OSDev.org

after Searching for a long time a came to the conclusion that none of the quick-start OSDev projects where what i wanted... so i made my own and called it the Operating System Development Kit. its pretty ugly and so far can only say Hello World, and i'm still having trouble with large amounts of text, but it kinda works. basically i started with an old version of the MSVC Kernel FrameWork and removed a lot of the odd stuff like the PE2HOD converter. i added a Multiboot header copied off of a PE kernel tutorial and then used mkisofs to create a GRUB boot ISO which can be run on Boch's. its easy to use, and it only requires Visual Studio.

i know for those of you who are more experience than me (i only have one year of experience) might think of a million things i can do better with it, if so please let me know. i would love to have someone more experienced on the dev team.

the basic intent is to have an easy to use starting point for making any kind of OS you wish. weather it is just a Hello World OS or a full graphical OS it should be a good starting point with a decent libc.

you can read more about it on the Project Site

Rather lofty goals, but we'll see..
Just keep in mind though, most OSDevers use Linux, including myself.

I don't exactly understand the concept... I guess what confuses me most is that this project should aide a developer in building any OS that he might conceive, but the kit includes a prebuilt libc? What if someone wanted to, for whatever reason, make an OS without libc support, or otherwise wants to use an incompatible variant of libc?

In short: how do you make a development kit that doesn't impose any particular code segments in the end result? Wouldn't you, instead, want to make an IDE suited for OS development?

KotuxGuy wrote:Just keep in mind though, most OSDevers use Linux, including myself.

Prepare to stand corrected (although you're right - an "OS Devkit" should be somewhat cross-platform): http://forum.osdev.org/viewtopic.php?f= ... ux#p153122

@OP:
The problem I see is that the point of hobby OS development is in some ways to reinvent the wheel, but reinvent it by making it completely different, so trying to make some sort of common framework is, oddly enough, against the grain. The greatest common factor of "whatever OS you wish" is an empty text file. Not to mention that pre-written code is probably not going to help newbies nearly as much as a step-by-step tutorial like JamesM's tutorial or bkerndev.

i agree that reinventing the wheel is really the main goal in many ways, however my goal is not to make a full operating system or even any type of operating system, i just want to create an environment for OSDev newbies so anyone can easily compile the there Operating system into a bootable ISO which they can then test however they wish.

i actually dual boot Linux and Windows so i might work on a version for Linux to if i could just find a good IDE for Linux. so far the only one i can find that is similar to visual studio (by similar i mean it has a full development environment with separate projects) is Code::Blocks. this isn't quite as nice as visual studio though.

my goal is similar to that of what OSKit was designed for except i try to keep things a bit more user friendly. in short i want to make something that is designed for quick development for experimentation or a full OS. the developer could change the assembly code starting point if they wish or if they don't want to mess with assembly, they can leave it. take a peak at the code to see what i mean. the only confusing part is the Assembly code which is why i tried to do a minimum of assembly code.

I've considered making a OS Dev Kit similar to your ideas(note, there does exist a really old thing called OSKit which includes a lot of your goals, but I believe the code was all GPL and it's a dead project since like 2001).. My main problem was just the amount of different ways someone could make their OS work.. With a OS Dev Kit you must abide to the way it expects an OS to work, which is not how most Hobby OSs work.

thats why i was mostly planning on making things like keyboard input and text output work. i'm not really planning on making anything that can't be changed. i'm pretty sure that most OSDevelopers though don't care much about how low level Text output is implemented so that could be almost fully implemented. keyboard and mouse input on the other hand does change from OS to OS so that would probably be pretty loosely implemented. i'm glad everyone has some input on the subject, i will be sure to keep all this in mind. because of this i might not implement a full libc, but instead just the beginnings of one so that it can be written by the developer.

Hi,
There is a similar project , Oskit
--Thomas

Thomas wrote:Hi,
There is a similar project , Oskit
--Thomas

It is old and dead by now. But its code is good, clean, provide wide range of support for many hardware, FS, other stuffs.