OSDev.org

This is a important ?, as we can not hope to keep up with the moving target that is todays desktop OS, so we need to find out what Hobby OS are better at, than your windows or linux etc, or if there is anything that they are better at.

Eg: a quick example is you can not get raw TCP packets under windows or injecting packets is hard under windows, also things like detecting rookits , POS Software (eg: Dos being better) etc .

If we can answer this ?, without baised, we may beable to make Hobby OS more successful and usefull by designing OS to meet these requiments, but at the same time still learning about the inner workings of a modern OS

Hobby OSes don't have legacies yet. No layer on layer on layer of compatibility, no sledgehammers required to fit new things in old designs.

I think that covers about every common desktop OS...

Good topic, Dex. And a very pertinent point. I myself wonder sometimes if I can ever catch up to the big fish. It's getting way more obscure with every passing month. I just, (yesterday) install Karmic Koala, and I'm stunned by how amazingly fast, and ...user friendly, and 'pretty' it is.

Every other OS in the major league is really moving up, and leaving the hobby age way behind. In order to hope to compete, there is, apparently only one hope: UDI. I've seen several other members stressing it, and there's a family member of mine who has ported it, and I myself have read the specification.

Modern hardware is advancing *too* rapidly. Very soon, everyone will have multiple, SVGA/XVGA resolution monitors, or touch screen monitors of some sort, and any OS even hoping to get a second glance will have to support hardcore multimedia on all kinds of radical display set ups. It's disconcerting.

I think, and would like to point out that it is both foolish, and also no longer a luxury for any hobbyist OS developer to be thinking in the 'I-move-alone' mode. Kernels coming out today have an enormous load of drivers to support off the bat. It's striking me more and more that the competition (where hobbyists are concerned) should not be at the driver level. Because if everyone independently develops his/her own drivers, we'll all fall in the end.

In fact, for any sort of strong OSDev community to be able to continue into the future, we will have to work hard at supporting having an abundance of highly portable drivers for immediate use.

If not, the big league OSs will simply dominate forever. And then, as time passes, more and more, it will really become nothing more than a dream for anyone to think they can break into the x86 kernel development market.

What are Hobbyist OSs better at? As a general question, we offer a fresh outtake on what's already there. Inevitably, we will all have to support the same hardware, and the same tasks. Technically, we don't do much different.

The thing that we have going for us isn't really what we are better at, as far as I can perceive the situation; Our main advantage comes from our pure hindsight where current OSs are concerned. We can see the falter points in Linux (/dev/this, etc), and Windows (evolutionary development on a badly programmed core), SkyOS (closed source, proprietary in an age where hardware is impossible to support without community help) and every other OS out there. And we can pick up the slack where they left off.

What are we better at? I don't think there's anything much for us to boast about, since most of us, myself included have not much of anything impressive to show. And again, technically, a fully mature kernel will eventually have to do the same things as every other kernel out there. Differing only in methodology.

Why should we fight over something as simple as drivers?

This post is very sentimental, and whatnot, but a highly practical limitation is the lack of a champion operating system for the OSDev community. If there is no operating system mature enough to use UDI, then there is no operating system mature enough on which to test drivers written for a UDI environment, and as such, there will continue to be no UDI drivers.

To that end, I'd like to ask everyone to probably, should pcmattman, eddyb and JamesM, etc decide to allow it (they technically have invited everyone to join the #pedigree channel and come do testing), make an active effort to support the Pedigree OS project, which is, to my knowledge, one of the only viable projects on our board.

I believe Mojo is also very mature, what with it having a full GUI, etc. It's time we kind of made ourselves a force to be reckoned with. When the FOSS/GNU horde wanted recognition, they took a project and championed it, and religiously promoted it with everything they had. They went the full mile.

Until we have something to represent what we are, and show that we're efficient, we'll forever remain just lingering around the board, discussing the dreams, and never getting to see them get big.

I don't know if anyone else is feeling a sense of futility, but I definitely believe that if, within the next 2[/i] years, I don't get a solid kernel out, I'll never have the chance. I assume that at the rapid pace of hardware development we see today, within 2-5 years hobbyist OSDev for x86 will REALLY be nothing more than a waste of time.

