ghostofsparta: I have been honing this idea which has been lingering in my mind for years now. It borders within a concept of type of OS entirely different from the conventional ones. I'm now self-convinced that it may even give Microsoft's Windows a run for it dominance. Infact it's so beyond both conventional 'styled' and futuristic types (Iron-man-ish). It's far from what everyone is used to from the likes of Wintel to the Linux based ones. But the issue I am having is I can't discuss my 'novel trove' with even my programming pals without divulging the main core concept in the hopes of getting some hints/advice on how to go about it. So since I'm left with the option of having to venture into programming myself, I have done my assignment here and there where I figured the basic differences between C, C++ and C# and why C is the most needed to learn for OSdev. I also learned that one shouldn't go into programming having OSdev in mind rather one should learn the languages "In and Out" first of all before thinking coding 'garri' oriented apps much less of an operating system. I'll have to regard myself a n000b, hence using this thread to solicit ideas, guides and encouraging opinions on how to go about developing an operating system from scratch to perfection. I creating this page as a forward to another page where game developers and programmers were being compared, I am aware it's not an easy task to do, some responders may say something like "dude, you need to ask yourself what purpose does your OS caters for" while some folks may say "stick to libraries". All I am asking as a n00b is what are the conditions prior to plunge into developing a totally bootable OS, hardware and software wise.

Thanks

2buff:

A book on OS systems....or Google.

ghostofsparta:

Both of course, but are they books on OS systems? I mean not OS manuals but how OS can be built?

Kambo: your post doest take long to betray you as a jjc. you dont even know how to align curly braces in c,c++ yet you want to "write operating systems". chill. bang your head on the languages first,then pick a book and stretch your mind a little

Kambo: it will take you (100 * 100 years to code it up.)

2buff: That being said, As far as the world is concerned today there are only 2 major Operating systems. Windows and unix(i.e Linux).
Both Apple's Mac/iOS and Google's Android are nothing but very different custom linux builds

WhiZTiM: this is going to be an interesting topic.
Windows 7 is believed to be above 80 million lines of codes. How are you gona beat that?

WhiZTiM:
Operating systems are the most difficult software to write from scratch.

WhiZTiM:
Not to discourage you, go and read about Kernels, Filesystems, ... Before going far... I did suggest you check out Minix. www.minix3.org, its probably the smallest and useable industrial strength Operating System available...
Its a few tens of thousands lines of code. If you understand the codes and working architecture fully, then you may proceed in taking out the next 7yrs of your life developing an OS.

WhiZTiM:
Before studying minix, you have to be a Veteran in post graduate level Mathematics, Systems Programming, algorithms, etc.

WhiZTiM:
Ever heard of "ReactOS"? ...

ghostofsparta:
No, what about it? I'm sure it won't be different from the rest of the pack in terms of its GUI

ghostofsparta:
Why should I beat that? Am I suppose to be intimidated with the 80million code lines everyone is discouraging me with, you just typify those Nigerians who believes since it took the United States 200 years to attain a perfect quasi democracy, therefore Nigeria must tread through such long years. For your information, there are certain missiles like the Iron Dome and certain TomaHawks whose sophistication is beyond 200 million self-generating ultra-AI oriented code lines.


Who doesn't know?


@bolded^^^: Only useful piece out of everything you wrote


Tell me, is Bill Gates or Kevin Mitnick a mathematician? ofcourse who doesn't know one needs to employ certain levels of arithmetic to represent a problem in programming. And remember when I said it still depends.


