video This guy makes some good points about Linux

[Growled Member]

I love Linux and I want it to work out and do good but it always falls so short. For me, Android is what Linux should have been.

[Glassed Silver]

[...]

KDE is a bit snappier and more familiar looking as a Windows guy but comes off a bit unpolished. Fedora's KDE spin is ****ing me off with constant SELinux warnings about this and that, that are not seen in Gnome.

I want to give KDE another go in OpenSUSE, but the install always takes a dump because it can't delete my fedora partition for some reason. Neither can Gparted.

SELinux should really be disabled.

It's only painful and doesn't make much sense in most environments.

Really, disabling it is like hotfixing your system.

Refer to this:

Glassed Silver:mac

[ichi]

And that's precisely why it's not going anywhere at the moment. I love choice, it's great. But it's just as possible to baffle a user with excess choice as it is to not give them enough. And that's my main problem with a lot of FOSS communities there's overemphasis on providing a million and one competing solutions for every idea, which ends up with a lot of spotty quality with the average user not able to tell what option is good and what one is crap. It's my opinion that the Linux communities need to pool their resources and work towards some form of unification. I'm not saying there should be no choice of course but they need to start focusing on quality and usability more.

Where else is Linux supposed to go?

Cooperating on a single project might sound great and surely you might get something better faster than people spread on different projects, but there are some reasons why that won't happen and also why it shouldn't happen:

-People work on different projects because that's what they want to do. You could get some Linux companies to join forces, but a hell of a lot of the work spent on Linux projects comes from volunteers, and you sure won't be getting them to stop working on whatever they want.

-If people had all focused on one single project we would be missing parallel projects that turned out to be awesome. If everyone was working on, say, GTK, we wouldn't have the cool crossplatform framework that is Qt (and no one can guarantee that GTK would actually be betten than it currently is).

The way I see it Linux as a whole is sort of an evolutionary playground that advances slowly because of the diversified and duplicated work. What's worth sticks and what's not eventually fades.

The good thing is that surprisingly it does actually move forward, and so far they have achieved a nice desktop (despite it's many flaws) and an awesome server OS. It's certainly not the best way to get a product out of the door, if Linux was an actual single company they would have gone bankrupt long ago, but since it isn't... well, that's the way it rolls.

If you want an unified desktop your best bet is pushing the vision of one distro (Canonical, RedHat, whatever...) and forgetting about the rest, because I don't see all the Linux developers working together on a single project ever.

Heck, you can't even get users to agree on one single distro yet you expect those who are actually spending time coding to do so?

[Noir Angel]

And that attitude is precisely why Linux is always going to languish as a niche product with little real world usage. Too many people obsessed with competition not prepared to work together for a common goal.

[ViperAFK]

Too many sound programs and video players is a lack of cooperation.

Really, all you need is vlc anyway, so I don't see the point in complaining about too many video players :) Also many of the other players are just basically gstreamer or mplayer frontends, meaning there's many video player front ends, but really when it comes to video players you have mplayer, vlc, and gstreamer based.

[Motoko.]

And that attitude is precisely why Linux is always going to languish as a niche product with little real world usage. Too many people obsessed with competition not prepared to work together for a common goal.

Only canonical seems to have a vision being fulfilled with Ubuntu in the consumer space and they're doing great with that. With every release they seem to be doing something right but it'll take a more larger scale of developers to get into a direction they can really branch out from, it's as you say.

[migo]

If the Linux world were organised in some fashion like BSD, the situation would be a lot better.

You have pretty much 3 options.

Want to run BSD on your toaster/graphing calculator/thermostat? Use NetBSD.

Want a completely secure and stable BSD for a server/workstation/corporate environment? Use OpenBSD.

Want BSD on your desktop/laptop and you care more about new features and less about security? Use FreeBSD.

Then, from FreeBSD there's DesktopBSD and PC-BSD, but both of those are 100% FreeBSD, just prepackaged in a more consumer friendly fashion.

There's no glut of extra, unnecessary BSD distributions.

It would be nice if Linux could be something even close to that.

As far as openSuSE and Fedora go, they're both testing grounds for their commercial releases. They're making money, so obviously it makes sense for them to provide their own distributions. Any other company that's commercially successful with a Linux distribution, makes sense to have their own. But when you get to hobby distros, or volunteers working on a project, what's the point of having 500 different distros? The less people you have working on a given distro, the less hardware variants it gets tested on, and the lest people it's potentially viable for. I have no shortage of Linux distros that I've tried that won't even boot to the Live destop on my netbook, even though as an Atom based Netbook pretty much all the components are constant.

