Operating System Success

[spix]

Excuse me, I've had a few beers (actually all of them >:( )

I've got a few questions. You've all been a great help the last 3 years I have been OS dev-ing so thanks for that..

Most of us have our own projects, our own hobby operating systems, and so you may have asked these questions yourself.

I have read about past exercises in trying to get the OS dev community together to make a kick @$$ OS and how it failed, and those who tried have grown wiser.

Recently there was a poll about whether your OS was original or a clone, Some answered yes, some no, and some a Frankenstein kind of thing.

Currently it seems that most opensource operating systems are clones of something. Although linux seems to have out grown the "clone of UNIX" and unix is becoming a clone of Linux.

Most people get bored of OS dev after the hello world kernel, or soon after, then there are some who have been here forever (eg, Brenden Pype, Solar and Candy) Brenden with BCOS, Pype with Clicker, Solar has found a niche with PDClib and I am not sure what Candy does. Then there is Dex with Dex4u and the other regulars I have missed out. All these projects are truly hobbies and there seem to be no public aspirations for something greater.

I wonder what are the private aspirations? Does anyone here secretly hope that there OS becomes the next big thing? Sure, it's a hobby, and the fun is in the coding, everything else is cream on the cake, but does anyone have a secret burning desire to become something else?

If you have these private aspirations, what are you doing about it? These dreams don't get realized by themselves, you've got to make them happen.

My OS is getting close to my original goal, and I am wondering if I can create a following like Theo did with OpenBSD.

I am probably just rambling. I am curious though, does anyone here have any desires beyond a hobby? and what are they doing to achieve them?

Andrew.

[Brynet-Inc]

Well I can't speak for anyone else, I was only part of an OSDev group for a few months and in the end the other developers lost interest.. So I was unable to continue on my own..

Mort-OS looks very cool, My goals were and are to be nearly what you have accomplished already.. Working in some state at least, Portability of applications for example.. Small foot-print UNIX with basic device support and a /bin full of goodies

Like one day booting the OS on real hardware and watching it work to the extent where you are joyful and proud of yourself, (Mounting your first device.. filesystem support.. syscalls, etc)

Theo's OpenBSD rocks, I use it.. It's his daily mission to protest against companies refusing to release documentation on hardware or make their binary blobs freely redistributeable.

(Your sever is awfully slow pal... :'()

[spix]

Thanks for your reply

Like one day booting the OS on real hardware and watching it work to the extent where you are joyful and proud of yourself, (Mounting your first device.. filesystem support.. syscalls, etc)

This is what I meant by being a hobby and enough to keep many of us interested, I am excited when I get something work in Mort, and a expect many of us are excited when we solve problems, it's the appeal of OS deving.

Theo's OpenBSD rocks, I use it.. It's his daily mission to protest against companies refusing to release documentation on hardware or make their binary blobs freely redistributeable.

Don't get me wrong, I love OpenBSD and use it everyday (my autobuilds are done on OpenBSD 3.9) but I wonder how much of the success is due to Theo. They forked from NetBSD a long time ago, and in my experience, OpenBSD is a hell of a lot more stable, and easier to use than NetBSD.

Theo is a great leader, I expect that is the cause of his success.

You know I am not even really sure what the question is I want to know the answer for, I guess on the surface is what practical steps people take to make there OS succeed apart from the code/development side of things.

(Your sever is awfully slow pal...

Sorry, I wouldn't know.. I only have 512k/128k adsl, it seems pretty quick to me, but then it's hosted by my ISP :-/

Thanks again for the reply, good luck with your projects..

Andrew

[Brynet-Inc]

(Your sever is awfully slow pal...)

Sorry, I wouldn't know.. I only have 512k/128k adsl, it seems pretty quick to me, but then it's hosted by my ISP :-/

Ah, Well I went all out and purchased 20mbit by 1mbit, I get around 1900kb+/sec from most servers, 50+kb/sec from yours..

Generally the response time is the only delay (loading your OS screenshots times out a few times..)

[spix]

Ah, Well I went all out and purchased 20mbit by 1mbit, I get around 1900kb+/sec from most servers, 50+kb/sec from yours..

