What Are The Biggest Problems With Linux?

[earthwirehead]

As far as *Desktop* Linux is concerned...

... the biggest problem is a certain anti-competitive corporate monopoly making it difficult to impossible for hardware vendors to support any operating system that doesn't have a trademarked name starting w/ "W".

The next biggest problem after that has everything to do with applications. I love LibreOffice Calc, but there is simply no way I could use it at work, even if they would let me. For all of of Excel 2010's horribly botched interface, it still has power and features it's OSS equivalents lack. Similar arguments have been made comparing Adobe's products w/tools like GIMP and InkScape (which seem fine to me, but I'm not a professional designer). When people are looking for reasons to not like Linux in the first place, losing functionality in tools they depend on is not a winning argument.

Which is a pity, because Desktop Linux has the best opportunity for increasing market share it's every likely to get. Windows 8 is the most colossally hubristic mistake Microsoft has made since Bob. The learning curve of adopting the Metro interface versus the learning curve of switching to Ubuntu or Mint looks like a coin toss to me.

The apps aren't going to get better until more people can make a living writing them. But how do you pay people to write code for less than 2% of the market--particularly when the code, strictly speaking, is given away? It's a vicious cycle--I wish I knew a way to break it.

[elanthis]

This: Why Linux is not (yet) Ready for the Desktop (a.k.a. Linux problems), 2012 edition

This is amazing. Thank you for this link, it'll save me a lot of time explaining to people why we don't waste time porting games to Linux.

There's one problem no amount of money can solve: for most Open Source developers Linux is a playground, a thing they don't care beyond their aspirations, thus we have constantly broken features and API breakage every odd moon cycle.

There's that, plus the complete lack of real QA. The Microsoft Azure QA team alone is larger than the entire developer and tester pool of X.org, Mesa, GTK, GNOME, and the DRI/DRM bits of the kernel. You would crap yourself if you had any idea of the size of the team responsible for just DirectX.

FOSS, for all the talk about it being open to anyone and having "millions of eyeballs," simply does not have a very large developer pool, and has an even smaller "support staff" pool. Real professionals with actual skill/talent are doing their work for pay. The good ones are doing it for a lot of pay. They do not have free time after working 40-60 hours/week to hack on a hobbyist OS or software, and even if they did they probably want to spend their non-working hours doing something other than slaving away on more software. Take that fact along with the fact that the largest FOSS companies have teeny tiny profits compared to even the run of the mill proprietary software companies (Red Hat's recent $1b _gross revenue_ is about the same as the _net profit_ of some of the smaller well-known software companies), and you have the cause for the practically barren developer pool on the important FOSS projects, and the reason why the folks working on those projects keep complaining about being so seriously under-manned.

This doesn't help the student that this article was for, but the sad fact is that there is no real help. One student working to better Linux is not what Linux needs. An army of professionals being backed by a multi-billion-dollar software empire is what Linux needs. Only, no multi-billion-dollar Linux empire is ever going to exist (ignoring future inflation) because there's no way to make that kind of money on software that nobody outside of a few large server farm operators have any reasons to pay for.

[devius]

CAD is more complicated due to political matters rather than technical.

No it isn't. There are already 2 native linux programs that are 100% compatible with the latest DWG2010 specification and that work 99% like AutoCAD: BricsCAD and DraftSight. If you know the AutoCAD commands you'll have no problem using any of these programs, although only BricsCAD has basic support for 3D modeling. BricsCAD is closed source and commercial, DraftSight is free but has less features and is less polished.

[e8hffff]

This is amazing. Thank you for this link, it'll save me a lot of time explaining to people why we don't waste time porting games to Linux.

You could easily write a list of horror point in Windows too. Simply the user experience in Linux is far from horror. It usually is a story of empowerment, freedom, and functionality.

There's that, plus the complete lack of real QA. The Microsoft Azure QA team alone is larger than the entire developer and tester pool of X.org, Mesa, GTK, GNOME, and the DRI/DRM bits of the kernel. You would crap yourself if you had any idea of the size of the team responsible for just DirectX.