--RFC
gravaera

I think you're missing two important points: First, there are lots of people who want to write their own OS, not just participate somewhere (they could take Linux if participating was enough for them). And second, most of us don't strive for world domination.

Good ponts gravaera, i know what you are saying ( but from my point of view), i do not beleave or have never have, since starting OS Dev beleaved that its posible to make a sucseeful desktop OS.
It is just impossable, other than from a learning point of view, there are alot of great desktop OS out there eg: linux, there is little need for more desktop OS's.

We need to find something that the big boys can not do, because the user base is too small ( this still could be many 10,000 of users) or its too risky to there image.
What a sucseeful hobby OS needs to do, is to be usefull to a number of users, even if that OS does just onething, then it is more usefull and a more sucseeful hobby OS, than a hobby OS that trys badly to be a desktop OS.
From this point of view, if you do not count linux, then one of the most sucseeful hobby OS's is FreeDos, now if you take that as a example, its taken a team of Dev many years to get it to a stage, where its usefull, now just think about it, as in OS terms its the most basic and simple of designs .
Its take that long to get it to a usefull stage, what chance have you or anybody else got of coding a full Desktop OS ?.

Also why is FreeDos sucseeful in hobby OS terms ?, because you can still do things on it that you can not do on the other big three OS.
We need to come up with the modern day Dos, if we are going to succeed.
We also need to see success in hobby OS tearms, as how usefull they are.

Eg: Can they fit on a floppy, and with less than 4 click be set up as a web server and run on anything from a 386 ?.

Kevin wrote:I think you're missing two important points: First, there are lots of people who want to write their own OS, not just participate somewhere (they could take Linux if participating was enough for them). And second, most of us don't strive for world domination.

Hi: thanks for posting, and I appreciate your sentiment; I believe though, that you didn't quite understand the full ramifications of the ideas I presented. Maybe you could re-read my post and try again? And as a side note, while not everyone may be striving for the world domination you speak of, I don't believe it's fair for you to imply that doing so is innately 'wrong', or 'not good'. I'm very sure there are folks who would like to see their hard work moving into a grand scale, and for such, a strong, recognizible front offers benefits. Maybe your goals are short sighted, but implying that anyone who dreams a little higher is somehow on the bleak end of some spectrum maybe isn't too open minded.

I also never implied that anyone should abandon his/her private project. I myself am not part of the dev team for either project I mentioned. All I ever do is test once in a while on real hardware wherever I am able. I simply tabled the notion that maybe lending help in strength, via testing or some other more vigorous effort, to any of the matured projects could be a good strategic move.

And technically: any project which has matured to the point of having, say, a workable kernel which can run programs, and has, maybe a full posix compatibility layer of some sort, and can compete at a respectable level, will most likely inspire its dev team to move on to the wider world, seeing as the project would have, at that point, potential. Who wouldn't continue their kernel's development if after several months they discover that "Hey: this is now a stable, working, fully workable kernel!".

First, I think the important question is this:
What IS a hobby OS?

Is Syllable a hobby OS? Is Haiku a hobby OS? Is ReactOS a hobby OS? They're definitely not mainstream, and they all tend to have small development teams. At what point does an OS stop being a hobby OS? What goes between hobby and mainstream?

(Incidentally, both Syllable and Haiku are getting mighty impressive)

gravaera, I'm not saying that the big project you're dreaming of with most of the forum members is bad or wrong or anything. What I'm saying is just that I don't think it's what reality looks like because most members are not interested in such a project (of course, go ahead and try to prove me wrong here). You're right, we might never acquire parity with the big ones with our hundreds of small OSes. But that's okay, it's not even my goal. And we don't need to be efficient, we're not a company.

I deliberately haven't commented on the UDI thing because I already have too often. I still think it belongs to the same same world as your project: It's just a dream. As far as I know, none of the OSes here implements UDI. So what does it buy me? I use what actually helps me. Currently this is not implementing UDI that noone is using anyway, but sharing CDI drivers with (amongst others) Pedigree. It's the small solution, but it works for me and it's reality.

