draft0

MenuetOS

This is a really interesting one. Or two. I'll start with MinuetOS. Written in Assembler for 586 systems, open source, very impressive, performance- and size-wise. Including the included applications it fits in 1.44 MB. The boot time on my Core2Duo isn't easily measured because it's completely done before my monitor adjusts the resolution from the boot menu to the desktop. But when I chose the same resolution it uses for its boot option screen, I learned that it boots up completely pretty much instant even if the monitor doesn't have to switch modes. A guess writing 100 % in Assembler makes this possible. Very promising software, but a small community. The developers have been very dedicated when the 32 bit version was still under active development. So it is very stable, packed with useful little applications, especially for development and debugging, but also for daily use and casual gaming. This 32 bit version is not actively developed anymore.

There is a 64 bit version, which is not open source. I'll mention the reason in the paragraph after the next paragraph. It's where the development of MenuetOS is happening nowadays. I don't know how far it has changed since the 32 bit version. I didn't try it, because I'm looking for a 32 or 16 bit system. So, the source is not available, but the information available should be suffice to write drivers and applications for it. Edit: I've tried the 64 bit version briefly. Since there has been years of relatively active development between the last (pre 1.0) 32 bit version and the latest 64 bit version, the changes and advantages are very noticable. If a closed source OS is an option, the 64 bit version is surely what you'll prefer using.

There is a CD image of ~ 22 MB of additional applications that I didn't really try. But it's worth mentioning that there are more than the very basic applications available. There's a media player, even some networking tools (FTP, telnet, but no SSH, as you might have guessed). There is a very rudimentary shell. But it is obvious that the developers' focus was on GUI programmes, as even the ping tool is used by opening a window, entering the target in a text field and starting the ping with a button. Most things worked flawlessly during the time I've tried MenuetOS. But there are a couple of things that may be considered annoying by a demanding user. The most obvious to me was that the mouse is very hard to use. Stopping the pointer in a particular spot takes practice. It seems to drag behind quite a bit. It takes time to do something with a mouse. And when setting the speed slower the pointer sometimes starts moving into the opposite direction. Not completely a matured OS. But many parts are matured and it is usable.

KolibriOS

And then there is KolibriOS, a fork of the 32 bit MenuetOS. It doesn't claim to be a fork, though. Its heritage isn't mentioned on the web site. Just once in the web forum by somebody asking about it. Every mention of it in the source code has been removed, but copyright notices have been added by the "new owner". That makes this fork a pretty vile act. I don't know what the motivations behind this have been. It is pretty clear that the KolibriOS developer didn't start KolibriOS from scratch, no matter how much they try to make it look like it. There are people who learn about and get into KolibriOS who don't know where it came from. So, after this fork the developer of MenuetOS has decided to not publish the source code of the new 64 bit version of MenuetOS. I felt it relevant to get into all that because making a decision for KolibriOS and against MenuetOS can be seen as a political statement. (And I'm not saying this in any way as a reference to current international political happenings. KolibriOS is sometimes called "a Russian OS". But people who speak out against using KolibriOS because it's Russian and therefore evil should either produce some sort of evidence, hint or at least rationale for why the OS or its developers have any connection to the war in Ukraine or shut up.)

So, back to the OS itself. It's not fully compatible with MenuetOS. As with previous versions of MenuetOS, some programmes still run, some have problems and many just don't work any more on the new version. I don't know if there are even some programmes left that have run on both systems.

There are still 1.44 MB images for floppy disks being published. But for my first impressions I chose the large image that comes with many more applications. That image is over 130 MB, which is likely mainly due to the applications that are not written in assembler. But still, a multiple of the CD image with additional software for MenuetOS. I'd have to compare them in detail to be able to say why that is. The set of applications that is included in the large KolibriOS image is very good for basic requirements and significanty surpasses what's usually included in OS images. There are editors, file browsers, development tools, quite a few casual games and a few not so basic applications. Some not written in Assembler just for this OS, like DOSBox and Netsurf, others impressively small and fast. Netsurf being the only web browser that's available for the OS makes clear who might want to consider using this OS and who might not.

The whole experience of using KolibriOS was a bit smoother than using MenuetOS. That is, almost everything was perfectly smooth, nothing ever crashed. But, some of the not so basic applications are more resource hungry than what you would expect from an OS written in assembler (which is because those applications aren't written in assembler and probably ported from another platform with functionality as the only requirement). For example, the video player would use hundreds of MB of RAM for the playback of some video files. The whole system would become less and less responsive, even though there were still GBs of RAM unused. I don't know what codex' are implemented and how. But some formats wouldn't play at all, bringing the entire system to a halt until I was able to kill the media player.

The size of the complete KolibriOS image (way over 100 MB, as compared to under 25 MB for MenuetOS) shows how the goal of keeping everything small and fast has been overlooked more and more when porting software to run on KolibriOS. But also that more application have been ported. Those are not part of the OS if you use the normal floppy or CD image. But there is one with everything pre-installed. The fact that those extra applications don't run as fast and aren't as small shouldn't be a problem and isn't a fault of the OS. But the fact that no alternatives have been written specifically for Kolibri might be restricting its use in practice.

