We report regularly about Visopsys – one of the few hobby operating systems that survived where virtually all the others (SkyOS, Syllable, etc.) died out. They’ve got a new website, which seems like a nice occasion to give it some attention again.
The bulk of Visopsys is a fully multitasking, 100% protected mode, virtual-memory, massively-monolithic-style kernel. Added to this is a bare-bones C library and a minimal suite of applications – together comprising a small but reasonably functional operating system which can operate natively in either graphical or text modes. Though it’s been in continuous development for a number of years, realistically the target audience remains limited to operating system enthusiasts, students, and assorted other sensation seekers.
It’s a great little project and deserves more attention.
Just tried it. It seems that the latest version has trouble with the realtimeclock and will not boot the cd. The last version, 0.71 boots, and installs to a hard disk, but the boot sector seems not to be written to the disk. Visopsys needs some work.
Back to Haiku.
“Syllable… died out”
Looking at http://www.syllable.org and it doesn’t look dead to me. It looks about as active as Visopsys.
Syllable has come a long way, and is every bit as usable as, say, Haiku these days. Of course, my definition of “usable” is probably not the same as most people’s; I have an old PIII laptop that I keep around for running native BeOS and old Windows games, and Syllable makes for a great modern OS for that machine. About the only modern GNU/Linux distro that can run acceptably on it is either Slackware with a minimal X WM (Fluxbox or xmonad), or Puppy Linux, which I don’t care for.
Of course, Syllable’s usefulness is limited to basic tasks or dogfooding the Syllable OS, and since I’m not a developer on that (or any) project it’s more of a happy curiosity than anything else. I like it though.
So all these OSes are just used for old hardware which can be easily surpassed by a 40$ RaspberryPi which uses a lot less energy (~5W?).
So why ? why still using these old OS while there is an alternative with many usable apps and support ?
Sometimes I ask myself why some developers who are expert enough to create an OS can’t start writing useful things for Linux that only a small fraction can do it: OpenCL acceleration of apps, OpenGL Shaders, Graphics drivers, vector graphics rendering library, optimization, refactoring of huge and badly architectured apps, …
They can even try breaking usual habits in computing and come up with new solutions not tested before.
(For BeOS, I remember the database-like filesystem, and the ultra massive use of threads)
IMHO, these guys are losing their life while they can become gods in an other world and another community.
Stop it ! That’s my words !
Kochise
Syllable was a hobby; hardly “losing their life”. If anything the stuff I did in Syllable has been a great help to me professionally, and I’m still very proud of the code I wrote.
Vanders,
Don’t take it personally, everything posted on here is bound to get criticized by someone.
It’s really nobody else’s business how we challenge ourselves. Over a decade ago I also had my own little OS too, and it ended as I moved onto other things, but it fulfilled my personal objectives of learning how to do it. If anything, I feel like I’ve “lost” more of my life at work than spending time doing the hobbies I enjoy. If it were possible I would have liked to get that time back and spend more time doing projects of my choosing since it’s my life. I suspect the majority of us who work on personally fulfilling projects feel the exact same way.
While I agree that the knowledge acquired on such projects can not be equaled to “losing their life” and many times prove to give new insights to coders about OS inner machinery, it is also a fact of life that the choices we made affect what we can achieve. Perhaps, it would be more useful to get acquainted to other subjects but who is going to be sure which one would be that on a field that changes so fast? The amount of inner knowledge we have to “throw away” on computing is overwhelming but we have no option but deal with it, somehow.
“a field that changes so fast”.
In hardware terms, yes. Software, however, is one of the most reactionary disciplines that exists. To be successful, anything new must function exactly the same as the previous generation. Innovation does not happen that way.
Don’t just take my word for it. Alan Kay and Rob Pike have expressed much the same opinion.
(Rob Pike on Systems: http://herpolhode.com/rob/utah2000.pdf)
(Alan Kay with DDJ: http://www.drdobbs.com/architecture-and-design/interview-with-alan-…)
I worked in Syllable soon after it’s fork from AtheOS. I did a little work on CD booting then was part of the Syllable logo team. I think we did a lot of good work and I am proud of my contribution.
When I see Syllable logo I still think to myself, “Man we did great work!”
Hi,
Why did the Wright brothers waste their time creating their own/new type of vehicle from scratch, when they could’ve just improved an existing truck design?
– Brendan
an OS is way way way way more difficult than a vehicle.
