Diseased Open Source Software Community - it's about ethics in Code of Conducts

[AmpleApricots]

I'm at a point where I'll just claim the very inflammatory statement about systemd that whoever thinks it's a genuinely sane, useful program has clearly not understood anything about the advantages of a heterogeneity in a software landscape, and in a wider sense about unixoids or the unix way, at all. And yes, I am willing to die on that hill and no, I don't care that copy/pasting systemd unit scripts from Stack Exchange without understanding what's happening makes your life as overpaid tech-soy easier. Also the corpo Red Hat (now owned by IBM which the young whippersnappers don't know, have literally invented the concept of the hardware lock-in and walled garden, Apple learned from them) pushed systemd and the maintanance nightmare that thing is hard. As other people said in this thread, Red Hat's sole goal is to own the core of Linux Userland in an effort to control that entire ecosystem. They haven't even been particularly coy about it in recent years. Tranny distro jannies are the useful idiots for their corporate overlords, as usual. And once again, they're probably even doing it for free.

I remember the ARM Mali/Bifrost GPU drivers of that M1 guy being celebrated but last time I tried them (which is admittedly a few years ago) they were barely working and pure shit performance-wise. They didn't even support putting the GPU into an idle state which I'd consider basic functionality and an absolute priority to implement before all else. I would not be surprised if they've never moved on from that barely working state and even though tons of so-called Linux devs absolutely love their Apple black boxes and don't actually use linux seriously, I have my sincere doubts the Apple Silicon will ever be well implemented in Linux. Not because it isn't possible, but because Apple will keep moving the goal posts with new hardware and these people just cannot help themselves running after the newest shiny instead of finishing their work in a consistent way. I'd not be holding my breath. I don't think there will ever be a truly useful ARM desktop that'll be genuinely better to use than x86 and I have no idea why Torvalds is so blind to that reality. We might get one last shot with RISC-V before we're walled in non-free hardware forever. Linux users have this weird Stockholm syndrom with poorly written drivers for exotic, non-x86 hardware where they claim they're just dandy but when you actually investigate and look at it objectively, you'll figure out that nothing really works. I'd always test myself before believing that "X is well supported". More often than not, besides bare-bones hardware support where all has to be done by raw CPU power, it simply is not the case.

To the sound discussion: I wrote my asound.conf many many years ago and just made slight changes with new speaker configurations and such. The only thing that truly is a pain in the arse is bluetooth, which is kinda inherently shit and doesn't really work properly in any OS. (and yes, I use bluetooth headphones/speakers with alsa only) JACK, Pulseaudio, soundwire etc. are all just layers on top of alsa and I while I agree that some aspects of ALSA are bad, layering more code on top of it can never be the solution.

 

[Dread First]

I'm at a point where I'll just claim the very inflammatory statement about systemd that whoever thinks it's a genuinely sane, useful program has clearly not understood anything about the advantages of a heterogeneity in a software landscape, and in a wider sense about unixoids or the unix way, at all. And yes, I am willing to die on that hill and no, I don't care that copy/pasting systemd unit scripts from Stack Exchange without understanding what's happening makes your life as overpaid tech-soy easier.

I don't even fucking understand why there's all this autism over init systems/service managers. When I first got started with Linux in high school, the big init systems were SysV, Upstart, and systemd right around the corner with Fedora being the only distro running it. From my teenage perspective, there really was no fucking difference in terms of performance between any of the init systems once it's all installed, and you're no longer regularly fiddling with it to fix something that went awry during the migration process.

I am not skilled enough to talk about the technical merits of systemd, OpenRC, Upstart, SysV, or anything else of the sort. But what I will say is that the collective autism of people like Lennart Poettering et al. could've more than likely be better spent on relevant shit that Linux is still terrible at a decade and change later. Y'know, like not regressing in terms of functionality on every X.0 release and waiting for an X.0.1 or X.1 update to fix all the bugs that should've been caught in beta testing.

Also the corpo Red Hat (now owned by IBM which the young whippersnappers don't know, have literally invented the concept of the hardware lock-in and walled garden, Apple learned from them) pushed systemd and the maintanance nightmare that thing is hard. As other people said in this thread, Red Hat's sole goal is to own the core of Linux Userland in an effort to control that entire ecosystem. They haven't even been particularly coy about it in recent years. Tranny distro jannies are the useful idiots for their corporate overlords, as usual. And once again, they're probably even doing it for free.

