What went wrong?

File: jewcloud.png (34KB, 760x960px) Image search: [Google]

34KB, 760x960px

Cloudkikes and social/mobile walled gardens made concept of computing freedoms of irrelevant.

Linux tries to solve problems nobody perceives as a problem anymore.

You posted an image of a sad penguin. Penguins are happy creatures, Anon. We penguins are all happy here.

File: IT'S_OGRE.png (1MB, 1078x683px) Image search: [Google]

1MB, 1078x683px

Poorly coded buggy kernel.

>>57234139
the kernel is good.

the community and all distros suck because of autists not knowing how to UX and ubuntu being kinda dead on the desktop side (when will they get their software manager to not lag af?!?!?)

>>57234139
Weebs poisoning what was once a good community.

>>57234139
Too easy to fork projects

2017, year of the Linux Desktop

File: bkL5gI1.jpg (79KB, 800x383px) Image search: [Google]

79KB, 800x383px

>>57234181

>>57234139
have shitton of bullshit, like 1000 distros, 25 DE, 14 WM...
Why cannot Linux be one OS ?

File: meme_os.png (51KB, 1401x628px) Image search: [Google]

51KB, 1401x628px

Too many distros. If loonix basement dwellers would have focused all their energy on 2-4 distros then linux would have more marketshare than homOS X by now. Sadly they didn't and all this fragmentation has pretty much made linux unusable as a desktop OS. This is ignoring things like having to use a CLI for simple tasks which most people have a ton of trouble doing.

GNU code is bloated and buggy.

Poor architectural choices like ALSA and systemd.

Most distributions having poor defaults.

Insufficient enterprise solutions and management tools.

Subpar development tools and SDKs drive developers away.

End result: it got relegated to applications where any cheap, quick and dirty solution will do, like web servers, number crunchers, cellphones and embedded boxes. Linux is kludgey and not a serious operating system by any criteria.

>>57234299
>pass

File: share.png (106KB, 1695x415px) Image search: [Google]

106KB, 1695x415px

>>57234139
linux or BSD won in every domain except desktop, which is a dying market

desktopfags are unable to accept that almost no one gives a fuck about them except for the tiny demographic of PC gamers that DirectX has captured

>>57234372
Go back to your desktop thread, jimmy. Don't forget to clean your basement for gbp.

>>57234344
>99.8% of the world's supercomputers
>a kludge
lol

File: penguin-on-an-iceberg.jpg (958KB, 3910x2612px) Image search: [Google]

958KB, 3910x2612px

>>57234139
global warming

>>57234386
>I'm sucking dicks badge next to my post

>>57234139
Drivers in the kernel space.

>>57234404
You are retarded if you think they run any publicly available distribution.

Bad marketing and investment capture.
I mean just look at MacOS, it's what linux should be with funding.
But instead you've got a clusterfuck of branched ;;distributions;;
that clusterfuck is just the result of a failed product
>ok im shit lets just let everyone have it for free

>>57234404
Could easily be done with OpenSolaris now, Since all of the pre-2000 super computers ran solaris already.

>>57234415
There is nothing wrong with that.

File: 1469553116195.jpg (233KB, 719x685px) Image search: [Google]

233KB, 719x685px

>>57234494
>Drivers in the kernel space.
>There is nothing wrong with that.

Retard community that doesn't understand that 99% of users need a good GUI

>>57234181
>implying windows is any better
>implying BSD is any better
>implying solaris is any better

>>57234404

As I said, mere number crunchers.

>>57234516
Rather have drivers then fonts in the kernel space.

>>57234558
There is nothing wrong with that.

>>57234558

That's a red herring, even if Windows' architectural decision was unjustified, that wouldn't justify Linux' poor decisions. Try again.

>>57234533
>This is an ancient bug that was actually attempted to be fixed once (badly) by me eleven years ago in commit 4ceb5db9757a (“Fix get_user_pages() race for write access”) but that was then undone due to problems on s390 by commit f33ea7f404e5 (“fix get_user_pages bug”).

