OSDev.org

Do some people do OSDev because they want to achieve independence from mainstream platforms as much as possible and to be more obscure?

I admit this is one reason why I want to leaen OS Dev.

Maybe you should learn what a web browser needs from an OS. Your OS will be independently useless if unable to run one.

As a single person, it is impossible to develop a completely independently OS that is also fully functional. But there are efforts like SerenityOS that already have a lot of contributors and are still independent from mainstream software projects. Most other OSes (such as our own OS, Managarm) do not even try to achieve that: they just port existing software to work on top of their core system.

OSDev is a very very big topic.

My previous operating system projects have been for embedded systems (generally microcontrollers, in application specific systems). The main reasons for using a self developed OS has been to avoid licence restrictions, resource requirement reduction (the OS only has the features needed for the application), and when there are a novel set of conditions which need to be met (embedded kernels often have unusual requirements like hard real-time schedulers etc).

Developing an OS for the desktop is really quite a different beast, and my main motivation is to see how well paradigms which work well in the embedded space fit into the desktop space.

So I supposed my reason is intellectually curiosity... Hence why I called my OS CuriOS

For me, I am writing an OS for multiple reasons,
To improve and reimagine Unix
To get away from bloated GNU as much as possible
To get away from the strange creature called the GPL
I don't intend to make everything GNU makes (as least not anytime soon), I mainly want to make a kernel, bootloader, shell, basic set of utilities, a utility to write to a filesystem without mounting anything, a linker and assembler, and eventually, a build system, and a filesystem. Very ambitious of me... But writing an OS to get independence is not a bad reason, that why GNU got started.

I was motivated by a desire to have power over my computers rather than giving it all away, but I didn't really get started with OS dev until I learned to relax. I don't think that's entirely a coincidence. It certainly helps to take breaks by playing games on my Android tablet - with all that implies! It also helps to run WIndows and put up with it, rather than fuss about trying to get Linux to behave in the ways I want. In fact, the last notable bit of development I did was during the 8 months I ran OS X instead of Linux.

I used 9front full-time for about 5 years; it's a small OS by most standards although certainly not single-contributor. During that time, my Linux machine was an appliance for web browsing, socializing through a certain game, and running a couple of other games when I felt like it. More text-based socializing could be done from 9front with the help of Bitlbee gateways. 9front's web browsers didn't really suit me, and in fact the web has become much more inimical and anti-freedom in the last 10 years as web devs rush to use all the web features they can in the name of progress. Some sites aren't so bad. I know a guy who AFAIK still surfs Tumblr in Mothra on 9front, but that was a little too limiting to me even in 2015. Other 9fronters eventually implemented virtualization so they can run a modern web browser in OpenBSD, virtualized. It's not full-scope virtualization; they gave up on Windows guests, but it runs this one relatively large OS (much larger than 9front).

Security issues also doesn't help independent developers. It's not paranoid at all to say that it's no good developing an encryption scheme on your own. You need to work it out with other people who will look at it differently and thus see flaws you didn't. In the world at large there is clearly a degree of paranoia over security but there are also serious issues which are just as annoying to handle. I'm not sure what I'm going to do in my Forth and Plain English Programming systems, other than importing libtls as a binary blob.

My own goal - at least as far as I will admit it - is to research new language, kernel, and document system models, or at least implement some older and largely forgotten ones to see how well they would work on modern hardware. I am hoping to use at least one of those topics for a master's thesis eventually, but so far, I can't settle on any one thing long enough to get anywhere.