IBM killed off any good will that they had left from the community when they shitcanned RHEL8's entire life cycle because they wanted to "rebrand" CentOS. On the flip side, that was my introduction to Rocky Linux and their community is absolutely a joy to be around. Then again, making the best of a bad situation tends to highlight the good in us all.

Linux users have this weird Stockholm syndrom with poorly written drivers for exotic, non-x86 hardware where they claim they're just dandy but when you actually investigate and look at it objectively, you'll figure out that nothing really works. I'd always test myself before believing that "X is well supported". More often than not, besides bare-bones hardware support where all has to be done by raw CPU power, it simply is not the case.

This is the biggest reason why I've been trying my hardest to migrate over to a *BSD userland, right down to only using hardware that the FreeBSD project fully supports. 2010-2011 might have been a little over a decade ago now, but I can surely attest to being an autistic teenager dealing with crappy wireless and printer drivers that were most likely updated last in 2007/2008. FreeBSD 13, which came out last year, finally added proper Xbox 360 (but not Xbox One) controller support. Third-party drivers have existed for years, but the official FreeBSD drivers are actually stable and lack the quirks of xboxdrv on Linux. FreeBSD takes a metaphorical eternity to add hardware support, but when it's implemented, it well and truly works. It ain't "barely functional rubbish" as is so often the case with Linux drivers.

To the sound discussion: I wrote my asound.conf many many years ago and just made slight changes with new speaker configurations and such. The only thing that truly is a pain in the arse is bluetooth, which is kinda inherently shit and doesn't really work properly in any OS. (and yes, I use bluetooth headphones/speakers with alsa only) JACK, Pulseaudio, soundwire etc. are all just layers on top of alsa and I while I agree that some aspects of ALSA are bad, layering more code on top of it can never be the solution.

This extremely long-winded blog post about the OSS/ALSA split was written in 2007. It's haunting to me that nothing has changed in the scope of things relating to audio on Linux in well over 15 years. I'm still not smart enough to work out how to do OSS shit, but I'm more than willing to give OSS the old college try than I am ALSA, JACK, Pulseaudio, etc.

 

[Rusty Metal Skull Gun]

They didn't even support putting the GPU into an idle state which I'd consider basic functionality and an absolute priority to implement before all else.

In a similar vein, I read a Phoronix article yesterday that said that Nouveau drivers for Nvidia cards were working, but very slow because they don't support reclocking the card. Which has been a problem for over 10 years now! I don't get why this is such a humongous problem... just run the proprietary driver through IDA Pro, and claim you found it through chance. Or do passthrough in Qemu, and get it logging what it does.

Tin-foil hat time, but I sometimes wonder whether these projects are deliberately sabotaged somehow by the company, to keep you using their proprietary stuff. ReactOS is another example: it's 27 years old, but still utterly useless.

 

[Maldavius Figtree]

Can you autistic faggots stop sperging out about systemd and whatever else, this is not Reddit or hacker news, you’re not impressing anyone and taking up space that could be used for laughing at trannies.

Regarding the stable diffusion webui controversy, automatic1111 is undeniably a racist and edgelord (both good things). He’s made rimworld mods making fun of George Floyd, game mods to remove black people, and one of his old steam usernames is „remember the 6 gorillion“.

 

[Rusty Metal Skull Gun]

If you took away the autistic sperging from the Open Source thread, there'd be nothing left.

I don't think there will ever be a truly useful ARM desktop that'll be genuinely better to use than x86 and I have no idea why Torvalds is so blind to that reality.

Well no, there won't ever be, unless something happens so that RISC chips are massively faster than x86. ARM's main advantage is its low power consumption, and when it comes to desktop no1curr. x86's thing is backwards compatibility, which is massively important.

 

[byuu]

Damned if I can get my bluetooth audio to work with anything but PulseAudio. I still miss OSS on OpenBSD.

BlueZ dropped support for ALSA a while ago.
Which is so fucking shit in general that I have given up on using bluetooth devices in Linux.
Non-bluetooth wireless headphones are much better anyway.

