Why no new operating systems?

UNIX is 50 years old this year. Windows is 33 years old. OSX is 30 years old (if you consider NeXTStep to be proto-OSX). Linux, a derivative of UNIX is 28 years old. IOS is based on NeXTStep, Android is based on Linux.

Why have there been no real new and innovative operating systems for 30+ years? There have been some interesting contenders (BeOS for one) but nothing really broke through.

People buy operating systems to run applications. It would be a truly huge cost to convert the large mass of applications to run on a new general purpose operating system. So programmers simply change the existing operating systems to meet new needs.

  • As they’ve developed over the years, they’ve all converged into being more or less the same WIMP (windows, icons, mouse pointer) and the differences are mostly window dressing

  • as they’ve developed over the years they’ve grown bigger and bigger and added more APIs, libraries, and functionality. anything new would be starting from scratch while the established players have a multi-decade headstart

  • any new OS would really have to be something remarkably different in order to even hope to attract users away from the established platforms. Not only that, it would have to be useful as well.

IOW, it would cost you a lot of money with no guarantee you’ll attract a user base.

Software development on any large installed base is almost necessarily evolutionary rather than revolutionary.

It’s absurdly expensive to port applications to a new OS, so OS changes almost always keep around the good stuff and add on, and only very slowly drop old compatibility.

Also, honestly, the fact that Android is based on Linux which is based on Unix doesn’t mean that they are the same OS. I’d be surprised if as much as 1% of the source code that is used to ship modern Android is in common with Unix from decades ago. It’s a ship of Theseus.

To a certain extent, I think we’re seeing a distributed development of the web application operating system. There are common elements that are starting to be used somewhat universally (the icon that looks like a head to click on for user profile, the hamburger icon for configuration, the gear for settings, envelope icon for messages). Maybe we’re going from needing an operating system to present to us the common elements of computer use to web developers using derivatives of those common symbols for elements in web apps)?

Most of the from scratch OSs are also in embedded system and their numbers greatly outnumber the general purpose computing world.

As those who remember the UNIX wars will attest that calling all of them the same “operating system” is a bit of a stretch too. OS like MINIX, Plan 9, or Window NT are very very different even if they all did have a Posix layer. Every modern Intel CPU runs Minix 3 OS as an example but MINIX 3 was first released in ~2005.

People create novel OSs all the time it is just that they often provide no compelling value to many users except for perhaps the type-1 hypervisor like Xen which could be argued to be as much as an independant os as a kernel like Linux.

Windows as a brand may be over 30 years old but the most current Windows OS most assuredly is not. I’d be shocked if there is much shared source code at all between whatever is the latest release of Windows 10 and Windows 1.0. Actually, there are already significant changes that have been made in Windows 10 since release.

Likewise, while some parts of the Linux Kernel are probably decades old, much if not most of it has been updated/replaced over the years.

And that’s one answer - we can’t assume having the same name on the box means the contents are remotely similar.

There is also the fact that we have now mostly works most of the time. Where is the incentive to start over from scratch?

Well, there are two “Windows” families of note for desktop/laptop:

  1. “Windows on top of DOS” = Windows 1.0-3.11, and the later (less DOS-reliant) Windows 95 to Windows Me, and

  2. Windows NT, starting with NT 3.1 (an arbitrary start point to be aligned with the consumer version) all the way up through NT 3.5, NT 4.0, Windows 2000, XP, Vista, 7, 8.x, and now 10.

Windows NT is almost 30 years old, but it does share some code still with older strains (e.g. Win32 APIs.)

You could say that about nearly any innovation in history. Where would we be if the Wright Brothers said, “Well, trains work most of the time. Nobody will want an airplane.”

We could use a reliable, easy-to-manage consumer-oriented OS. Windows has good features but is onerous to maintain and hell to troubleshoot. I have not used Linux on a PC but I don’t think it’s something your grandmother would use.

we do. it’s called iOS.

Ha ha ha! jz78817 made a funny joke.

There are new operating systems being developed all the time, as said rat avatar. You have your Inferno OS, Amoeba OS, Fuschia OS, L4 OS, Fluke OS, you name it. Haven’t heard of them? There is a difference between an OS written for research purposes, and something proven for decades that you are going to pick to operate your spacecraft or critical infrastructure, or even just on your desktop or laptop computer. For your space shuttle you are not going to pick something completely experimental, and for browsing the web and playing video games Microsoft Windows and Linux work fine, but good luck trying to get your hardware to work with something exotic. Over time, even semi-abruptly, new ones can work their way up the popularity ladder.

Any new system inevitably involves trade-offs: planes, for example, can’t carry the volume of material a train can, and they’re a lot more expensive per pound of passenger or freight.

What trade-offs would be necessary to get to this “reliable, easy-to manage consumer-oriented OS”? For example, a lot of the complexity of modern consumer-oriented apps is because of all of the plugins and add-ons: a website may use Flash, PDF, and several different audio/video codecs, but managing all of those isn’t easy. If your new OS won’t run the nifty app that Aunt Susie wants to run (Angry Birds?), then it’s not oriented towards her kind of consumer.

There are plenty of other OS’s around but none of have gotten much traction. Everything from Plan9 to Harmony. Harmony will likely be a real player soon in terms of absolute numbers.

There is a heck of a lot more to an OS than what picture you put on an icon to represent what option. Even if those web applications have similar-looking buttons, they will look very different under the hood, and none of them is anything resembling an “operating system”.

Oh, and
[Moderating]
Most of the answers to this are going to be matters of opinion, so let’s move it over to IMHO.

There is also the question of when are different, but related, software stacks different operating systems? Android, ChromeOS, Debian, RedHat, and Archlinux all use the Linux kernel, but are they all the same OS?

Debian, RedHat, and Arch I would all call the same OS. They have different package managers, and do some things a bit differently, but they basically run all the same software at the kernel, library, and user levels. Generally, people just refer to these as Linux distributions, and most people know they’re pretty much the same, until you get to the details.

Are Android and ChromeOS just Linux distributions then? They certainly run the Linux kernel, but that’s pretty much where the similarity stops. There are goldfish, zygotes, and other strange Android things that never appear on a traditional Linux system. The counter argument is that because they share a kernel, they can even run the same binaries, so of course they’re the same OS. But really, are my phone, computer cluster, and car all running the same OS, just because they have the same kernel? (the answer might be yes)

If I remember correctly, iOS and MacOS share a kernel, so are they the same OS, too?

Depending how you judge things, in the last 10-15 years we’ve had the introduction of Android and iOS. Both of which went from non-existent to running on billions of devices.

The future might see the rise of Google’s Fuchsia OS. It might be based on Linux, it might not, it might just be a research project, it might unify and replace Android and ChromeOS, or just fade away to the Google graveyard.

No, I agree. The OS performs a lot of different functions but one of the big ones is to give the user some consistency in the environment. I’m not saying that web apps are going to replace the OS but some features of web applications are becoming similar to present to users the consistency that used to be performed by the OS.

I have customers that do pretty much all of their work within web applications so the web browser and web apps are performing most of the UI functions for these users. Again, this is not replacing the OS but it is replacing some of the common OS functions. Some functions of the OS are becoming more distributed over workstation OS, browser, web app and server OS.

I figure it’s because the existing ones do all the functions that OSes do, and are extensible enough to support stuff in the future like VR interfaces, etc… I mean, what’s JoeBob’s x86 OS going to do that Windows, iOS and Linux don’t already do?

And even with the case of Harmony, the only reason it’s coming about is because of Chinese concerns about being forbidden from using Android in the future. It’s not like it’s going to do something Android doesn’t.