LOL. I live in rural Australia, I wonder if the ISPs around here even have that kind of bandwidth..

Generally the response time is the only delay (loading your OS screenshots times out a few times..)

The screenshots are on Flickr, so I dunno what the problem is there.

Andrew

[Pype.Clicker]

Most people get bored of OS dev after the hello world kernel, or soon after, then there are some who have been here forever (eg, Brenden Pype, Solar and Candy) Brenden with BCOS, Pype with Clicker, Solar has found a niche with PDClib and I am not sure what Candy does. Then there is Dex with Dex4u and the other regulars I have missed out. All these projects are truly hobbies and there seem to be no public aspirations for something greater.

I wonder what are the private aspirations? Does anyone here secretly hope that there OS becomes the next big thing? Sure, it's a hobby, and the fun is in the coding, everything else is cream on the cake, but does anyone have a secret burning desire to become something else?

I once filed a "this project needs help" request on SourceForge for Clicker, and there were several guys popping in my mailbox saying "that's amazing, i want to join!". Then there was the discussions about high-level stuff, security, device drivers, GUI, etc. but no code. You can talk to Solar or Candy about that: it's always the same. Unless you have the time available to keep all your dever's around you, it's almost impossible to find guys who adhere to your visions enough so that they indeed start writing code that fits your system.

Most of the benefit i got from MT has been one-shot contributions (mostly GPL/PD code written by third parties that i ported to clicker myself), and atm, i no longer feel like porting stuff to clicker myself. The Clicker source base has drifted from what i'd like it to look like, there are plenty of things that should be rewritten to be clearer/more efficient here and there ...

And then comes the x86 nightmare. A piece of ugly "architecture" that never draws a straight line on what should be supported. Segments and TSS will one become totally obsolete, and it's unlikely i'll have 64-bit code running by that time. ACPI is auwful, USB is complex, 3D is undisclosed.

Then comes the almost-30-blues things. In a couple of years, i'm likely to have babies around ... in 6-7 years, those kids might ask me "you can really create a new program for the computer?" and i'll say "sure, kiddo ... look" ... and what am i going to show them ? Clicker? they'll hardly be proud that their dad has drawn a blueish raster on screen on which you can move a small square, will they?
What else? the small games i wrote when i was 20? yeah. that could be ... but then they'll ask "why did you stopped making games?"

And honnestly, i don't feel like answering "because i wanted to focus on the moving square" ...

[elderK]

Interesting question .

and I hope you dont mind if I jump onto the jabberwagon .

I can wholeheartedly agree about the 'Looking for help' thing, a lot of people hear 'Custom OS' and go 'oooerr' I want to try this!, Of course - Once they learn about the work involved, you can usually expect a good number of them to jump ship.

My aspirations are simple (Yep, thats my life ideal.).

Baseline is, I want to make something that I can feel proud of, and that someday I can use.

The ambitious part is - Im developing my project, to ... not limit the user in any way, if they want to do something or change something, they can do it.

And, My all-time big wish is, that someone will find the sourcecode useful and say, 'Hey, Youve done a pretty good job. Well done.'.
When I hear that, I know that all my time and hair pulling was worthwhile. .

Besides, I have to admit - Im obsessed with it.
Ive worked on a lot of things, out of curiousity and impulsive-interest, once theyre written, the interest usually dies away.

Not so with Kernel developement, Its always in my mind, so much that Ive virtually dedicated my life to learning as much about the ugly, horrible IA32 architecture. (But, I Didnt say that Mr. Intel sir, Who just gave me a bunch of Manuals for free, WOHOOO MOOCHING!)

It gives my little life some meaning, and it keeps my mind working on things, so I dont think about things that make me sad. .

~Zeii.

[Habbit]

Ive worked on a lot of things, out of curiousity and impulsive-interest, once theyre written, the interest usually dies away.