 

[The Ghost of Kviv]

I don't think there will ever be a truly useful ARM desktop that'll be genuinely better to use than x86 and I have no idea why Torvalds is so blind to that reality. We might get one last shot with RISC-V before we're walled in non-free hardware forever.

ARM desktops are close enough for a lot of folks. If all you do is emails, web browsing, watching videos, listening to music and document editing, ARM is enough. The biggest issue is graphics drivers, but it's easy for companies like AMD to modify the memory and I/O to slap on embedded graphics.

RISC-V is not going to succeed because it's <i>libre</i>, it's popular because nobody needs to pay license fees to make custom controllers. Nvidia uses RISC-V on their GPUs for power regulation; if they had to license it, they would just fall back to some other cheap ISA like MIPS.

ARM's main advantage is its low power consumption

Sort of, low power consumption because it's a smaller ISA that isn't afraid to throw out old features. Head to head using modern feature sets, Zen and ARMv8 are in the same ballpark in power consumption, this example is on Windows, with its usual bloat:

 

[Dread First]

Well no, there won't ever be, unless something happens so that RISC chips are massively faster than x86. ARM's main advantage is its low power consumption, and when it comes to desktop no1curr. x86's thing is backwards compatibility, which is massively important.

To play devil's advocate here, we haven't undergone a substantial computer architecture shift for almost two decades (even longer if you include everything before x86_64). x86 at this point is so thoroughly optimised, documented, and iterated upon that it's now impossible to ignore its flaws. I think it's highly likely that a transition to ARM won't be so bad when you factor in the boosts to ARM that the Raspberry Pi platform, Apple's MX chips, and future offerings from AMD and Intel have given it.

The real issue with ARM platforms is the huge variety in implementation. Phone ARM is fundamentally different to desktop ARM because desktop ARM requires more power and permissions to match the same parity as x86. Phone ARM platforms, by stark contrast, are much more likely to remain walled gardens due to practical security concerns.

I think we're gonna enter a world where the same processor that powers my PC will inevitably make its way to my phone, and I won't be able to achieve 1:1 feature parity on both. Quite a surreal thought to ponder.

 

[ZMOT]

Can you autistic faggots stop sperging out about systemd and whatever else, this is not Reddit or hacker news, you’re not impressing anyone and taking up space that could be used for laughing at trannies.

we're just laughing at systemd users, because they more or less trannies. it's 2-for-1!

Tin-foil hat time, but I sometimes wonder whether these projects are deliberately sabotaged somehow by the company, to keep you using their proprietary stuff. ReactOS is another example: it's 27 years old, but still utterly useless.

given the zeal of the average nvidiot, I wouldn't put it past them. but then I think that's mostly a /v/g/ thing. otoh troons like to go full in on brand identity which could make up for it. maybe worth looking into.

as for reactos, I think it's a mix of lower demand, navigating around potential legal issues (if not outright sabotage), and windows being windows.

 

[The Mass Shooter Ron Soye]

Some new cringe just dropped. I think this is the best thread for it.

Native Americans ask Apache foundation to change name

Open source org called out for ignoring own code of conduct

www.theregister.com

Native Americans ask Apache foundation to change name • The Register

archived 13 Jan 2023 11:38:14 UTC

archive.ph

Native Americans urge Apache Software Foundation to ditch its name • …

archived 13 Jan 2023 11:56:16 UTC

archive.ph

Indigenous tech group asks Apache Foundation to change its name

Identity as a "last tribe" is simplification and erasure, indigenous group says.

arstechnica.com

Indigenous tech group asks Apache Foundation to change its name | Ars…

archived 12 Jan 2023 23:55:25 UTC

archive.ph

Make sure to check out the comment sections.

 

[AmpleApricots]

I don't even fucking understand why there's all this autism over init systems/service managers.

Honestly, I have no idea. I feel I would give less of a shit if the systemd advocates wouldn't be like a of trannies that have a meltdown every time you don't respect their pronouns. Also systemd has this knack of invading other core software components and make them dependent on it for no good reason. That all said, it's still pretty easy to avoid so I'm not really frothing at the mouth over systemd.

This is the biggest reason why I've been trying my hardest to migrate over to a *BSD userland, right down to only using hardware that the FreeBSD project fully supports.