It's made it impossible for me to recommend any Linux distro to anyone else, because even if I get it working on my own system, I have no idea if it will work on theirs. That's a bad situation for Linux to be in, and it'd be much, much better if there were a few distros that were tested on many hardware configurations.

[Glassed Silver]

I agree, BSD is awesome.

I'll likely use it or Solaris for a server in near to mid future.

Currently tinkering with BSD in VMs and adore it.

I'm especially intruged by ZFS 28.

Might also play around with Oracle Linux, due to btrfs testing - NOT going to deploy it in a productive environment in the foreseeable future, as it's too young and immature for me, but I'm looking forward to it! :)

Glassed Silver:ios

[ichi]

And that attitude is precisely why Linux is always going to languish as a niche product with little real world usage. Too many people obsessed with competition not prepared to work together for a common goal.

Maybe, maybe not. The point is that if Linux desktop ever goes anywhere near mainstream that'll be because of the push of a single company with one single distro, not because everyone suddenly decide to work together.

Just look at the smartphone market: it wasn't Linux as a whole who took it by storm, it was Google's Android. Does anyone really care about any of that guy's complains when using an Android phone?

And besides the desktop, Linux is certainly not niche anyway.

[migo]

Actually, Android has some serious fragmentation complaints. Sense on an HTC, TouchWiz on a Samsung, Blur on a Motorola, Flyme on a Meizu, and so on.

The big difference for Android success is Android comes bundled with the hardware people are going to buy anyway, and for a while it was the only viable alternative to the iPhone. The problem Linux runs into is that for the start of the Netbook craze, the community didn't set aside the differences and pool their resources to make a UI and overall experience that worked very well on a netbook, and instead ceded that market to Windows.

It's only now that Netbook sales have almost dried up that we're finally seeing some distros that work well.

Linux might get another chance with gaming - no way it'll be up to snuff for tablets even though Unity and Gnome Shell are both designed with tablet use in mind - and that's only if Microsoft fumbles or Valve really pushes ahead with their Linux support. For that though, not only will there have to be a solid Linux distro that looks good and performs well on all hardware, they're going to have to get the graphics driver issue sorted out, and I'm not even convinced they'll be able to do that with Intel HD 4000 (or 5000/6000 when it comes out) graphics even though Intel provides the source code to the drivers.

[.Neo]

The big difference for Android success is Android comes bundled with the hardware people are going to buy anyway, and for a while it was the only viable alternative to the iPhone. The problem Linux runs into is that for the start of the Netbook craze, the community didn't set aside the differences and pool their resources to make a UI and overall experience that worked very well on a netbook, and instead ceded that market to Windows.

Ubuntu was pretty much there really. I'd say the real problem lies with commercial software support and the already available software just not being good enough feature-wise.

[Anibal P]

Actually, Android has some serious fragmentation complaints. Sense on an HTC, TouchWiz on a Samsung, Blur on a Motorola, Flyme on a Meizu, and so on.

The big difference for Android success is Android comes bundled with the hardware people are going to buy anyway, and for a while it was the only viable alternative to the iPhone. The problem Linux runs into is that for the start of the Netbook craze, the community didn't set aside the differences and pool their resources to make a UI and overall experience that worked very well on a netbook, and instead ceded that market to Windows.

It's only now that Netbook sales have almost dried up that we're finally seeing some distros that work well.

Linux might get another chance with gaming - no way it'll be up to snuff for tablets even though Unity and Gnome Shell are both designed with tablet use in mind - and that's only if Microsoft fumbles or Valve really pushes ahead with their Linux support. For that though, not only will there have to be a solid Linux distro that looks good and performs well on all hardware, they're going to have to get the graphics driver issue sorted out, and I'm not even convinced they'll be able to do that with Intel HD 4000 (or 5000/6000 when it comes out) graphics even though Intel provides the source code to the drivers.

The fragmentation myth, how cute, you sound like an Apple drone, you seem to HATE choice if it's not the one choice Apple allows, guess what, most normal people are not like that at all, normal people like choice

[ichi]

Actually, Android has some serious fragmentation complaints. Sense on an HTC, TouchWiz on a Samsung, Blur on a Motorola, Flyme on a Meizu, and so on.

Android has problems with OEMs not updating their devices to the latest ROM, but the point is you don't see there any of the complains from that video, eg. the audio stack. It doesn't matter if there's OSS, Alsa and Pulseaudio out there, Android's sound just works.

Same should happen on the desktop: Canonical (or anyone else) should just pick their choices and polish them until everything works fine. If someone else wants to keep working on eg. a different sound stack then so be it, Ubuntu wouldn't be affected by that in any way.