>In the meantime, the s390 situation has long been fixed, and we can now fix it by checking the pte_dirty() bit properly (and do it better). The s390 dirty bit was implemented in abf09bed3cce (“s390/mm: implement software dirty bits”) which made it into v3.9. Earlier kernels will have to look at the page state itself.

http://www.cyberciti.biz/faq/dirtycow-linux-cve-2016-5195-kernel-local-privilege-escalation-vulnerability-fix/

>>57234558
Don't bother responding to him, anon. Nobody here, at all, is possibly using a microkernel to do general-purpose work. At best, they could be using an ancient version of Linux hosted under a microkernel, and even that's unlikely.

>>57234260
>>57234299
Freedom is about choise. Remember this. Is never too many distros because 80% of this is debian with different wallpaper.

>>57234299
The issue with this argument is
>most distros are ricing and out of the box shit
>downstream contribution is nonexistent
>many built upon debian or arch etc.

If you want to find wasted work hours then look around DEs:
https://news.ycombinator.com/item?id=12578204
Thanks to GNU and Red Hat other part of the system is barely fragmented.

>>57234494
It's the source of the most lunix bugs.

>>57234530
>implying most of the users wouldn't be ok with linux. Many of them just browse the facebook, make prints and small docs

>>57234575
windows had a CVE thanks to that.

>>57234533

Windows is shitware, but BSD is at least on par with Linux, and Solaris is way more advanced.

I can think of 3 better kernels just off the top of my head: XNU, seL4 and Minix.

Consider taking operating systems.

>>57234533
The summary of my impression [after reading the OpenSolaris code] was that I was... surprised....the [OpenSolaris] code, as I saw it, was neat. Real neat. Extremely neat. In fact, I found it painful to read after a while. It was so neatly laid out that I found myself admiring it. It seems to have been built like an aircraft. It has everything that opens and shuts, has code for just about everything I've ever seen considered on a scheduler, and it's all neatly laid out in clean code and even comments. It also appears to have been coded with an awful lot of effort to ensure it's robust and measurable, with checking and tracing elements at every corner. I started to feel a little embarrassed by what we have as our own [Linux] kernel. The more I looked at the [OpenSolaris] code, the more it felt like it pretty much did everything the Linux kernel has been trying to do for ages. Not only that, but it's built like an aircraft, whereas ours looks like a garage job with duct tape by comparison....[OpenSolaris] looks like an excellent design for a completely different purpose. It's built like a commercial design for commercial purposes that have very different requirements than what most of us use Linux for, but it does appear to have been done so very well. It looks like a goddamn Star Destroyer, and the Linux kernel (scheduler) suddenly looks like the Millennium Falcon. Real fast, but held together with duct tape, and ready to explode at any minute.

>>57234578
I'm aware of the bug, and what it is. If there's a bug in an open-source OS that has been the target of more research and security scrutiny than any other OS in history (except perhaps seL4, which can't do anything general-purpose without hosting a VM), how do you expect closed-source or less-popular systems to be better?

>>57234613

>real fast

That's pretty much the only good thing about Linux. That's why it's ubiquitous on supercomputers. They don't need quality, stability or security. They just need raw speed.

>>57234645
Linux is stable and enough secure too. I don't know what do you mean by "quality" and my guess you either.

>>57234645
How many supercomputers do you run?

>>57234613
The sounds an awful lot like a macfag pretending to be a microshill talking about macs and OSX...

>>57234683
quality is the ability to be instrumented, traced and debugged. Things that are easily done in Unix but no Linux. Ever wonder why dtrace isn't part of Linux?

>>57234260
>have shitton of bullshit, like 1000 distros, 25 DE, 14 WM...
>Why cannot Linux be one OS ?
>Why cannot there only be one car manufacturer? I require a small city car with two seats and that's what they should be focusing on so I get the best possible car.
So how are you going to deny people from making alternatives? Nuke from orbit?

If you want to make a change you start using the one that's best for you, so you support it. Then the one that's best for most people will eventually be the most popular. The most popular will get developed since people realize it's important to focus on keeping it active and working.

>>57234701
except its from a linux kernel dev.

>>57234645
>Stability
As far as I'm aware this hasn't ever really been a major issue for Linux
>Security
Only people who call Linux insecure are clueless people and the BSD crowd, who build all sorts of userland breaking security mechanisms into their OS not to counter any realistic threats, but just for the sake of sake of making it more secure.