We don't all need to write a full fledged OS from scratch. If people really share their code (advertise their projects, and making sure the components are portable), it becomes much more viable to work on those things you are good at and plug in whatever else you find useful from others work. Open source is not just about freedom, its also about allowing people to improve on other people's work.

What I see in pedigree, the main reason they have so much running, is that they have stolen about every good unix tool in the book - bash, python, gcc. There is no need to reinvent the wheel, so why bother? True, there are people who are doing that for the sake of it, but that is what they want. And maybe the end result is better than bash (all in the eye of the beholder), and other people start using it. The great pyramids were not built with just one pair of hands - it would be unwise to try. If you would for instance have a crane, you could stack those massive blocks much easier.

Regarding UDI - it's the same as SFS. Both are good concepts, both lack the support to make them a de-facto standard. Their success comes largely with the effort spent in them. There has been a lot of pushing around SFS some time ago, and it has an (albeit small) software base, making it actually usable. Similarly, UDI is not going to leave soon because it has been mentioned too often.

What I think we should do, is focus more on the parts of the OS we like, instead of redoing the same things all over again. There is a good share of hobby kernels around, you can build on that, and then focus on the parts you like, and grab the parts you need from elsewhere. That will give a lot more diversity, and make it easier for people to rise above the hobby level and make something interesting.

I personally have some plans for porting the UDI framework. Not for the sake of it, but because it gives me an opportunity to provide you all with high quality graphics without you having the need of knowing all the intricate details that I do. I'm not alone here, the other gurus of wisdom are contributing their own masterpieces. There's a C library in the make, there are portable x86 disassemblers and emulation libraries. All those people share their work for you. There are enough helping hands to take you out of the stone age, you just have to invest a part of your pride, so that a greater pride may return for you.

A lengthy $.02

One thing I'm disappointed about is that theres no good off the shelf microkernel that I could find; There's L4 derivatives like L4ka::Pistachio, but theres absolutely no documentation - I don't know how to write my own root servers.

And it's very disheartening - because I don't really want to touch kernel space; I've done that, it's boring now - and I'm sure L4ka::Pistachio is great, but I don't know how to use it!

First off. @dex: Please install a spell checker and work on grammer.. You have no idea how much I cringed while reading your posts.. (and I'm not a grammer nazi.. but it was just painful)..

Anyway.

I would qualify a hobby OS as something that is not used as a primary OS by anyone. Or by less than the amount of developers. (Yes, we all know you can be self-hosting. But have you overwritten your main harddrive with your OS yet?)

The point you must understand though is that most people develop OSs because they are fun. If developing an OS isn't fun for you, your probably not developing it very actively.

earlz wrote:Please install a spell checker and work on grammer.

Kind of ironic...

earlz wrote:I would qualify a hobby OS as something that is not used as a primary OS by anyone. Or by less than the amount of developers.

The thing is, that definition makes all of the negative statements about hobby OSes self proving. Once a hobby OS becomes good enough, it simply ceases to be a hobby OS. That's like saying that because we don't see homeless millionaires, it must be impossible for homeless people to ever become millionaires.

NickJohnson wrote:

earlz wrote:Please install a spell checker and work on grammer.

Kind of ironic...

Dex wrote:

NickJohnson wrote:

earlz wrote:Please install a spell checker and work on grammer.

Kind of ironic...

Yea, can't mess up the grammar in smileys. I seriously skip over parts of your post because its just unbearable to read. My conclusion is your either 10 years old or you have been speaking English for less than 6 months. (at least, that's what your posts sound like)


and @nick,
Yes, I would consider something no longer a hobby OS once it's main stream. Look at linux. That was once Linus' hobby OS. As was the Apple OS of the olden days. A hobby OS stops being termed a hobby OS whenever it ceases being a hobby basically. What I mean is that when an OS starts being used by multiple people as primary platforms, the OS is critical. There are actual maintainers by that point. Though it may be taken as a hobby for the developers of it, the OS is not used by its users "as a hobby."

I would say the defining point of a hobby OS is, that it is a hobby. It's being developed because you want to, for fun. The lack of use they get just stems from this. Eventually, when your "hobby" grows large enough, it ceases to be just a hobby and becomes serious business.

Just my $0.01