The following screen shots are all of KolibriOS.

Solar OS

Solar OS is a one to few person project. An OS with a GUI with an early 90s feeling. Closed source, if I didn't miss something. There is an API reference, but not much help beyond that. I didn't see a software repository mentioned anywhere. I guess the most likely usage is with your own custom applications or ported software.

Solar OS's advantages are definitely its small size (< 1.44 MB) and how fast it runs (a 386 with at least 8 MB RAM should be enough). I booted it from a USB flash drive (using grub and Syslinux) on an Athlong 64 X2. It was as fast as it could be, of course. Although, movement of the mouse pointer is very laggy with screen resolutions higher than VGA. With HD resolutions it can't really be called usable with a mouse anymore. I don't know if that's a driver issue with my graphics hardware or what. This is not a problem with other systems that use a simple VESA driver.

But, there wasn't much to do to test how well, fast, comfortable, beautiful and error-free everything runs. There aren't many applications included. The things that are included do a good job of demonstrating the OS. There are things like a process manager, a picture viewer, a very simple text editor and other tiny demo and system applications. And there is a help decument, debugger, things like this. For a proper test, I'd wish for more applications to use for practical work. I don't know what exactly. It depends on what you need the os for. Maybe writing your own application would be best included in a test that aims to collect for experience about the OS.

It crashed after less than half an hour trying it out. Because it stopped responding before I made any screenshots and I didn't yet bother to boot it up again, see the screenshots page on the project web site. There are many screenshots that paint a good overall picture of the OS and its GUI.

Visopsys (VISual OPerating SYStem)

Visopsys is a real hobby OS created from the ground up. It's very likeable like this. It has a GUI that is kind of a mix of a MacOS-9-style and Windows-95-style desktop. There are a couple of accessories and a limited shell. A very promising OS from what I've seen and read. But also limited in its usability as of now, from my own experience.

I've tried Visopsys 0.91, the currently latest version on the download page. Running it in a virtual machine is recommended because of the limited hardware support. But I think that it is fair to hold an operating system to a standard that requires it to be able to run on real hardware, not on another operating system. I tried it on five computers before I got it to boot. I'm not going to list all the hardware components (I'd have to look up all the motherboards - meh), but the CPUs to get an idea for the age of the system components. First a Core2Duo. It printed bunch of error messages to the screen that I didn't further investigate. Initialisation failed and stopped. Then an AMD K6. I got a message informing me that Visopsys will boot in text mode because of a lack of SVGA support and then it rebooted. The used VGA card is a very old one. So that might be correct. Then I tried an Athlon 64 X2. It did something, the loading bar appeared, then it rebooted. I didn't get anything else out of it. Then I wanted to give The K6 a newer graphics card. But while digging for one, I found another PC with a Celeron E3200 and tried that first. And it worked. So I didn't bother checking the K6 again.

Boot time to live mode is about 3 seconds. Incidently, I also learned something about booting from CD. Booting from any ATAPI CD or DVD drive took many times longer than booting from the same CD in a SCSI CD drive. Like 10 times longer. That surprised me. I didn't think there would be such a big difference. Maybe my IDE cable has a problem or I did something wrong. Anyway, booting from USB or an IDE hard drive was quick enough to not bother to measure how quick exactly (about 3 seconds). Edit: A likely explanation for the load time difference between the SCSI and the ATAPI drives is that the ATAPI drives were all modern, fast-spinning drives that take a long time to spin up before any data is read, as apposed to the old, slow SCSI drives that I have, that allow reading from the disk much sooner.

Installing to a hard drive is straight forward: Optionally choose a language other than English (this also changes the keyboard layout), coose a partition, click a few next buttons, enter a password for the default user account and reboot. There is also a partitioning tool available from the installer. But I didn't find a way go get back to the installer without rebooting. After booting the installed Visopsys and logging in, or after booting in live mode, you get a simple desktop with a few icons and a task bar at the top.

There isn't much to do. As far as I know the included programmes are pretty much all that are available for Visopsys. (Please do correct me if I'm wrong!) There are system settings, a very limited shell with a limited number of commands with very limited options, a few casual games, an image viewer and a few other simple tools. The most useful things included apart from the file browser are probably the text editor, the calculator and the telnet client. There is no web browser and no multimedia applications.

During the time I used Visopsys, nothing crashed or went fundamentally wrong. But I noticed a couple of small things that would need to be sorted out to create a comfortable user experience. The most obvious one is that while dragging, the dragged objects sometimes were lost and dropped too soon. Especially dragging windows vertically has to be done very, very, very, very slowly. It's as if the mouse button had a loose contact otherwise. Another thing I noticed when choosing a new desktop wallpaper. Once the directory in which to look for image files to be used as wallpaper has been chosen, the file names seem to be cached. Image files that I've copied into that directory after that were not listed before rebooting. Little things like this add up to create a bit of an annoying experience.