So what. Azure is a lame project anyway. Direct X is justified and I give it credit, but it's come from Microsoft's deviant manipulation of hardware makers. Just because Direct X is nice doesn't make me want to use Windows. Far from it.

Linux is the future. I can't say that about Microsoft. Microsoft is as good as its cunning works and as time goes I see more leaving Windows.

This doesn't help the student that this article was for, but the sad fact is that there is no real help. One student working to better Linux is not what Linux needs. An army of professionals being backed by a multi-billion-dollar software empire is what Linux needs. Only, no multi-billion-dollar Linux empire is ever going to exist (ignoring future inflation) because there's no way to make that kind of money on software that nobody outside of a few large server farm operators have any reasons to pay for.

That could be one avenue. I agree.

The simple fact is we are going into a world of austerity. Many businesses will fail. Many may start. Many people have habits that are linked to Microsoft software, a culture to stay with MS, but this is crumbling. More and more people don't need to use Excel, nor AutoCad, nor Outlook, or Adobe. The web is also changing. Open source and free projects are gaining ground. It's also upto user to push businesses to accept more then the old titles. Like if a Government department wants MS file types then demand they support open source.

[pat99]

There is no big single problem...

It's more like a thousand papercuts... At first, you're like: "hey! this is quite fantastic!". Then you can't play that game because your video driver stopped working with the last X server update. Then the HDMI plug on your specific laptop is not working that month. Then sound latency is horrible for a couple of weeks because of a regression in pulse audio and or alsa and or phonon and or gstreamer. Then your printer does not have an ink monitoring utility. Then you have to fiddle with every single f****ng KDE application because their default window sizes are all 109x208 and you start to wonder if there are any QA going on on these distros/desktop projects. Then that big DE throws the baby with the bathwater and tells you to start using that awesome tablet/phone interface. Then that new feature that will make you PC so much better is in a git repository somewhere on the net and will be integrated in 2014 in the next major release of Archbuntubayon 22.4 but by then the developpers will have changed their mind and that feature will be replaced by that newfangled wizzbang dbus interface that works once every moon cycle. Then that software you have used for 10 years will be forked because a developper nobody cares about somewhere in Braziluzuela has developped a shitty attitude and won't talk to other people. Then you have to stare at a cashew at the corner of your screen that you have no idea what its use can be. Then there are no minimize buttons suddenly on your window decoration. Then they rename udev to systemd-tools and you have to fight for hours figuring out why pacman won't upgrade your OS. Then the KDE guys do a system tray protocol and the Gnome guys think :"What the hell! We'll do our own instead of reusing their stuff!". Then.... Then... Then...

Then you reinstall a pirated copy of Windows...

I started using Linux a long time ago with a Slackware '96 cdrom i got for 10 bucks at a computer expo while i still was in school. I've used KDE since the 1.1 series. All the way to 3.5. Then KDE went into a 5 years alpha phase and got decent again by the 4.6 series. The other options were XFCE (like going back to the fvwm days) or Gnome (like having a cheap and uncool Mac). It was 2007 and i was so tired of all this shit that i just said goodbye and never looked back. Sorry.

I tried many distros recently just to check if maybe thing were getting better... then all the things i described in the first paragraph happened to me in a 2 weeks period after trying Kubuntu, Ubuntu, Sabayon, Arch, Chakra and Fedora. Then it hit me right in the face: Linux on the desktop is dead. Deal with it. Until people start cooperating it's never gonna get better. You have to work so hard to get anything done on a Linux desktop... It's just not worth the effort.

They say if you scratch a cynic, you find a disapointed idealist...

[Teho]

Take that fact along with the fact that the largest FOSS companies have teeny tiny profits compared to even the run of the mill proprietary software companies (Red Hat's recent $1b _gross revenue_ is about the same as the _net profit_ of some of the smaller well-known software companies), and you have the cause for the practically barren developer pool on the important FOSS projects, and the reason why the folks working on those projects keep complaining about being so seriously under-manned.

