This is really not entirely accurate. The original Lisa/Macintosh used the CPU to do most of the work – Macs came without graphics or sound cards and handled the GUI just fine on the back of the CPU. Windows relied heavily on the graphics card for its GUI, and OS X did the same, implementing all manner of eye candy therewith.
You look at the contemporary ARM-based SoCs: they have a variety of logic blocks that would have been the support chips and graphics cards on older devices. Apple puts a so-called “neural engine” in their recent SoCs to support facial recognition, something eight cores working their hardest together would struggle with.
To say that CPUs have gotten powerful enough to handle all the workload of modern devices is simply not correct.
Your response doesn’t fill me with hope for new exciting operating systems, but the IT pendulum swings back and forth all the time. Maybe there will be a need for something we can’t even dream of yet.
This is excellent and just what I was looking for. Thank you!
Google did a good job creating Android. It is built on Linux but the user interface is entirely new and designed for phones.
I tried learning Linux a couple times and always ended up frustrated. There were so many duplicate programs. There’s even two different help systems. I lost count of the desktops. There’s at least six that I remember.
I remember (1998) losing a full day trying to edit the config files to set up my DSL and internet. I finally took the pc to our Linux support guy at work. He had six years experience in Linux and it took him a couple hours to get DSL setup and working. A major PITA. The PC did work well after I took it home.
Interesting point. In the sense that your browser supports application launching and task switching, it’s fairly reasonable to call Chrome or Firefox an OS analog. Obviously, at the moment, they’re pretty bound to the underlying features of the OS to provide a lot of features (principally I/O) but there’s really nothing from a user’s standpoint that the OS offers that a browser doesn’t as well.
Obviously, that’s the whole motivation behind Chrome OS, though I don’t know how tightly they attached the browser to the kernel. (I assume that it just runs as an application in an Android derivative OS.)
Modern browsers are p-code systems, an idea that’s been developed constantly (if not consistently) since the 1960s but is always seen as too inefficient for real-world use by a certain set of managers and developers: They run architecture-indepdendent code from the outside world in a sandboxed environment which provides a rich native API but very little access to the underlying system in order to enforce a security policy. Sun tried to do this with the JVM and Java applets, but they proved clunky enough that they didn’t catch on, except in all of the places they did, of course; really, p-code systems in general are seen as inefficient failures except for all of the places they’ve become ubiquitous and sufficiently efficient to gain widespread use.
I don’t really miss OS360 but SWylbur never worked weil on DOS or the WINs I tried.
$700k? Piffle. I read that the MEDIAN pay for FazeBook employees in SanFran is US$240k. A few * ricos* may skew the curve, but this still indicates many in the half-million-plus range. $700k? Feh.
It’s like I’m watching the automotive industry circa 1900, before steamers and electrics were vanquished by petrol and diesel ICEs - blame electric starter laziness. The world yet evolves. PREDICTION: Implants linking human consciousness to the Net will be here soon. What OSs can/will integrate us? How hackproof will they be?
As has been said already; all those legacy applications. I saw the switch from 8-bit to 16-bit programs on the IBM PC, and that took a while. Or old CP/M programs reworled for the PC. Then the big switch to Windows. It used to be that there were dramatic changes every few years, but from Windows 2000 onward the OS have been similar and the apps as well, with the proviso that applications needed to be upgraded because of internal changes in Windows, such as .NET. The recent incarnations of Windows have been aimed more at making them more Web-oriented and able to work well on tablets and smartphones. I mainly process text on a PC, and Windows has not got any better over the years. The same for many applications; upgraded over the years, with some bugs still not fixed. But I can live with it, and I am not prepared to dump my entire suite of software for something new. Which is the main reason I have not gone over to Linux.
Linux is the obvious alternative, but not exactly easy to get started with. However, as a free system, it parallels the situation in the PC world where many free apllications are available. But while Linux is free and constantly updated, it has not gone beyond a niche player even with big companies. I wold love the idea of presenting Redmond with the rigid digit and going off the M$ grid and living free, free of the endless paid upgrades, but the true cost comes in the time to learn the new system and to find substitute programs. At the end of the day, what do have that is better? windows is reasonably relaiible these days, and so are most application programs.
And another issue. It is becoming increasingly less important which system you use because some much now is done by interacting on the Web. These days you spend most of your time working with the browser instead of the OS.
Linux may still be unpopular as a desktop OS, but it is the most common server OS, to the point that Microsoft started including a version of Linux as part of Windows. It’s popular with power users as well, which is why Microsoft is integrating a lot of the functionality of Linux programs like bash into Windows 10.
I also have seen the desktop OS as popular with people who are turned off by the problems with Windows 10, but are worried about staying on Windows 7. You have Steam’s Proton and SteamOS for gaming, and Wine for other Windows apps you might need. While it’s usually the more technically minded power user who does it, it is increasing in popularity.