I tried FreeBSD the other day and the wireless and GPU drivers are direct linux imports. I liked FreeBSDs consistency and avoiding of shitty software but the Kernel's hardware support sadly is nothing to write home about. In general, I actually do like the Linux kernel. It has a lot of cool features and has resisted influence of the usual suspects pretty well so far. It's also in fact relatively easy to avoid a lot of the shitty userland software. Pity is that it is necessary and the general trend is towards that shitty software.

BlueZ dropped support for ALSA a while ago.

It works alright, but there are some caveats

ARM desktops are close enough for a lot of folks.

ARM usually equals "low power" in my mind too but I have a Kaby Lake core M Thinkpad tablet here that actually uses less power on average than most current high end ARM tablets while not being really noticeably slower in reality for most of the common tasks you listed. (and still being faster than the common ARM SBCs that actually can run Linux) Contrary to these high end ARMs, it also actually can run any common OS, including Android if I really want to. There's no such thing as free lunch, no matter the architecture. If you want fast, you need to pump the watts into it. Look at the power consumption figures of the M2 chip if you can find them. The numbers are impressive, but not *that* impressive. People for some reason completely ignore or belittle the low end x86 offers, but they're quite competitive. They just usually end up paired with shitty, cheap hardware with too little ram, slow eMMC and such. If you actually pair a low power x86 with decent hardware like in this tablet, it's a pretty smooth experience. It just rarely happens for market reasons.

I stick to it, ARM is not really gonna go anywhere in normal desktops. They'd at least need to adapt a standard like UEFI. This is not gonna happen since the ARM SoC makers do not care about desktop, which is a tiny market compared to the disposable smart device of the month. I'll take my puzzle pieces now.

Some new cringe just dropped. I think this is the best thread for it.

Based indians. If you want to add another layer of funny, google who holds the leadership positions in many of these tribal organizations. They're surprisingly often a bunch of old, incredibly white dudes.

 

[Mikoyan]

Some new cringe just dropped. I think this is the best thread for it.

Native Americans ask Apache foundation to change name

Open source org called out for ignoring own code of conduct

www.theregister.com

Native Americans ask Apache foundation to change name • The Register

archived 13 Jan 2023 11:38:14 UTC

archive.ph

Native Americans urge Apache Software Foundation to ditch its name • …

archived 13 Jan 2023 11:56:16 UTC

archive.ph

Indigenous tech group asks Apache Foundation to change its name

Identity as a "last tribe" is simplification and erasure, indigenous group says.

arstechnica.com

Indigenous tech group asks Apache Foundation to change its name | Ars…

archived 12 Jan 2023 23:55:25 UTC

archive.ph

Make sure to check out the comment sections.

Amazing. I was just thinking the other day "Why haven't they come for Apache yet?"

 

[UERISIMILITUDO]

Some new cringe just dropped. I think this is the best thread for it.

I just today finished reading 1984, and I'm not trying to seem deep with this post, but I can't help but think of such things as similar to one of the aims of Newspeak: To remove as many meanings from as many words as possible.

 

[The Mass Shooter Ron Soye]

ARM usually equals "low power" in my mind too but I have a Kaby Lake core M Thinkpad tablet here that actually uses less power on average than most current high end ARM tablets while not being really noticeably slower in reality for most of the common tasks you listed. (and still being faster than the common ARM SBCs that actually can run Linux) Contrary to these high end ARMs, it also actually can run any common OS, including Android if I really want to. There's no such thing as free lunch, no matter the architecture. If you want fast, you need to pump the watts into it. Look at the power consumption figures of the M2 chip if you can find them. The numbers are impressive, but not *that* impressive. People for some reason completely ignore or belittle the low end x86 offers, but they're quite competitive. They just usually end up paired with shitty, cheap hardware with too little ram, slow eMMC and such. If you actually pair a low power x86 with decent hardware like in this tablet, it's a pretty smooth experience. It just rarely happens for market reasons.

I stick to it, ARM is not really gonna go anywhere in normal desktops. They'd at least need to adapt a standard like UEFI. This is not gonna happen since the ARM SoC makers do not care about desktop, which is a tiny market compared to the disposable smart device of the month. I'll take my puzzle pieces now.

