Could you write your own operating system from scratch, Sup Forums?

Could you write your own operating system from scratch, Sup Forums?

>BigGuns
w-what is it

Yes, but would take some time since I have a job and thus limited free time to spend on it. I'd rather contribute code to open source projects that I actually use, though.

If you were to pay me to study and work on it full time for as long as it took, yes, probably.

In my spare time, hell no.

Why would you even bother when there are better alternatives.

The real question is are you autistic enough and have nolife to create an OS from scratch?

>Why would you even bother when there are better alternatives
That's not always true though

"Operating System" is a little vague. Some OSs are little more than file management systems.

no, but since this is anonymous and nobody can make me prove it then I will do like these other faggots (who either suffer from the Dunning Kruger effect or are lying like me) and claim that I can.

>Could you write your own operating system from scratch, Sup Forums?
Can and did way back when for course work.

I actually did when 16 yo. Asm/C, had VESA gui, mouse, keyboard support, FAT12 FS, multitasking.

sure, but it's boring as fuck

Artillery simulator where you bomb African villages back to the stone age.

in a z80 system , yes
in a x86 computer not yet

on what system?

Yes planning of making one in motherfucking java

>bomb African villages back to the stone age
They're already there desu

Given enough time, probably. Would I though? Nah.

Fuck off, terry. Get a fucking life.

Question: can I emulate ARM on a x86 computer in order to boot into a homebrew OS which I would develop? I want to write a C compiler and micro kernel that runs bash and a simple file manager and text editor.

Yes, I'm doing that right now.

Even with paranoid schizophrenia people can still be massive attention whores. Fuck off with this piece of shit OS whose only selling point is 'built by one person'

>Why
Fun

Most people here could probably learn how to do it. It just takes some time.

I guess I might try some day. So far I'm writing my own display server and I'll probably write a sound server soon.

that's pretty cool, actually.

>only selling point is 'built by one person'
You don't seem to understand TempleOS.
The main goal of TempleOS is to revive playful programming at a low level, similar to a Commodore 64 or other 1980s home computers. They usually shipped with a coherent programming environment like BASIC and were simple enough to be understood fully.
TempleOS is limited to 100,000 LOC for that reason.

We had to write a kernel in C++ in operating systems class. That was C++ without a standard library, RTTI or exception handling of course.

t. CIA nigger

Genuine question: how useful is it to know how to do such thing?

I'm actually doing that right now, but not as part of a class, just in my free time.

With Qemu yea, but I forgot how to do it

I went through university only picking courses that I thought would be absolutely necessary for a good paying programming job, i.e. "useful" courses. I got the job in the end, but now I'm complete garbage when it comes to stuff that's not cranking out code, such as security or even running a server. If your end goal is just to make money at some big company, there's very little outside of being really good at Java, C# or really good at doing webdev shit that you need to know, at least in this day and age. But if you want to be a good developer, knowledge like how to write your own OS would be really great considering how much other knowledge you'd pick up along the way about how computers work.

I think knowledge like writing your own OS might also be good for working at the big-name tech companies like Google or Amazon, but at that point it depends on whether you want to go through the suffering at working at such places.

like what?

>CPP.Z
go fuck the fuck out of yourself, posting any version of TempleOS without HolyC

Get ARM DS-5

Not with the knowledge I have now, but give me an x86 assembly tutorial that isn't dogshit and enough time, and probably.

>give me an x86 assembly tutorial
welcome to your 20 year tour, enjoy your stay

Yes, and I have done so. A simple one (basically a FreeDOS clone, though I never got around to implementing everything), but an OS nonetheless.

kek

Only a micro kernel for a VM

>Can and did way back when for course work.
what class

Like the system libraries that game developers statically linked against up until PS3/Xbox360/Wii U generation.

Yes, and I'm kind of tempted, but I'd have to design the CPU, GPU, SPU and glue logic first.

Look. Imagine if you took knowledge of RISC-V, DRAM, SCSI/SAS, Firewire and USB-C, modern compression and coding schemes (tagged asymmetric numeral systems, arithmetic coding, MDCT, JPEG, FLAC, Opus; turbo and raptor coding) and reasonable but not completely voodoo (say, 180nm) process knowledge, back in time to the 1970s/1980s. Imagine you got in on Ethernet really early, influencing the early internet to adopt something closer to IPv6 instead of IPv4, and DNS the way Usenet did it.

Imagine you started making an obscure minicomputer workstation, then scaled it down to an early home microcomputer that would look like a cross between the Amiga and the Archimedes.

64-bit, dual core CPU; a glue chipset featuring a fast bus with lots of DMA. Use USB-C as a peripheral socket, but also internally, but have the bus look more like IEEE 1394, attach SCSI with it serially. Make a 120MB superfloppy using a voice coil. Influence early optical drives.

A 32-bit fast float DSP-based soundchip with a bank to that DMA and some internal buffer with a multiplexed oscillator capable of both synthesis, wavetable and sample playback, but also operations capable of operating on multiple channel outputs like phasemod (OPL style), multiplicative FM, ringmod, hard/soft sync, a digital IIR Butterworth multimode filter, out of which you could also make limiters, compressors, gates, delays, reverbs, phasers, flangers, etc, with 48-128 voice polyphony depending on what you're doing. Make it a 16-bit stereo DAC, that'd be fine, could always upgrade it later.

A GPU with a hardware compositor (beyond hardware sprites) capable of going from monochrome, through tiles, bitplaned palettes, to 24-bit true colour if you feel like spending the bus bandwidth; a blitter; and primitive drawing with gouraud shading and texture mapping with optional hardware bilinear filtering taps and antialiasing.

Design your own multisync CRT monitor to display it on. 4:3 is attractive, but why not pioneer 16:10? Have the monitor connector differential twisted pair analogue RGB on an XLR8 connector. There is no real reason you can't do 1920x1200 in nu-1984 with that.

ALPS-switch mechanical keyboard; make your own layout, there's no standard yet. Consign caps lock to the dustbin of history. 3-button mouse with a scroll wheel.

Take all that shit, and make a preemptive multitasking operating system based on a microkernel capable of soft and hard real-time; inspired by UNIX and BeOS, but not either, maybe a bit closer to QNX.

Did that last summer, pretty fun even though debugging was hard at first.

Isn't it kind of nice to be able to throw off your legacy and take a different path? To get lost in one rabbit hole after another?

You'd be able to do vector fonts ahead of anyone else. Sod the Mac, you'd have desktop publishing locked down. You could pre-empt MIDI and push your USB-C/Firewire bus instead, make the first DAW, maybe even ahead of the Fairlight CMI.

What would it look like? What would it sound like? How would you do the OS and/or GUI? What would you program it in? Take something like Rust back to the '80s.

Could I do it? I could do some of it. I could learn other bits along the way. There are esoteric details, like optimising compilers, that are much harder than they appear, but if you're not aiming for perfect...

You could emulate it as if it were a real machine. The Pico-8 enjoyed great popularity. The SPU of this machine would make one hell of a practical softsynth, along the lines of what-if Bob Yannes went all-out on the SID instead of having to do it in a few weeks on shitty old NMOS.

I find the idea strangely enticing. I may actually do the sound part of it.

I find it quite amusing that Sony took 3 attempts and still couldn't get malloc() working right on the PlayStation.