13 millions lines of Code for Linux, thousands of devices and drivers, hundreds of badly designed hw but obliged to support, hardware itself moving so so fast into new architectures (HSA ?), …
The golden days when someone could stay few weeks in his room them come out with an OS are over.
In the other side, the gap between open source systems and proprietary ones are becoming more and more wide.
Adobe software started to be accelerated using OpenCL/CUDA/HSA … Graphics drivers still not very well supported like in closed systems…
BTW, for the usefulness of “a life”, I mesure that by how many other lifes it impacts. If you make the life of millions easier, then you have succeeded in your life. I don’t mesure that by how much it has helped me only improve.
I don’t disagree, but if you limit yourself to Intel cpus with integrated graphics then it most certainly is doable now that they publish both drivers and documentation for their GPUs. Nothing wrong by defining a narrow range of hardware to support.
What do you need to create a basic diskless OS?
– basic ethernet (you can borrow this from another open source os)
– basic usb drivers for mouse/keyboard/flash (you can borrow this)
– basic graphics buffer (you can make a simple framebuffer for a start)
You could use BareMetal OS as a starting point.
Or better yet, use the the free books from Niklaus Wirth about the Project Oberon[1], alongside with the available source code and try to make a make a single user desktop workstation OS using a safe systems programming language.
[1] http://www.inf.ethz.ch/personal/wirth/ProjectOberon/
I think that’s quite a disturbing philosophy.
disturbing ? why ?
it’s even an evolutionary fact..
http://en.wikipedia.org/wiki/Altruism#Evolutionary_explanations
I have to disagree on both points:
1) ‘OS is more difficult that vehicle’, no, some OS are simple, some vehicles are simple, some OS are complicated, some vehicles are complicated. Generally, they’re as complicated as you want to make them.
2) The days of making an OS in a couple of weeks are over? No, only if you want to consider an OS to be what the big players want you to think it is. An OS does not need to have thousands of drivers to be useful, only to support thousands of different types of hardware. Target specific hardware, don’t overcomplicate, and making an OS is as simple (or as hard!) as it ever was.
Hi,
I wouldn’t know – I’ve never built a vehicle.
An OS doesn’t necessarily need to support more than 1 architecture. Device drivers are also mostly optional; in that (assuming 80×86) you can get a long way with generic “frame buffer” video, no sound, standard AHCI and/or ATA drivers, a few drivers for the most common (wired) ethernet cards, 3 USB controller drivers, USB mouse, USB keyboard and USB flash.
Initial knowledge, quality and features of the OS and level of hardware support all vary significantly; so time to implement also varies significantly. At one extreme something similar to DOS could probably be slapped together in 1 month, and at the other extreme no OS is ever “finished” (there’s always more work you can do, even if you’re Microsoft).
Not really. E.g. Linux is able to work extremely well if you choose your hardware with a little care. The progress its made in the last 10 years has reduced the gap by a lot.
Every day I make the lives of over 7 billion people easier (by trying not to murder them).
– Brendan
I have built many times in the past, especially in my youth. Those are simple machines with wheels. In order for you to move the vehicle, you have to push it. It is really simple, there is no combustion engine or any complexities you may found from other vehicles.
allanregistos,
Haha, well I don’t think that would qualify as a vehicle in most people’s heads, the same way a bootloader wouldn’t qualify as an OS in most people’s heads
Building a simple vehicle is easier than building a complex OS. And building a simple OS is easier than building a complex vehicle. I just don’t find the comparison meaningful.
In my mind innovation doesn’t necessarily stem from complexity, I’d even say something is innovative if it reduces complexity.
Total and utter bollocks. A modern car is the culmination of 125 years combined automotive industry experience. It takes several thousand engineers 3-7 years and about $2-3billion to develop a “new” car model. However the “new” model is usually nothing more than a very minor iteration of an earlier design using existing technology (“a dot point release”). Every single component (“subroutine”) of a car from the windscreen to the door mirrors to the airbags has in turn taken the suppliers decades of engineering expertise and millions (billions) of dollars to develop. Even a minor design change such as a new grille and bumpers (“a code patch”) can easily cost $50million to implement.
In other words it takes thousands of engineers and costs $billions just to produce a “dot point release” of a Ford, VW or Mercedes.
Edited 2014-01-21 11:49 UTC
13 million lines of code divided by 100 programmers = 13,000 lines each. This is hardly a monumental task. David Cutler only used a small team to create Windows NT.
A modern PC OS only needs to support one architecture: AMD64.
Hardware vendors write most of the drivers and provide documentation for Windows.