My case is worse: interest usually dies away once things are HALF written (hashsum, adsvote and i hope not the sfs driver)

[Dex4u]

First let me congratulate you on your OS and a great question .
But if you where looking for inspiration, i think you will be disappointed. OS Dev is getting harder each year that passers, as your OS takes one step forward, hardware and peoples wants from a OS, take ten steps forward. For example it tuck me 3 years to run a simple web server on my OS, but i can buy a single chip with full server for about ?3.
As for getting help i may have been lucky, but "Team Dex4u" has stayed together form the very beginning and the original group is still together.
That is down to the fact that OS Dev's are by nature individuals, with there own ideas of how things should be done, "Team Dex4u" did not try to tire them down to one idea of a OS design, but instead use there ideas, So each member as there own OS projects, if they do not like a way Dex4u does something, they do it differant in there own project, if by doing this its found to work better than the way Dex4u does it, we will change to there idea or the other way round etc.
One thing Dex4u, was never designed to be, was a desktop OS, but more a modern day ver of what Dos was used for in its latter days, eg: a good platform to get directly at hardware.

But my advice to people who want to hit it big with there OS is forget x86 and go for ARM or some other processor etc.

I also must add its getting hard each day to motivate my to add more code to the x86 ver of Dex4u .
Good luck with your OS.

PS: As a side note, i have been working on my main OS for about 4 years, but i also make a 2k dos clone, which tuck about 3weeks to make, the dos clone has been down loaded 10 times more than my main OS and i get more people say they use it.
So effort does not always mean success.

[Brendan]

Hi,

And honnestly, i don't feel like answering "because i wanted to focus on the moving square" ...

I take a slightly different view... I look at all the people around me and wonder what they've actually achieved. It's almost always the same story - went to school, (optionally) did some further training, found a spouse, had some kids, worked for <insert names here>, then retired and died. It's all "good", but I wonder if that is all life is about...

I want to create something - something that people can look at and say "Wow, he did that?". I know what I'm capable of, and I'confident that given enough time I can acheive something special, but...

The main problem is society and economics. Once you've left school you don't get free computer hardware, free housing, free food, free electricity. Until you retire you're more or less pressured into spending half your time working and the other half recovering from work, eating, sleeping, paying bills, etc. There is some time left, but not enough to actually achieve anything of significance - even if you did have a "complete" OS, just finding time to keep it up with technology would consume every spare minute you have. Eventually you retire, and you've finally got time to work on things, but the best part of you're life is gone and you've probably given up all hope long before then anyway.

The only real hope is to find a way to get paid for developing your OS, or winning a million dollars in a lottery. I'd say winning the lottery would be more likely...

Despite this, there is always a very tiny chance that you can get enough time to create something that looks promising enough to convice people to either help fund your work or help develop your work. For example, if 1000 people donated $20 a year, you'd be able to work on your OS full time, make it more promising, and (hopefully) increase the number of donations. This may not sound impossible, but there will always be much larger and/or older projects that look much more promising. This isn't the only way though - maybe the OS is good enough for you to get a research grant, maybe it's good enough to attract investors, or maybe it's good enough for people in a niche market to pay for.

So I guess the next question is, how much time do you need to spend to get your project to the point where it;s "good enough"? There's a huge number of factors involved here...

There are some things you can do to significantly reduce the amount of time it takes to reach the "good enough" stage, like porting large chunks from other OS's (for e.g. porting Xfree86 instead of developing your own code), or making your OS a clone of another OS (so you don't need to spend time making sure your design is solid).

These things can significantly reduce the amount of time it takes, but IMHO they can also significantly reduce the chance getting some form of funding. If the OS is too similar to any existing OS then people won't be curious enough to donate, it won't be interesting enough for research purposes, and probably won't be better for any niche market.


Cheers,

Brendan

[Pype.Clicker]

it's almost impossible to find guys who adhere to your visions enough so that they indeed start writing code that fits your system.

Oh, i need to credit whyme_t, of course, for being a strong exception to that. He did a very nice job on PCI/PS2/RTL8139 while i was busy working on kernel internals, which fortunately enough kept the system alife during the past 3 or 4 years ... but now he's got job & assignments of his own

[xyjamepa]

that what i was talking about in osdev and the future...
after all this years and after all this job and researchs
benefits you can get ... no one can ever pay you some money
for your hard work
it's a good thing to show your kids your os and they would be so proud of thier dad,and also a good thing that some body
downloaded your work and said that's nice and amazing
and he used for learning stuff.

go for ARM or some other processor etc.

why is that for?