The big difference for Android success is Android comes bundled with the hardware people are going to buy anyway, and for a while it was the only viable alternative to the iPhone. The problem Linux runs into is that for the start of the Netbook craze, the community didn't set aside the differences and pool their resources to make a UI and overall experience that worked very well on a netbook, and instead ceded that market to Windows.

It's only now that Netbook sales have almost dried up that we're finally seeing some distros that work well.

Which was my point: same as Google pushed Android on smartphones you won't see Linux succeeding on the desktop in any other way that coming from some company pushing it. If you see OEMs shipping laptops with Ubuntu at all (like those Asus just launched) that's precisely because of Canonical.

Regarding the netbooks, well, OEMs going with weird obscure distros like Linpus or Xandros without even verifying that the software did actually support the hardware it was bundled with certainly didn't help, and that had nothing to do with comunity's work.

I didn't get to use Xandros (although from what I heard on some netbooks the webcam didn't even work OOTB) but on Linpus the software selection from the repos sucked, and even getting some video codecs to work was a PITA. I just replaced Linpus with Ubuntu on my Acer One and everything worked perfectly fine, go figure.

[.Neo]

The fragmentation myth, how cute, you sound like an Apple drone, you seem to HATE choice if it's not the one choice Apple allows, guess what, most normal people are not like that at all, normal people like choice

It's becoming really tiresome on this forum that people who are critical about whatever product are immediately being called a fanboy from the other camp. The true myth is that the average consumer actually cares about having tons of choices and care about something being open source or not. They just want something that works and couldn't care less about the rest as long as it continues to work properly.

[Deihmos]

Everytime I tried linux (ubuntu) I found the audio to sound terrible so I always get rid of it. I also encoutenered tons of bugs and I also find it to be really slow compared to WIndows 7.

[Jason S. Global Moderator]

[migo]

Android has problems with OEMs not updating their devices to the latest ROM, but the point is you don't see there any of the complains from that video, eg. the audio stack. It doesn't matter if there's OSS, Alsa and Pulseaudio out there, Android's sound just works.

That's true, there's less fragmentation than in Linux, but it's still there. If every phone just ran stock ICS (they're STILL doing overlays, sheesh!) things would be a lot better.

Same should happen on the desktop: Canonical (or anyone else) should just pick their choices and polish them until everything works fine. If someone else wants to keep working on eg. a different sound stack then so be it, Ubuntu wouldn't be affected by that in any way.

They should do the same with DEs. Would have been better for them to stick with Gnome than to have Kubuntu, Xubuntu and Lubuntu. That's unnecessary extra support work for them.

Which was my point: same as Google pushed Android on smartphones you won't see Linux succeeding on the desktop in any other way that coming from some company pushing it. If you see OEMs shipping laptops with Ubuntu at all (like those Asus just launched) that's precisely because of Canonical.

So Canonical went out to Asus and asked them to build hardware that was known to work flawlessly with Ubuntu or?

Regarding the netbooks, well, OEMs going with weird obscure distros like Linpus or Xandros without even verifying that the software did actually support the hardware it was bundled with certainly didn't help, and that had nothing to do with comunity's work.

Even Ubuntu at the time had problems though, I remember deliberating between Ubuntu, Knoppix and PCLinuxOS because each one had its own flaws that were quite annoying that the other distros had solved perfectly. It's not like there was a good Linux distro back then, at least not unless you were lucky with hardware combinations, and that was an issue with the community's work.

I didn't get to use Xandros (although from what I heard on some netbooks the webcam didn't even work OOTB) but on Linpus the software selection from the repos sucked, and even getting some video codecs to work was a PITA. I just replaced Linpus with Ubuntu on my Acer One and everything worked perfectly fine, go figure.

That also brings up the question of why people feel they need to make ****ty distros that are forks of a much better supported one.

[migo]

Ubuntu was pretty much there really. I'd say the real problem lies with commercial software support and the already available software just not being good enough feature-wise.

Commercial software support is something of a community problem. Linux users didn't really want anything to do with WordPerfect, despite being a native port (for 8) because it wasn't open source. Stallman isn't really helping there either.

[simplezz]

I see a lot of people, including the guy in the video, complaining about too many video apps, sound architectures, distros, DE's etc... Well that's what FOSS and GNU/Linux are all about - Choice

Some like VLC, I like Mplayer, and the list goes on. There's something for everybody, that's why GNU/Linux is so appealing. Why should everybody conform to the same software or way of doing things? If you want that, stick to Windows or OS X, don't foist it upon us Linux users.

[migo]

If they were all self contained apps it wouldn't be so much of a problem, but Linux has layers upon layers of dependencies.

