The Linux Foundation banned Charles Max Wood for having supported President Trump's election campaign in 2016

(post is archived)

You are currently inside a comment thread.

Click here to see all the comments (13).

[–] • 0 pt

Let the games begin. The UwU-spamming Microshaft soy-drinking dragon tales watching beta cucks at Linux Foundation will kill Linux, and something new will rise.

Anybody for a completely new kernel standard built on security, privacy, and freedom?

Every good OS movement needs an icon or mascot. Linux took penguins, Apple took Apples, and Microsoft took multicolor flags. I propose that the icon/mascot for this new paradigm in technology should be a dragon. Not the eastern (Chinese) type, but the western type. Need examples? Spyro. Just not too colorful. Why is it a dragon? Because it's going to reduce the pathetic Linux kernel to ashes. We'll create the best kernel in the history of kernels. We'll create an operating system to dethrone Windows itself!

But what should we call it?

link

[–] • 1 pt

(edited )

I've been eyeing making a new "kernel" completely not built around security but very much around freedom. I got started writing some new C tools to make merging projects easier. My jssocial is actually a prototype for the concept except in JS.

The thing is it's not really a kernal concept. I call it "just running code," because there would be no official kernel or even source code that would be exclusively "JuSo OS" (just software OS).

The idea is based on the fact that Linux, because it is written in C, has one namespace for its code, because that's what C projects do, and yet has more people working on it at the same time, and has more stability, and is able to run on more devices than pretty much any other coding project. So the idea is that encapsulation is evil, and the reason most projects can't take on the complexity of an OS is because of encapsulation.

So what the C tools would do is make it so C has the ability to override functions instead of just complaining about a second definition, and then provide a way that people can organize their preferences for priority in a useful way. Then that compiles down into a single binary. The way linux launches is that the boot loader is able to arbitrarily identify a function it wants to start from, so that wouldn't change at all.

It would be smart to use it first to do single purpose applications. Then basic games. And then start looking at executing userland code from it. And all of the parts of the OS could be programmed by different people and swapped out for alternatives much easier than is the case with linux. Saying you want your OS to run a different malloc is as easy as listing a different git repo in the manifest.

A config for a particular install of an OS would just be an ordered list of git resources on the internet, and some environment variables to effect the C preprocessor.

In order of difficulty from most hard to simple is:
Making a bootloader to load it.
Preprocessing C to avoid intersection complaints.
Compiling C to not run in userland.
Grabbing code from multiple sources based on a list.

Most of these are not hard, at all. That's all you need in the tools that will define what JuSo OS is. Then using them to make an actual OS is everyone else's problem. But because it's a single namespace you just take it function by function.

I've been eyeing making a new "kernel" completely not built around security but very much around freedom. I got started writing some new C tools to make merging projects easier. My jssocial is actually a prototype for the concept except in JS. The thing is it's not really a kernal concept. I call it "just running code," because there would be no official kernel or even source code that would be exclusively "JuSo OS" (just software OS). The idea is based on the fact that Linux, because it is written in C, has one namespace for its code, because that's what C projects do, and yet has more people working on it at the same time, and has more stability, and is able to run on more devices than pretty much any other coding project. So the idea is that encapsulation is evil, and the reason most projects can't take on the complexity of an OS is because of encapsulation. So what the C tools would do is make it so C has the ability to override functions instead of just complaining about a second definition, and then provide a way that people can organize their preferences for priority in a useful way. Then that compiles down into a single binary. The way linux launches is that the boot loader is able to arbitrarily identify a function it wants to start from, so that wouldn't change at all. It would be smart to use it first to do single purpose applications. Then basic games. And then start looking at executing userland code from it. And all of the parts of the OS could be programmed by different people and swapped out for alternatives much easier than is the case with linux. Saying you want your OS to run a different malloc is as easy as listing a different git repo in the manifest. A config for a particular install of an OS would just be an ordered list of git resources on the internet, and some environment variables to effect the C preprocessor. In order of difficulty from most hard to simple is: Making a bootloader to load it. Preprocessing C to avoid intersection complaints. Compiling C to not run in userland. Grabbing code from multiple sources based on a list. Most of these are not hard, at all. That's all you need in the tools that will define what JuSo OS is. Then using them to make an actual OS is everyone else's problem. But because it's a single namespace you just take it function by function.

parent
link