You should note that companies like IBM, Intel and Google also support and invest heavily on Linux and they are among the largest companies on the industry. Altough they don't really focus on desktop they still indirectly work on many parts of the stack. The desktop Linux is close to feature parity with Windows on almost everything on the plaform level. If hardware manufacturers would take more active role in developing drivers for Linux we would be just about set. OpenGL as spesification is much more widely adopted by industry than Direct3D for example and I fail to see any truly blocking technical difficulties that would stop Linux from entering the market.

[soupbowl]

Agree with you pat, though everyone will scream "THEN USE A STABLE DISTRO!"
I have and had similar problems, update my ubuntu loose moose support and hdmi support.
Little stuff like that happening every few months, is to annoying.

The next thing you hear is "USE LINUX SUPPORTED HARDWARE!"
well my mouse and my wifi were supported, the wifi had great drivers, then one update
it quit working and has yet to get patched, its been a year at least and there have been
bug reports.

[c0sf]

[Native] Applications

I do not think that's a problem with linux itself, but a problem with open source applications available, specifically they features, missing or incomplete.
many open source applications tend to be very "weak" compared to their proprietary counterparts.

[sloggerKhan]

Disabled by default (man nouveau), add this option to you xorg.conf(.d):

Option "GLXVBlank" "true"

^ Basically see my opinion about sane defaults and ease of configuration for multiple hardware/software situations.

[elanthis]

You could easily write a list of horror point in Windows too.

... can you?

Simply the user experience in Linux is far from horror. It usually is a story of empowerment, freedom, and functionality.

The kind of empowerment that users actually want is the kind that doesn't require them to spend months mastering being a Linux sysadmin and instead spending their free time doing things they actually care about.

Sure, you (and me and everyone else on Phoronix) are massive dorks who actually do enjoy tinkering around with OSes. The other 99% of people think that just about anything _other_ than tinkering with OSes is fun.

So what. Azure is 7 a lame project anyway.

The "unsexiness" of that project is exactly the point I was trying to make. Cool projects in Linux should -- according to the rhetoric of FOSS -- be attracting talented users in droves, since after all the projects aren't "limited" by funding or schedules or management and anyone can contribute. Instead, Linux projects get tiny handfuls of devs at best and many of those potential developers are going to work for a company on a project like Azure because that company can actually pay them what they're worth.

Direct X is justified and I give it credit, but it's come from Microsoft's deviant manipulation of hardware makers.

Yeah... Microsoft doesn't work the way the average Linux hipster thinks it does (yes, I too once wrote things like "M$" thinking I was so clever and smarter than the "Windoze" sheeple). In any case, regular users don't make OS choices based on hate, and they sure as hell don't give a crap if a company is or is not manipulating hardware manufacturers. If things worked the way you think they do, Microsoft would have died out in 1980's.

Microsoft is as good as its cunning works and as time goes I see more leaving Windows.

The actual stats show that the only people leaving Windows are the people going to Macs. Of course, they also show that a lot of the hardcore Linux users of 1990's and early 2000's also moved to Macs. Linux was at 1% in 1999. It's at 1% now. Projecting historical evidence indicates that it'll still be at 1% in 2025.

And no, Android doesn't count. Android is Linux the same way that the Sega Saturn was Windows.

Like if a Government department wants MS file types then demand they support open source.

I actually worked for a very large government installation for years, doing in-house software development and some light sysadmin work on the Linux server farm. This was back in my Linux fanboy days. It is actually one of the larger catalysts that caused me to turn from a "Linux is the future" proponent like yourself into a "Linux is a nice Web server OS, but thank God there's someone I can give money to in exchange for a less frustrating desktop experience" believer.

Government jobs will best illustrate for you just how great Windows is for idiots who can't tell their assholes from floppy drives and how awful Linux is for "idiots" who can't figure out how to read unified diff files generated by dpkg when foobar-1.7.2b changes config file compatibility with foobar-1.7.2a.