File: 1477299522642.gif (1006KB, 500x300px) Image search: [Google]

1006KB, 500x300px

>>57234715
http://akari.osdn.jp/

Ubuntu 10.1 went and sold the transparency feature to Microsoft, and now "Multiple desktops".

what was it, 12.1 when Linux got baugt out by the bankz?

>>57234687

Just two currently at my uni's applied math and computational physics lab. They're Crays running on CLE. It is a heavily modified Linux with a PBE scheduler to make it suitable for supercomputing, because the default scheduler is dogshit.

>>57234735
Ok... There's loads and loads of those people with skills and knowledge from end to end. The least you could do is point to who we're talking about rather than trying to portray this as the opinion of the linux kernel dev community as a whole.

Also, try to remember that OpenSolaris is pretty much dead by now, XNU (the OSX kernel) is pretty awful in many places and the NT kernel isn't much better than XNU.

>>57234797
>>57234687
PBS*

>>57234809

He was talking about the scheduler in case you didn't notice.

>NT kernel
Best memory management around.

>XNU
Architecturally one of the best out there.

>>57234797
You do realize that the default scheduler isn't meant for supercomputing tasks? Pretty much all supercomputers (or at least those run by people who know what they're doing) are using custom schedulers made specifically for supercomputing.

>>57234837
>he
DID YOU JUST ASSUME THEIR GENDER?!?!?!

>>57234848

No shit, Sherlock. Did you even read my post?

>>57234608
I've taken operating systems, and it turns out that we spent more time talking about how to build systems, not comparing various microkernels designed for intellectual masturbation. For XNU and solaris, you're trying to make the "macs don't get viruses" argument for systems that attackers just don't care about. seL4 is the only one of those kernels you mentioned that doesn't have programming bugs, but it does basically nothing. If you actually knew about OS's, you'd know that it, along with any other L4 derivative is used for two things: embedded standalone systems, and hosting (usually Linux) VMs. Let me know when you have a web browser running on seL4, and then it would be relevant to this discussion.

>>57234848
>You do realize that the default scheduler isn't meant for supercomputing tasks
Linux is never production ready. Switch to Solaris.

>>57234797
>because the default scheduler is dogshit
supercomputers are batch processing systems. you submit your job and it returns at some point. they are not interactive, so they do not care about latency, but they have a lot of users, so they do care about throughput. all schedulers have to trade off between latency and throughput. batch processing systems are highly specialized, so they require highly specialized schedulers. it would be insane to ship a general-purpose kernel with a batch scheduler

>>57234837
>He was talking about the scheduler in case you didn't notice.
The scheduler itself hasn't been directly mentioned in any of the posts in this chain and the first message was about the general code quality of the kernel as a whole, not the scheduler.

>Best memory management around.
Great, but a kernel does a whole more than just memory management.

>Architecturally one of the best out there.
Pretty hard to believe considering how badly kernel extensions have been architected and documented.

>>57234848
The default Linux job scheduler sucks for more than just high-performance computing though.

>>57234608
Minix3 is bretty awesome, isn't it?

>>57234590
Freedom is great, but that's not the point. The point is: that fragmentation evidences the Linux community's inability to supply general solutions. Linux is only capable to deal with very specific problems, and only after much tailoring.

>>57234890
>b-but Linux can run a web browser
Fuck, and you guys call Windows users dumb desktop mouse-draggers...

>>57234905
See >>57234870

>>57234516
The kernel is literally the part of the OS designed to facilitate communication between the hardware and software, it makes perfect sense to load the drivers into the kernel.

>>57234837
>He was talking about the scheduler in case you didn't notice.
it's a little weird that schedulers took over this conversation because they're frankly not very important. kludge heuristic schedulers are near-optimal unless your usecase is extremal. no one has ever bothered with e.g. sophisticated machine learning in schedulers because there's no way in hell to make the optimization time pay out given how close to optimal the kludge heuristics are

>>57234938
It does not, drivers are usually very buggy. You don't want that kind of shit running in ring 0. It needs to be able to fail gracefully and recover IN USER SPACE.