But I don't want to talk down the state in which the project is in. Many things are working slawlessly and depending on what you need your computer to be able to do, it might even fulfill many of the needs. A web browser is not one of them, though.

I've attached some screenshots that I took. The one where the Snake icon is selected. should have a window with the Snake game in it as well. I don't know why it isn't invisible.

There are too many Linux distributions to list and too many Linux-based systems to try all of them out just for fun. There also wouldn't be much of a point to doing this. And there's nothing to say about Windows or MacOS in this entry. But I'd like to write a bit about operating systems not very many people have heard of. Specifically, I took a look at several lesser known desktop operating systems. This entry is like a forword to my coming entries about those alternative PC operating systems.

The less big OSs

There are a few other operating systems that aren't used on as many computers, aren't as well-known as the big three (Windows, Linux, Mac OS) and their derivatives, but are still important for their use case or niche. There's the extended family BSD, MINIX, the PlayStation system and many other Unix-like systems, there's Illumos and other actual UNIXs, a variety of DOSs and many even less widely used OSs for small niches. Those all deserve an article. But this is not one about any of those.

The discontinued ones

There are many UNIXs and Unix-like operating systems that aren't actively developed or not even officially supported anymore but still in use in various industries. (Never touch a running system!) There are many old systems that are still sometimes used by retro computing fans, users with nostalgia attacks and people who just never updated their machines since the 80s or 90s. I'd like to write an entry about those systems. But not right now.

The unnoticed ones

There's also a huge landscape of operating systems on embedded systems that are not primarily seen as computers, such as DVRs, cars, household appliances, toys, parking meters and many many industrial tools and machines. An article about how this landscape of proprietary in-house systems turned into a stable of Linux systems would also be worth writing. Or one that takes a look at some sleek systems, from tiny, incredibly efficient environments for 90s microcontrollers to larger systems like Palm OS. But my motivation for this entry right here is about something else.

The interesting ones

And there are operating systems that don't fit any of the beforementioned categories, but still have a reason to exist or at least had a reason to be made at some point. Independently developed OSs for personal computers that emerged from a single person or a small group of computer enthusiasts as learning projects, to prove a point, or as a recreational coding project. Those are the OSs that I'm interested in this entry.

For a long time I thought that such operating systems don't exist, don't reach a state in which they are worth checking out or aren't shared publicly. I did wonder why nobody seems to write alternative OSs just for the sake of it. Maybe the landscape of existing OSs and the tiny adoption of any OS that isn't one of the big three (or four, or five, depending on where you make the cut) takes away any motivation to start yet another small OS from people who would otherwise pursue such an idea, I thought. But it turns out that I didn't look closely enough to find them. There totally are working alternative operating systems that were made for the sake of it, despite Linux satisfying all requirements of the project's initiators in most cases. Many of them may have 0 to 1 users and some of them may have never and may never be used as anybody's main OS. But some of them are very interesting systems. I'd like to point out a few of those.

My initial intention with this entry was to write a paragraph or two about one or two handful of OSs. But I found that there is more to write and more projects to share about this topic for several different reasons. I will write separate entries about a dozen alternative operating systems when I find the time to check them out myself. I'll use the category Operating Systems for these entries. Since I found more than I thought I'll restrict myself to testing only systems that check the following boxes:

Those are critera that I thought of and wrote down after searching and finding most of the projects I'll write about and after deciding which ones I find interesting. And I might ignore or change those criteria further when I feel like it. They are here just to give an idea on what sort of projects I'll pick out and write about.

My impressions/conclusion so far

There are more alternative desktop operating systems than I expected. I thought after the 90s the landscape has become very flat. But development of a number of commercial and non-commercial operating systems went on for much longer. There have been many large projects I had never heard about. Some of those have been officially ended over the years (like eComStation and Syllable Desktop), others are still worked on (like Redox, MenuetOS and MorphOS). I stumbled over so many names of OSs that are in a state in which they're not really usable as a daily desktop OS, that had been published on some now offline website or that have been discontinued years ago. I thought I'd try out most of what I can find and write an overview of the alternative desktop OS landscape. But that would be a project in it's own that would be huge and take far more time than I want to invest in it. Many names worth mentioning have not been mentioned here and many cool projects also won't be mentioned in any of the articles to come. For example I ignored commercial OSs, like ArcaOS. I just don't want to pay 129 $ just to test a system that I'm likely not going to use.

I can already, after having only briefly tested a few systems, say that there are usable, stable systems that hardly anybody has heard of that are as good as any commercial OS from the 90s. That's good enough for most desktop users. Each with their own set of features and unique selling points, there are real decisions to be made when selecting an alternative operating system that go beyond the questions one asks when selecting the next Linux distribution during distro hopping. I am happy that there are more than one or two systems in a usable state that focus on supporting old hardware and/or using less hardware resources than we got used to with Windows, Linux and MacOS. Whether you're looking for something to run comfortably on an 80286 or something that runs ridiculessly fast on a modern 64 bit system. There are systems to choose from.