702555988282
There is also ReactOS, which is not a completely new operating system in the traditional sense, but an attempt at a reverse engineering project for Windows. In other words, ReactOS does not contain any Windows source code, but is intended to work with software and drivers written for Windows, but is completely open source.
So far, neither these nor other attempts have succeeded in displacing Windows, Linux, macOS, Android or iOS and their variants from the mainstream, but there may be good reasons why it is worth pursuing this approach.
Operating systems bring a lot of ballast with them
The operating systems we use today are complex systems built on codebases that took decades to develop. Despite the shiny new interfaces and bolted-on features, their core consists of quite a lot of decades-old code. This isn't just due to laziness, of course. If something still works, why change it? Not to mention that newer versions of the operating system are supposed to run software that was designed for older versions.
However, all that baggage can lead to bloated and resource-hungry software, or an operating system that feels disjointed. Windows 11 is a prime example of this, with many new features and UI elements duplicating things still present in older parts of the operating system. The venerable Control Panel has been part of Windows since version 1.0, and even after Microsoft finally hinted that it would be removed from Windows, it changed course. Go one step deeper than the first level of Windows 11's context menu and you're right back in the era of Windows 95 or 98. I don't want to talk about Windows specifically here, as all established operating systems are generally like this.
If you could clean up the mess and create a new operating system that didn't lug around various anchors, you could build it on modern hardware, software engineering theory, and everything learned since the dawn of the computer age. It would be painful in some ways to lose all that heritage, but starting over with a better foundation for the future has its advantages.
A new operating system could be much more secure
Source: Unsplash
In the early days of computers, operating systems, and the advent of services like email, the concept of cybersecurity was still very vague, if anyone was even thinking about it seriously. After all, who would have thought that every member of the public would one day have their own computer? Who would have predicted that people would network all of those computers together? Not the people who laid the foundation for the technologies that underpin our modern operating systems.
Much of cybersecurity involves fixing unintentional vulnerabilities in outdated code, adding extra security measures like HTTPS or email encryption because they were not part of the fundamental DNA of those protocols. Exploits of outdated operating system components are discovered all the time.
For example, in 1996, Microsoft made a significant change to Windows NT 4.0 by moving the window manager, graphical device interface, and video driver to a kernel-mode driver called win32k.sys. Keep in mind that after Windows 98, Microsoft abandoned the DOS-based Win 9.x family and moved all of Windows to Windows NT. NT 4.0 was the direct predecessor to Windows XP, and every Windows system up to 11 today is still based on Windows NT. That change in 1996 was made in an attempt to improve the performance of the then-limited hardware, and today win32k.sys is used in various exploits with potentially serious consequences.
This probably won't happen
While I think there will always be attempts to break with existing operating systems, their inertia makes it quite unlikely that there will be a revolution instead of the expected evolution. Perhaps one day, when our robotic probes reach alien civilizations, a piece of Linux kernel source code will be the first thing those people ever see. Hopefully the code comments aren't too harsh.