>>57234921
>Great, but a kernel does a whole more than just memory management.
Sure, but that's one thing that can only be done in the kernel.

>>57234870
You do realize that a supercomputer scheduler isn't really suited for applications outside of supercomputing? Supercomputing really isn't anywhere near the biggest use of Linux so shipping with a scheduler meant for supercomputing by default would be pants-on-head retarded.

>>57234901
Switch to a dead OS? Are you serious or just trolling because I have a feeling you're doing the latter. If you're going to run Linux on a supercomputer you're not going to install bog standard Ubuntu on it, you're ether going to install a distro built for supercomputing or set up your own build specifically for your needs.

>>57234990

>You do realize that a supercomputer scheduler isn't really suited for applications outside of supercomputing? Supercomputing really isn't anywhere near the biggest use of Linux so shipping with a scheduler meant for supercomputing by default would be pants-on-head retarded.

You're outdoing yourself, Captain Obvious.

>>57234990

Solaris is far from dead, son.

>>57234938
>>57234961
from an architectural perspective the drivers should be in the kernel. from a trusted computing perspective the drivers should be outside the kernel. the question is whether you care more about architecture or trust. some people view the kernel solely as a trusted computing base, but the most important thing the kernel does is really multiplexing, and multiplexing is greatly complicated if you need to selectively expose hardware interfaces in non-privileged modes

File: illumos.png (39KB, 285x582px) Image search: [Google]

39KB, 285x582px

>>57234990
>working state: current

>>57234990
>distro built for supercomputing
Haha no, you're installing UNICOS in that shit.

>>57234987
Still doesn't mean that the whole kernel is made to anything close to the same standard.

>>57234932
You do realize that you don't actually have to use the default scheduler if you don't want to? Something that's supposed to serve as many applications as the Linux kernel does is obviously not going to be as good as something written with one very specific use case in mind.

File: 8B3S4Ef.jpg (61KB, 640x640px) Image search: [Google]

61KB, 640x640px

>>57235043
>guys my kernel is great as long as you replace the shitty parts, ok?

>>57235037
>their code is documented
:^)
(open)Solaris has the momentum but failed.

File: 1475855141214.jpg (78KB, 340x314px) Image search: [Google]

78KB, 340x314px

>>57234181
>I tried fixing it 11 years ago but couldn't

>>57235068
>wants to sort a petabyte
>"why is std::sort shit"

>>57234404

Whoa, that sounds like a monopoly. How is that allowed?

>>57235104
>a free monopoly

>>57235096

pibibyte*

>>57235017
>>57235037
>Solaris is far from dead, son.
It may not be dead quite in the same sense as Plan9, but in terms of actual users and the amount of active developers it's that not far off from where MorphOS is.

>>57235039
See >>57234384

File: 454d09bafcea790f647abe60877124bf.png (206KB, 400x400px) Image search: [Google]

206KB, 400x400px

>>57235121
>he thinks the operating systems we run on supercomputers are free

>>57235068
>My kernel is great because one part of it is great and that other kernel is shit because one (easily replaceable) part is shit
Sure thing famfam...

>>57234932
>The default Linux job scheduler sucks

Well the disk io is shit for desktop, but aside that it's good. You can choose 3 or more patch for it.

>that fragmentation evidences the Linux community's inability to supply general solutions

??? BSD community or Tanenbaum provided any (((general solution)))?

File: this_triggers_the_freetard.png (11KB, 750x530px) Image search: [Google]

11KB, 750x530px

>>57234384
>>57235135

Well, Android is soon replacing Linux with Fuchsia.

Servers are far from Linux dominated. Your picture only counts webservers, which is but a tiny fraction of all servers. In the enterprise market, it's pretty much 100% Windows.

Supercomputers are just number crunchers. They choose whichever kernel is cheapest (which is, you guessed right, Linux) and just attach their custom high-performance modules into it. The default Linux kernel doesn't come even close to high performance.

Mainframes = custom OSes

Gaming consoles... come on, seriously?!?

And embedded electronics are facing the consequences for just going for whatever's cheapest: we now have massive botnets bringing the whole Internet down. Linux won't remain a serious choice in this area for too long.