I'd say winning the lottery would be more likely...

i agree with you...
thanx.

[distantvoices]

Yeah, and there's me ... (dunno why ppl think I'm not here anymore, just because I've ceased to post regularly - no time, period. I'm after a masters degree)

@zeii: You are high on coffe, eh? *wink*

I reckon, for my micro kernel project there will be many more years full of developing and coding sessions. I don't think about success. I already know, that I can do it, so I continue to spin the thread, just to see where it leads me.

Sure, there will be Family. Yes, for me. Kids gambolling around and shouting *boo* into my ears just to see if I hear (maybe I won't, Time will tell) This matter ain't keep me away from development. It'll happen, in a slower pace. One can't stick to the computer if a kid wants to have "da time".

well --- future, the undiscovered land.

stay safe

[elaverick]

I did a lot of work in the Open Source Quake community before I started messing around with Kernel dev, the main thing there that usually guarrenteed sucess (sucess equating to people actually playing the games) was for people to actually release something.
I think the same could be said of the OS deving community as well. The operating systems that are picked up seem to be those that are actually feature complete (to what ever level the current version is desired to support) and released. MS for all their delays do actually manage to get products out of the door, the same is of course true of the various well known Linux and BSD distros. Other OS's that seem to have been well received are things like FreeDOS and ReactOS, both of which also manage to get releases out their and into the community.

I'm not sure if my OS will actually go anywhere as it's still in very early stages (as I am rapidly finding is my knowledge of what's actually going on... nothing like OS dev to make you feel humble), but if it does get the basics down then I intend to set very strict milestones for the development process and try very hard to avoid "feature creep". My first milestone is to get the OS self hosted... if I can do that then I will be a very happy bunny indeed...

[Candy]

Most people get bored of OS dev after the hello world kernel, or soon after, then there are some who have been here forever (eg, Brenden Pype, Solar and Candy) Brenden with BCOS, Pype with Clicker, Solar has found a niche with PDClib and I am not sure what Candy does.

I'm aiming at Alpha Centauri. I'm aiming so high it's ridiculous to hope I'll ever hit anything, but if I do, everybody's going to be amazed.

I wonder what are the private aspirations? Does anyone here secretly hope that there OS becomes the next big thing? Sure, it's a hobby, and the fun is in the coding, everything else is cream on the cake, but does anyone have a secret burning desire to become something else?

I'm actually hoping on people seeing stuff in my OS as good, as opposed to a mixed feeling about a number of things. I'm trying to think stuff out to the maximum level I can think of and then a bit further, so that all stuff connects and connects logically, all power is shown while no complex operation can be done automatically, all userland things are replaceable and separate, the operating system is as modular as I can get it with the least amount of physical system dependency, all is as free as it can be and nothing suffers from any legacy code slowing it down due to excessive conversions. That includes bringing a few ideas from areas they're being used to other areas where they could do good but weren't used.

As I said above, I'm aiming too high. That includes only considering stuff that's in full PD as fully open, the rest has at least one string attached, attempting to recreate everything and to understand everything to the lowest level I can comprehend myself. I'm partly satisfied with make, mostly satisfied with gcc (but at the moment not at all with ld), mostly satisfied with yasm and somewhat satisfied with common shells. I intend to recreate them all, plus OS kernel, libraries and so on. All together until it forms a full system again that isn't built layer on layer.

I am probably just rambling. I am curious though, does anyone here have any desires beyond a hobby? and what are they doing to achieve them?

I don't expect to finish. I don't expect to come too close to finishing either, since I'm not even coming close to any goal and as others have said before me, the ground on which our stuff is founded moves faster than we can move our stuff. I'm hoping to beat this by genericity and the principle of least innovation: all current systems will be based on existing systems as much as possible, including methods of thought and practical implementation.