This discussion looks off-topic but whatever. ARM's benefits are overrated and overhyped, and don't overcome x86's advantages for desktop use. RISC-V will have the same problems, but even worse since it's more of a free-for-all. Apple Silicon was severely overhyped and had a full TSMC node advantage or more over the x86 options it was compared to. That doesn't make Apple wrong for switching from Intel to their own chips, but it affects the fanboy discussions about how Apple Silicon is burying x86. AMD's Phoenix APUs on TSMC 4nm could compare favorably to M1/M2.

Better cheap new x86 SBC/mini PC options would be a killshot for RPi-as-desktop, and RK3588. AMD can't get the supply it needs or is uninterested in selling cheap chips anymore, so that's a problem. Mendocino looks like it will be overpriced, and those 4700S desktop kits made from defective console chips didn't have integrated graphics and were overpriced crap. There was this report about sourcing from Samsung's old nodes but it's not promising.

Intel has had interesting options in the market like Cherry Trail, later cheap Atoms, and recently Jasper Lake i.e. in ODROID-H3 looks pretty good and should always be a better pick than RK3588. Alder Lake-N is questionable. Premium branding for the 8-cores with high prices, and they slashed the memory channels in half. It has me looking for discounted Alder Lake-U instead, which has the benefit of hybrid x86 delivering more single-threaded performance. Intel dominates the good used and refurbished options, because they ship so much volume to businesses. Now that AMD's desktop chips are all including an iGPU (so far), they could be retired into tiny Mini-ITX boxes in the future (just limit the TDP).

Gimme dats

Amazing. I was just thinking the other day "Why haven't they come for Apache yet?"

They came, they saw, and soon they're going to conquer.

 

[Dread First]

As much as I like the Apache Licence 2.0, the Apache Foundation has been fucking abysmal with maintaining anything that isn't the web server. Does anyone else remember in 2010 when Oracle bought out Sun Microsystems, and they discontinued OpenOffice.org? Oracle gave the codebase to the Apache Foundation, and it was hosted on Apache Incubator for like... two years while LibreOffice was out and about gaining traction.

 

[The Ghost of Kviv]

As much as I like the Apache Licence 2.0, the Apache Foundation has been fucking abysmal with maintaining anything that isn't the web server. Does anyone else remember in 2010 when Oracle bought out Sun Microsystems, and they discontinued OpenOffice.org? Oracle gave the codebase to the Apache Foundation, and it was hosted on Apache Incubator for like... two years while LibreOffice was out and about gaining traction.

They also were supposed to maintain Google Wave. They forgot it existed for nearly 10 years, and dropped it from it's list recently.

 

[Coolio55]

They also were supposed to maintain Google Wave. They forgot it existed for nearly 10 years, and dropped it from it's list recently.

To be fair, Wave was kind-of poorly designed in a lot of respects.
All the sub applications in the promo video (multiple choice polls etc) relied on random companies continuing to host them forever or else all the content would be permanently lost.
t. skimmed the source code.

 

[teriyakiburns]

I remember trying Wave when it first emerged. It was slow as shit and a portent of things to come, given how much chatter it emitted back to google, every time you interacted with the window in any way. One ten minute session was enough to convince my little social circle to never use it again.

 

[Dread First]

They also were supposed to maintain Google Wave. They forgot it existed for nearly 10 years, and dropped it from it's list recently.

Good God, and here I thought the Mozilla Foundation was fucking awful with maintaining projects... at least they spun off Thunderbird into its own administrative entity and it's now able to properly develop and expand.

 

[Rusty Metal Skull Gun]

As much as I like the Apache Licence 2.0, the Apache Foundation has been fucking abysmal with maintaining anything that isn't the web server. Does anyone else remember in 2010 when Oracle bought out Sun Microsystems, and they discontinued OpenOffice.org? Oracle gave the codebase to the Apache Foundation, and it was hosted on Apache Incubator for like... two years while LibreOffice was out and about gaining traction.

They blatantly should have given the OpenOffice trademark to LibreOffice but didn't, for reasons nobody can quite fathom.

The Apache Foundation is a joke. It's the server, what's essentially an old version of LibreOffice, and wave upon wave of Java shit.