Conversation
Notices
-
Embed this notice
寮 (ryo@social.076.moe)'s status on Saturday, 07-Jan-2023 10:22:18 JST 寮 Technically, ARM is a RISC clone (not the other way around), but locked down and loysensed out to whoever pays them to use it.
It's just that RISC started gaining traction just recently.
Though not sure if I'd really support GNU in finishing Hurd after them having stalled it 31 years ago in favor for the Linux kernel, because even though Stallman is seemingly more tech freedom aware than Torvalds is, every single GNU tool is bloat, so I'd expect Hurd to be bloat as well when that gets developed again.
But I can sympethise with the Linux kernel being so bloated in a way, because there's so much more hardware today than there were a decade ago, they all require a separate driver which are all probably very bloated too, and on top of that most of them are proprietary blobs or poor attempts on rewriting the driver too.
But I could be wrong on that, because I need a proper comparison between Linux and Linux-libre in order to actually be sure.-
Embed this notice
Terminal Autism (terminalautism@social.076.moe)'s status on Saturday, 07-Jan-2023 10:22:19 JST Terminal Autism Yeah, hardware massively complicates this. Though ARM is no better, because ARM computers are special snowflakes and are completely closed, and you pretty much get what you get and can't expand it at all, it's worse than x86. I hope RISC-V catches on and does better. Hell, I hope GNU makes the Hurd kernel work at least on it, if nothing else, because as soon as free hardware becomes available, there will be no reason for them to focus on anything else.
But it's a real shame that things ended up like this. It's another example of how everything in computing tends to go in the worst possible direction. Everything that becomes standard is almost inevitably the worst shit that happens to be available. We could have PowerPC desktops now, but nope, it lost to the x86 and got wiped out. Maybe we could have SPARC, but it was never used outside of servers and workstations.
By the way, both OpenSPARC and OpenPOWER have OpenFirmware (Check it out, it's very cool: https://yewtu.be/watch?v=KvxxAeuhPp0 ), which is built into an implementation of Forth. You can run programs directly on the firmware, no other additional OS required. Why don't we have that? Hell, why aren't OSs just built on top of powerful languages like that in the first place, why all these layers of complexity? From what I have heard, a window manager has been written for it, and even a Forth-based Emacs clone and some games. It's just ridiculous that this is not possible on every single computer. To me it seems like the things that are allowed to become popular on the consumer market are deliberately picked specifically to restrict the user, and to prevent individuals from being able to do too much themselves.
You can also see the amazing things that were done with MIPS CPUs in 90s and early 2000s consoles. But nope, x86. And the only alternatives are shitty locked-down ARM single-board computers that can't be expanded and that run like, two OSs, if you're lucky. And with no standardization, so each OS has to support each device individually, on top of the OSs themselves being clunky and barely even designed.
Also, don't forget Mezzano ( https://github.com/froggey/Mezzano https://yewtu.be/watch?v=Wd_-h5kRQLo ). It's an OS written entirely in Common Lisp. Still only runs on VMs as far as I know, but it's very impressive that it exists in this environment of very complicated computers and also nothing but Unix clones everywhere, and also an architecture that wasn't exactly built for that type of language (probably quite the opposite). I assume not made by many people either, and if that's the case, they must be insanely productive. https://social.076.moe/url/50258
-
Embed this notice