But, yeah, there’s no real need to make more OSes when you have the paid one everyone uses, and the most popular free one that people can update to add new stuff to. The rest of the OSes are more niche for specific uses, or just for fun. Since everyone is using ARM or x86, and all the OSes run on them.there’s not much need for the constant development we had back when CPU choices were more diverse.
It was oddly Apple who created a “new” OS in building iPadOS, though it’s really just an extension of iOS with a different shell. Still, they can only pull it off because they control the hardware.
Bash is 100% GNU, not sure why you credit it and its ilk as “Linux programs”; like you said, they run on everything including Windows.
Also, not “everyone”, even excluding the niche users you dismiss, is running ARM or x86, even to run Linux, and not all operating systems run on those platforms- especially z/OS, VMS, and similar (hardly brand new operating systems!) But that is more a question about why there are not more cheap processor alternatives; a mainframe is one thing, but, as someone pointed out in another thread, there was a time when I could get a Sparc or Power laptop for a reasonable price, but not right now.
Yabbut z/OS (or whatever IBM is calling it this year) is basically “invisible;” unless you are doing something in a 3270 window (which I do sometimes) you almost never know it’s there. VMS is rather niche at this point and isn’t going to make any inroads into broader use.
Though you could say VMS has, in a way. The underbits of Windows NT are/we’re quite similar to VMS thanks to being spearheaded by Dave Cutler after MS swiped him from DEC.
Just stumbled into this discussion and reading it was quite interesting, although there seem to be a lot of people talking past each other and missing definitions.
I have thought the same 2 years ago, having grown up with Windows. I’ve been using Linux for a good year now and am by now convinced that I could set up the computer of my Grandad with Linux and he wouldn’t even notice. It is much better today than most people think, and btw I also game regularly and can play most games on Linux without performance losses.
Simple version: because the version of Bash included in Windows 10 runs on top of the Windows Subsystem for Linux. It is thus the Linux program Bash.
I am also unaware of any port of Bash that runs natively in Windows without some sort of compatibility layer, like WSL, CygWin, or MinGW. This makes sense as the GNU utilities were created to run on Unix-like OSes, even before Linux came around.
Finally, while Linux was created outside the GNU Project, it was the first truly free OS, and thus was the first OS under which the GNU goal was met: it ran the free software utilities on top of a free software OS. GNU/Linux, as the combination was sometimes called, was thus endorsed by the project.
Today, we usually say Linux to refer to GNU/Linux, as the two are so intertwined.
Yes, I was talking about portable code being widespread in application development. I was not referring to operating systems like Windows, MacOSX or Linux.
Android is relatively new as “OSes” go. It’s technically only 12 years old.
Android uses the Linux kernel and the Linux filesystem and most Linux shell commands work as expected. And the C/C++ portions of it are generally built by GCC.
So it has old underpinnings. And in many cases the device drivers for an Android computer are the same ones that would be shipped with Linux. (as expected, since Android is based on the Linux kernel)
And Android uses SELinux, a kernel level security mechanism.
With that said, unsurprisingly, these underpinnings are both very reliable and in some ways trash. The Linux kernel is both extremely well tested and reliable and also very difficult to read and maintain as it’s mostly all written in C, with some non-standard additions that means you must compile it with GCC.
The problem with it fundamentally is that it’s very, very, very complex and fragile and most of the complexity is not required to be in the kernel itself or necessary for the operation of a specific computer.
The obvious modernization effort is to rewrite just the kernel as a microkernel design, with feature rich frameworks to support sanely written, reliable device drivers. Fuschia is partially an effort towards this end.
With Fuschia + Android, this would technically be a young, new OS, with all components of it developed after 2005. Almost all Android apps would continue to work without modification.
ah yes, the geek obsession with “pure” microkernels. Purity always gives way to real-world requirements. Windows NT started out as more or less a microkernel, but performance shortcomings pushed them to move more stuff into kernel-mode. Ditto NextStep/OS X/macOS. and Linux is used everywhere while MINIX remains a toy for students.
QNX. It’s a microkernel and is widely used in some sectors. Device driver code is unable to crater the entire system which is crucial because next year’s hardware needs new drivers which means new bugs.
The other reason for microkernel is it is possible to formally verify the kernel/asil D certify it. The ONLY OSes with this cert are microkernel.
A microkernel isn’t necessarily any faster. The reason to use one is to isolate the critical code needed or the system crashes from the supporting subsystems that can be restarted if they crash.
So reliability and confidence in the code, not speed.(Linux kernel is immensely complex and so while it is highly reliable anyone inspecting it will realize it isn’t something you would want to be your life on. Autonomous cars, etc, are bet your life mission critical systems.)