>>57235071
>>57235135
I'm guessing you don't visit the illumos-gate github.

>and the amount of active developers
You don't need a lot of developers you need good devs.
meanwhile illumos has
True containers with zones
advanced instrumentation with dtrace and a kernel that supports it.
Advanced Filesystems with ZFS
and still Linux is playing catch-up to 2005 era Solaris

>>57235215
>Fuchsia
>implying
We barely know about it yet. Also:
>multi-billion OS
>replacing amateur OS
Come on, seriously?

>>57235270
>multi-billion corporation OS
fixing

>>57235270

Linux was just a temporary solution in order to achieve time to market. You have to be a literal moron not to see that.

>>57234139
The users care to much about nostalgia or some shit like that, keeping it stuck in the 80s.

File: 2016-10-24-075903_1920x1080_scrot.png (832KB, 915x947px) Image search: [Google]

832KB, 915x947px

>>57234139

>>57235248
Visited. I wasn't really got those orgasm what >>57234613 got.
>>57235270
Took the b8.
>>57235287
MS and IOS would bath in champagne if they try to fuckup android with their meme os.

>>57235270
>barely know about it
Wtf, it's right here anon: https://fuchsia.googlesource.com/

>>57235349
>meme OS
Keep telling yourself that.

>>57235408
Keep shitposting about it.

>>57235089
lmao pathetic

>>57234753
>g-guys, your patch broke muh userland! just because my code reuses a few dangling pointers and fails at bounds checks a couple times it won't run anymore!
Wow, you are literally blaming others for your own mistakes, how does it feel being a literal nigger?

>>57235491

You sound frustrated.

>>57235215
>On-Premises server OS
Nice try at picking selectively as on-premises servers is mostly things like the servers required to support a corporate windows environment (which is obviously going to be big due to Windows' stranglehold on the corporate desktop environment).

>Well, Android is soon replacing Linux with Fuchsia.
According to rumors, but there's really nothing actually substantial behind it. Rumors like that come and go all the time so you shouldn't really pay that much attention to them.

>Mainframes = custom OSes
Sure thing, but mainframe market is a pretty small and specialized market anyway. It's even considerably smaller than the console market.

>And embedded electronics are facing the consequences for just going for whatever's cheapest: we now have massive botnets bringing the whole Internet down. Linux won't remain a serious choice in this area for too long.
If you've had a look at the malware that has been creating botnets on internet connected embedded systems you'd know that they're not relying on security flaws in the underlying OS, but the shoddy work of the hardware vendors with factory default root accounts you can't remove. Because of this BSD-based systems are suffering from exactly the same problem as Linux-based ones.

>>57235514
Why would google replace linux with fuwhatever and what would be the advantage? Better drivers? Faster bionic? Barely can find any point. Maybe licensing but google now uses openjdk anyway. They should support the development of toybox not supporting memers like you to spread inane comments.

>>57235575
something something botnet

>>57235599
Solving with other botnet.

too many autistic people trying to do too many different things. for instance the polar opposite personalities of linus and stallman.

>>57235495
>So what if these things break the userland... It's not like anyone is every going to be doing much in the userland to begin with on my OS.

Also
>Calling someone a nigger as part of attempting and un-ironic insult
/pol/ pls go...

>>57235248
None of that really matters if nobody except the devs themselves use it and there's really next to no third party software for it.

File: 6540.png (42KB, 507x316px) Image search: [Google]

42KB, 507x316px

>>57235564
>Nice try at picking selectively
Thanks, I also made it very clear. I said "enterprise market" very clearly. I didn't attempt to be intellectually dishonest and count on the reader's lack of attention like you did here >>57234384 selectively picking web servers and trying to make that pass as the entire server market.

>there's really nothing actually substantial behind it
Why is Google developing Fuchsia then?

>but mainframe market is a pretty small and specialized market anyway
If we only count the big market, Microsoft blows every other competitor out of the goddamn fucking water hands down. Pic related.

>hoddy work of the hardware vendors
Most Windows problems are just shoddy work of hardware vendors who make half-baked drivers, but you seem to blame Windows for that nonetheless.

