how long before a programmer can make an OS?

[trashthing]

hi. how long should someone be programming before they make an OS?

[x_Ray]

I don't think it's a time frame kinda thing, just because someone may have the talent and skill to a accomplish such a task doesn't they would be interested in actually doing it. It could take a lifetime for a person to reach the point where they're skilled enough to make a worthwhile os, it might take another person a few days...no way to put a number on it really..

Maybe a better answer would be never, why reinvent the wheel? I think a better use of that talent would be to help better an existing open source os...

Just my opinion...

[Citadel]

You need certain skills, for example.

• 1) The C programming language.
  2) Understand Software architecture.
  3) Understand Hardware architecture.

You need to know C so that you can study the implemenation of Linux or basically any modern enterprise platform. An understanding of hardware architecture is required for a few of the low level modules. And an understanding of software architecture is required in order to design modules, for example you need to understand operating system concepts as well as software design using multi-paradigms and patterns.

This would realistically take many years, and in addition, these types of projects are not solitary endeavours, you need to work in a team, but you could be the project manager/director.

Based on the types of things that you would learn along the way to acquring the knowledge, you might end up implementing some other type of system rather than an operating system, the basic software theory can be used in any number of projects.

A decent place to start is college 'computer science' because it introduces you to a lot of concepts.

[cjanscen]

You are looking at programming wrong in general, meaning: The pinnacle of programming is not being able to code an OS....it is being able to code anything that you want to do. And besides, you could be K-or-R-knowledgable and not know how to program an OS, because an OS requires a ton knowledge about OS theory and hardware.
Like the post above said, computer science for a couple years and then think about it...
But also look around, I once came across a tutorial on how to make program a very basic OS using assembly and C.

[Strike]

10 minutes

[Sepero]

Why waste your time creating an inferior OS, when you can help improve an already superior one?


If the goal is only to learn, then yes, the answer is 10 minutes.

[Primus]

Originally posted by Strike
10 minutes

[mart_man00]

Maybe a better answer would be never, why reinvent the wheel?

This scares me alittle.

Did every one know teh world is flat? Eath center of the universe?

Somethings have to be question. Some of the fork of the day stuff is too far.

When the average nerd starts to say that and believe it theres something horrible wrong.

Why do you think the Linux and BSD stayed around so long? The people enjoyed it and questioned the hell out of everything. Plus then theres competition, normally a good thing.

The idea that we will all holds hands and have a nerd-topia is nice, but grow up alittle please. It doesnt work out, its just a nice thought.

A decent place to start is college 'computer science' because it introduces you to a lot of concepts.

Remember, theres a difference between a Computer Programmer and a "Software Developer" or a "Software Engineer". Ask a real CS student(some one that knew what they were getting into and consider MIPS the best thing know to man kind), they will explain.

I really took it the wrong way before someone straighten me out alittle. I hate cheesey descriptions.

[Fryguy8]

heh i'll be glad to explain

a computer programmer writes code, a software engineer deals with the bull**** of a whole team of programmers and spends less than 10% of his time actually writing code.

[mart_man00]

I ment the whole, thinks he knows everythign vs knows everything, but that works

Programmer makes a OS and real apps
Developer is a Indian that uses VB 9-5

[jlmb]

Originally posted by Sepero
Why waste your time creating an inferior OS, when you can help improve an already superior one?


If the goal is only to learn, then yes, the answer is 10 minutes.

What you guys mean by 10 minutes of experience? You guys j/k ? or just pointing out that you'll learn must of the programming knowledge needed down the road?

Making an OS, nowadays, is not worth it if you ask me, is better to learn from what have been done and start from there.

[sharth]

no, 10 minutes of coding

[x_Ray]

Originally posted by mart_man00
This scares me alittle.

Did every one know teh world is flat? Eath center of the universe?

Somethings have to be question. Some of the fork of the day stuff is too far.

When the average nerd starts to say that and believe it theres something horrible wrong.

Why do you think the Linux and BSD stayed around so long? The people enjoyed it and questioned the hell out of everything. Plus then theres competition, normally a good thing.

The idea that we will all holds hands and have a nerd-topia is nice, but grow up alittle please. It doesnt work out, its just a nice thought.


Remember, theres a difference between a Computer Programmer and a "Software Developer" or a "Software Engineer". Ask a real CS student(some one that knew what they were getting into and consider MIPS the best thing know to man kind), they will explain.

I really took it the wrong way before someone straighten me out alittle. I hate cheesey descriptions.

I almost sat down and wrote a long boring response to explain further, but then I realized it would be, well, long and boring. So in short...

Originally posted by mart_man00

When the average nerd starts to say that and believe it theres something horrible wrong.

I guess I'm not your average nerd, damn and I try so hard to

and for everything else:

Originally posted by x_Ray

Just my opinion...

[Sepero]

I concur x_Ray. That whole spiel mart_man00 made seemed a little too leet(or 1337) to me too.

[sharth]

Heck, i thought mart's was better than sepero's. Basically, you need a heck of a lot of time, and you need to PLAN first. If you don't plan ahead and knw what you're doing, half of what you do will just be a hack, and you will end up with code that nobody else can understand, you barely understand, and that nobody wants to help you with.

So, if you want to code an os, read rfcs, find bsd code, and spend about a year trying to figure it out, and then scrap it, start over, and see what you come up with.

Just a thought, I don't know how much code is simply pulbic knowledge (bsd code for example), so I would probably recommend using gpl since if you need a driver for a linksys nic, then its much easier to take written code, understand it, and then write a driver for your os, then it is to reverse engineer it all over again.