No, what about it? I'm sure it won't be different from the rest of the pack in terms of its GUI

lordZOUGA: you need a team sir and lots of passion and money too..

lordZOUGA: I'd advice you to study the linux OS. no need to start from scratch.

crossfire: Game programming is simply out of this world, look at the minimax algol. Try developing a simple chess program. Or simply look at the source of "stockfish". (Stockfish is an open source chess program).I've looked at an early version of LINUX called MINIX ( i'm talking of hundreds of thousands of lines of code) and my little lazy brain could wrap around some of the concepts and data structures (I'm a software developer, and a chess enthusiast), but anytime i look at the internals of "stockfish" i feel like a child playing with sand. It looks like this program was written by aliens . There is another one called "Robbolito" that's simply crazy inside and except for a few "anonymous" comments on one russian blog nobody really knows who wrote this program .
Now stockfish is a console app that runs through an already developed UI like "arena" so it doesn't touch on graphics. You need an advanced level understanding of maths and physics before you can develop game apps that manipulate the graphics card masterly like as soccer or even car racing games do.
Desktop apps are developed with C++, C# java, python, but games are predominantly created with c. Now unrestricted c is simply mind bending and very unforgiving. I always tell my students that c is the latin of programming languages
Lets not over flog this : game programming != computer programming.

WhiZTiM: ...
Well it is. Its a striving Operating System(written from scratch) that uses exact model of the Microsoft Windows NT Kernel... And so far, those guys are trying and have still been developing it!

WhiZTiM:
Since u said its not something so big that you are after, ... Just start by checking out these...
Writing MBRs
Bootstrapping and Boot loaders...
Processor instruction sets..
Processor' supervisor modes and interrupts (depending on what type of kernel you want to make)...
File system structures...
Hardware Abstraction Layers (HAL)...
Handling and decoding Hardware interrupts...e.g from input devices...
Display and Output devices...
Synchronization...
Virtual Address mapping.
Assembly language may be required at some point.
Providing services and API to run processes on user space mode.
Context switching etc...

WhiZTiM:
Afterall, Some embedded OS had their 1st version written in a few months... But took years of continous development to become acceptable..
Seriously, you gona need a lot time... Probably some finite percentage of your life time...But for the start, I suggest you should attempt to write a small OS of a few thousand lines for the RISC MIPS processor. It doesnt have overly complex instructions like the IA x86, x64 families. There are some MIPS environment emulator around. Google it. Goodluck.

Elxandre: @op. mr gates has an IQ of 160..thats a genius man..he could learn most things if he wants.

gbigbega: I see a 100% failure ab-initio. With such arrogance, you will get nowhere. You have so little idea of programming and your first shot at it is an OS? Excuse me. You must be an expert in different programming technology before you can even attempt it. Why do you think big companies like google didn't build theirs from scratch. Oh mind you, Bill Gate did not build Windows from scratch, go and make your research. He started with DOS, which he bought, then migrated to window which me stole from MAC when he was on a project with them. He however still did not build it himself. It took thousands of expect mathematician and programmers whose expertise focus only on secific areas to build it. Bill gate himself can never build a windows 7 kind of OS himself. If you think OS won't need much programming then you are clearly disillusioned. You will need thousands and thousand of codings to write the type that people will be interested in. You will also need millions of dollars to get it to the stage it will be acceptable in the market. I was in microsoft a while ago and one of their technology expects boasted that they spend about $2 billion or so (I can remember the exact figure now) on research alone every year. I think its easier to build computer model from scratch than building the OS that will run it. Seriously.

Having a good idea is not bad, but let it be a realistic one. Start by learning the language, there is a say that it takes a life time to learn all of C. Start writing small programs, see if you get to even write a sellable software. If you are able to do so, then you can gradually proceed. I have been programming for more than 10 years and have made a lot of money doing so but I will not even think of an OS talk more of attempting it but guess what, I have great ideas of an OS that will push microsoft and Linux out of the market too.

If you are born to do this, please don't stop. Buy a computer with configuration that will still be relevant in the next 50 years, buy a 50 years internet package for browsing, lock your self up in a cave that won't open until next 50 years, don't socialize, don't marry etc you may just be able to do it. If not, pleeeeeeasssse get a life.

Let the advice from true programmer be your guide, those who don't know what it entails are not qualified to encourage you into a path they no nothing of. Please humble yourself to learn and not waste your time. Your aiming for heaven his good, if you are deligent, you will land among stars.