>>57235575
Ummm... no 9-year old critical vulnerabilities?

Why would they even be developing Fuchsia otherwise? Maybe they're just playing around, right?

>>57234195
>>57234299

if your idea of good UX is what Microsoft and Apple do than I'm very glad that Linux refuses to follow suit.

>>57234139
For me personally? sudo pacman -Syu went wrong, twice

>>57235632

>we gotta keep backwards compatibility at all costs, even if it means not making any improvements
>making application software developers fix their sloppy coding is bad
This is how you sound.

>le /pol/ bogeyman
XOR yourself.

>>57235632
>None of that really matters if nobody except the devs themselves use it and there's really next to no third party software for it.

>What is samsung? Which just bought joyent (the main contributer to illumos)
>third party support for containers
>third party support for a filesystem

File: 1476162286762.gif (3MB, 400x327px) Image search: [Google]

3MB, 400x327px

ITT: lincücks get BTFO

>>57235679
well at the end of the day you don't control what users use. Users do. And they want something that runs their programs and doesn't break. So if you try to maintain your purity and break the application because the application did the wrong thing, all the user sees is that his program used to work, and you released a new version, and now it doesn't. He blames you, not the people who wrote the application. His first solution is to not upgrade to your new version, because remember, he doesn't give a shit about your idea of how code should be organized, he only cares about running his programs, and with an absolute minimum (ideally zero) hassle. In Windows land this is why XP still refuses to die.

This is why disregarding backward compatibility is shooting yourself in the foot.

>>57235663
They can have whatever project to have. With this analogy no one should use windows either because the legacy code gave older security holes than 9 year.

>>57235642
You do realize that Microsoft has pretty much a lock on the enterprise market with the way they go out of their way to ensure that if you buy one of their products, you have to buy a host of their other other products. There's really no way of having Windows desktops and not having a Windows server backend serving them or to get a Windows backend to serve anything but Windows desktops?

>Why is Google developing Fuchsia then?
Companies develop systems that never make it to market at all or only as re-used tech in other products all the time. Just look at all the other OSs Bell labs put together over the years or the MinWin kernel that was just used to develop tech that Microsoft would be used for Vista onward.

>Most Windows problems are just shoddy work of hardware vendors who make half-baked drivers, but you seem to blame Windows for that nonetheless.
Nice attempt at a strawman, but I specifically talked about the more iffy parts of the Windows kernel.

>If we only count the big market, Microsoft blows every other competitor out of the goddamn fucking water hands down. Pic related.
It's easy to be a big company when you have a 100% lock on the OS you're working on you know

>>57235679
>Try to insult someone by calling them a nigger
>Not seeing why someone would tell you to fuck off over this

Also
>Hurr durr... I'm going to break shit all the time just so that I can add all of these features that there's isn't even real need for

>>57235642
>selectively picking web servers and trying to make that pass as the entire server market
it's labeled as web servers in the figure that i copied from wikipedia. i didn't offer any commentary about it. you're inventing the deception so you'll have an argument to win

>Why is Google developing Fuchsia then?
google is a research company and maintains an overwhelming number of research projects, most of which never become products

>If we only count the big market, Microsoft blows every other competitor out of the goddamn fucking water hands down. Pic related.
you know that revenue is not proportional to share for at least one extremely obvious reason

File: tumblr_inline_nt0vfexkpK1spsojg_540.jpg (67KB, 540x534px) Image search: [Google]

67KB, 540x534px

>>57235297
lol'd

>>57235700
>Thinking that the only reason someone would buy a tech company is to get an OS they contribute to
Still doesn't really do anything to dispel the fact that nobody outside of the devs really uses it. Maybe you should come back IF Samsung actually does something with it...

>third party support for containers
>third party support for a filesystem
Yes, it's a common file system available for use elsewhere as well so there's really nothing impressive about an open source piece of software being developed for more than one OS.

>>57234139
Richard Stallman

File: lolcuck.png (287KB, 440x440px) Image search: [Google]

287KB, 440x440px

>>57235270
>Fuchsia
>amateur

>>57237246
Linux is amateur.

>>57234139
GNU/