Ha! Ha! Ha!
Right, that’s why I have spent the last month+ trying to find hardware that has not been crippled for the masses to run Haiku-OS on.
What I really would like is a motherboard that supports four(4) I7 cpus (32 threads) that will boot Haiku. Problem, no-one stocks such hardware where I can test it, and the online stores DON`T supply enough info before buying.
Haiku is not behind times, in many cases it is ahead.
Most times someone complains Haiku does not include some feature, 95% it is something I never used or every plan to use.
One thing Haiku does not do is the lastest fab feature, how many functions have been added to the top three OSes only to be dropped a few years later.
Today, I like Haiku for 90% of my stuff, and when I have to use a Windows, Linux or Mac I find they full of silly UI choices.
Edited 2014-01-20 14:49 UTC
There’s your problem – there aren’t motherboards that support four i7 cpus.
Four Xeons, on the other hand, is doable. What kind of info do you need on these high-end boards that isn’t provided by the stores?
Raspberry Pi has a lousy ARM11 CPU. Not sure it’s better. See the Linpack results for both Raspbberry Pi and Pentium 3.
As for OpenGL, OpenCL,shaders and graphics drivers for Raspberry Pi’s GPU, come on…
Well, I can’t speak for anyone else, but in my particular case, the Raspberry Pi isn’t usable as a desktop OS except in a pinch. It’s simply too slow. It does have a ton of uses outside that paradigm though. I’m using mine to prototype a commercial/retail grade safety timer system for work, and at home it’s been used as anything from a low power server, to a media center, to the heart of a ham radio digital rig.
My PIII laptop is (with the right OS) much faster than the Pi as a desktop system, it serves a couple of specific use cases I previously mentioned, and it’s just a hobby/fun system anyway. It has a built in battery that gives me a couple of hours untethered, and it uses much less power than the old Compaq tower I used to use for niche OSes and retro gaming.
As for the developers who create these awesome niche OSes…maybe they just do it for fun? Maybe they already do contribute to other “more worthy” projects and this is just a hobby for them? I suppose you’d just have to ask them.
I speak as a developer who would rather work on my own, unheard of project, than contribute to a well known project such as Linux.
Why? Well, it’s mostly this:
If you’re going to work on something for fun/interest/passion, I don’t think I’d find that being a small cog in a big machine. I also don’t think I’d find it working on 99% of projects out there, simply because I don’t believe in them, I don’t think that they’re improving computing, simple as that. If you’re working for money, then of course, that’s a moot point.
Linux and it’s associated projects are all well and good, but they’re not going to change how think about and use computers. I don’t think the Mac or Windows will either (they have in the past, but I doubt they will again for a long time).
But really, if someone wants to work on Visopsys, Syllable, then I think that they have a far greater chance of making something interesting and innovative than someone working on say, GNOME or KDE.
‘Design by committee’ is considered an insult in most circles, but that’s exactly what you get in most large projects (open source or not), the smaller a project gets, and the fewer people you have working on it, the more likely you are to be able to move fast and get unusual ideas implemented, and not left on the sidelines.
I think you have a point. If I can write my own OS, will just use ARM based CPUs to start with.
I think writing an OS is really a different field than graphics development. And I believe its fun to create an OS from scratch to see your creation blossoms as a full blown operating system is very rewarding. But I agree with you, they must use their talents to improve existing operating systems and their libraries.
I have exactly the opposite feeling of Linux/BSD dev’s: why are so many folk wasting time creating something that at heart retains the O/S philosophy of a 1970’s O/S when even some of UNIXs creators believe this is not a desirable situation.
We need more O/S development, not less, unless you really believe computing is a mature enough discipline that we’ve found all the answers.
Nothing significant has been added to Syllable for years. A lot of the main devs left or were driven out, and the obsession with REBOL has consumed so much time that I don’t see it ever gaining momentum again. REBOL is very much a problem in search for an answer. It’s lovely for Kai that he was able to showcase it as a first class language with in an operating system, but there is so much else wrong with the current releases of Syllable as to make you really wonder what the point in adding such an overly specific and niche language to an OS with its feet strongly in the C/C++ world was.
Don’t let the site confuse you. That last news item, that reads “Dec. 19”? That’s from *2012*.
I can *swear* it used to actually say 2012. Not sure what to make of that.
It seems there was some “problems” with the site hosting, it’s an old version of the site. Check here: http://forum.syllable.org/viewtopic.php?t=2000
I hope the project doesn’t die, alternatives are always necessary.