[ichi]

That's true, there's less fragmentation than in Linux, but it's still there. If every phone just ran stock ICS (they're STILL doing overlays, sheesh!) things would be a lot better.

That doesn't mean OEMs would be releasing updates faster. The actual problem with fragmentation is device drivers, which is what Google is trying to solve giving OEMs earlier access to the new platform with the PDK.

They should do the same with DEs. Would have been better for them to stick with Gnome than to have Kubuntu, Xubuntu and Lubuntu. That's unnecessary extra support work for them.

Canonical? Probably, but then I don't see them spending much time on those other distros. All their development is focused on Ubuntu.

So Canonical went out to Asus and asked them to build hardware that was known to work flawlessly with Ubuntu or?

I don't know how Canonical approached Asus, but they can provide a comprehensive list of certified hardware, they can provide support and they can help with any questions the OEM might have about the deployment.

Canonical has been in talks with OEMs for quite some time now, and I'd think that has something to do with them releasing laptops with Ubuntu and not Fedora or Suse.

Even Ubuntu at the time had problems though, I remember deliberating between Ubuntu, Knoppix and PCLinuxOS because each one had its own flaws that were quite annoying that the other distros had solved perfectly. It's not like there was a good Linux distro back then, at least not unless you were lucky with hardware combinations, and that was an issue with the community's work.

As a OEM you shouldn't be releasing a non working product and expecting anything but failure. It's up to them to assemble hardware and software together and make sure everything works.

If some of your product's hardware isn't working either get in talks with the developers to get it fixed or just don't release the product at all.

The community can't test a distro in every possible hardware configuration. You can blame devs if you download a distro and it doesn't work for you, but when it comes to OEMs it's their responsability to work with the devs if they really intend to get the product out.

I guess I was lucky with the Acer One since everything worked when I installed Ubuntu.

That also brings up the question of why people feel they need to make ****ty distros that are forks of a much better supported one.

Because they can, but that doesn't mean that anyone has to use those or even pay any attention.

If they were all self contained apps it wouldn't be so much of a problem, but Linux has layers upon layers of dependencies.

Anyone can release statically compiled packages, the problem is you would have to keep baking releases for every update in each of the statically linked libraries or else your product would have flaws that have been already fixed.

Anyway since dependencies are sorted out automatically by the package manager I don't see how that's a problem.

[migo]

Sometimes one program has a dependency for an earlier version of a particular library, while another program has a dependency for a later one. Unless you're using something like Gobo Linux (which almost nobody is as it's defunct), you're screwed for one of the programs. It's like requiring 4 different versions of the .net Framework to be installed, except on Linux you'd only get to pick one.

[ichi]

Sometimes one program has a dependency for an earlier version of a particular library, while another program has a dependency for a later one. Unless you're using something like Gobo Linux (which almost nobody is as it's defunct), you're screwed for one of the programs. It's like requiring 4 different versions of the .net Framework to be installed, except on Linux you'd only get to pick one.

You won't get into that problem using software from the repositories, that's the whole reason why you don't get bleeding edge versions until the next distro release.

For stuff like libc where you are more likely to run into that problem there are libcompat libraries that provide compatibility with previous libc libraries.

If for some strange reason you really really need wildly different lib versions for a specific application you can always chroot it, although that'd be a bit overkill :/ I haven't found such problem so far, though. Only with some HP products and only about libc, which is solved with the libcompat package.

[n_K]

The guy having problems removing his fedora partition - you need to boot from a livecd/usb and use gparted, right click on the partition and click 'unmount' if you can't delete it :).

SELinux - for most home users it isn't needed at all, if you're doing server work and whatnot though it's a VERY good idea to enable it (microsoft uses it on their *nix skype servers).

BTRFS is in theory stable now, it has an fsck tool called btrfsfsck NOTbtrfs.fsck so I symlinked it for my server install, haven't really tested it's speed in comparison to ext4 but it's reliable! Server's lost power on a few occasions and the VMs have just been shut off without powering down and they've always recovered fine.

'If for some strange reason you really really need wildly different lib versions for a specific application you can always chroot it, although that'd be a bit overkill'

That's useful for skype to prevent it sending data about your PC ;)

[allknow]

The fragmentation myth, how cute, you sound like an Apple drone, you seem to HATE choice if it's not the one choice Apple allows, guess what, most normal people are not like that at all, normal people like choice

? ohh.. so this is only choice and not fragmentation? add in on top of all the fragmentation all the different screen sizes and aspect ratios of each Droid maker and you got yourself one heck of a huge mess.

reminds me of how a girl says "I'm not a slut, I just like variety"