Xfce, LXDE, & GNOME Are Running On Ubuntu XMir

[finalzone]

You're right, I forgot about Mandrake and urpmi. I do remember Mandrake being a pain because third party RPMs were usually built for Red Hat. But that wasn't the fault of urpmi.

Red Hat had up2date then, they didn't ship with yum until RHEL5. I remember having problems with that, but I think it was mainly because there was a lot of software that wasn't in the official repos, so you would add an external repo, which would eventually get out of sync with the official one and cause problems. Ubuntu handled the external software situation a bit better by incorporating way more packages into the official repository - Ubuntu had 14000 packages vs Red Hat's 4000 - and by later introducing PPAs so they could host and rebuild user-contributed packages automatically.

Could you clarify your posts? Are you talking about Ubuntu LTS or plain Ubuntu? RHEL is an entreprise distribution.
Number of packages are bad metrics because of variable like subpackages, active maintenance, and legacy.
Anyway, this is off-topic and we should get back to display manager

[brosis]

without MIR noone would care about Wayland atleast for the next century and we would be using X anyway. After Ubuntu announced MIR suddenly there is some rushing for Wayland and, of course, KDE and RedHat keeping the shitstorm against Ubuntu/Canonical going.

FUD. It was told over and over that release of Wayland have NOTHING in common with appearance of Mir. Wayland is result of long process and is unrolling completely unaffected by Mir. Because, unlike the shitstorm Mir, that jumps in front copy-pasting everything, without any real difference in goals, except producing fragmenting, tons of bugs and requirement for all desktop environments to support YET ANOTHER display manager just because its Canonical owned, Wayland was coming long independent way and is official Xorg child. Exactly for stuff like Canonical does, people actually leaved microsoft.

Because we either end up with duplicated effort and more bugs, or end criticizing each other, or end up with some software available only for specific implementations. In all cases microsoft profits, and this is exactly what bacon and his chief ex-microsoftie spencer are so happy of. So, please, just do a good thing - stfu and find a distro that does not create problems in Linux adoption, but rather helps along. RedHat are kernel maintainers, they are opensource drivers maintainers, they have huge impact on average desktop Linux, unlike Canonical. Canonical has huge impact only on Ubuntu. If this gets too far, people will start claiming "Ubuntu = is Not Linux".

For Mir they can do a piece of crap not working at all, they are right because is one of the basis of the free software... They Can fork Wayland, patch Wayland they can do another project. You can't be critic about that if they think is a better way for them and the project.

They can fork. They can patch. But they can't do another project, that is incompatible to Wayland and mis-use the userbase to split the Linux ecosystem. That is antiethical and anti-ubuntu. If Mir would be significantly different from Wayland (which its not), they would not keep it secret(which they didn't), and they would build on TOP of Wayland where does not get fit (which they don't do, they build on top of Xorg).

I don't know why people are so upset about this.
Some DE will run on XMir/XWayland for awhile.
Unity will run on Mir.
Some DE will run on Wayland.
Some may run on all 3(technically 4).

When you are an application developer, what should you support?
BSD is different from Linux. Xorg is different from Wayland. Alsa is different from OSS. PulseAudio is different from Jack.
Mir is NOT different from Wayland. And its not your *another* music player or *another* text editor.

Fragmentation 101.

All will be included in the Ubuntu repo.

Exactly as it was with KDE. I.e. with zero support. Broken.

*buntu flavors have the freedom to pick which option works for them based on upstream and thier own work.
Users have freedom to pick which *buntu they want to run.
The world goes on.
(I'm not trying to argue, I won't respond to any review, especially nic-picky, of this comment, its just a statement of facts as they stand.

Ubuntu flavors right now are having a headache how to support them.
The users indeed have a freedom to choose OS, and its a good time to give Canonical a middle finger until they start behaving like they claim to behave, ie BY BEING UBUNTU (as in African word).

[chrisb]

Could you clarify your posts? Are you talking about Ubuntu LTS or plain Ubuntu? RHEL is an entreprise distribution.
Number of packages are bad metrics because of variable like subpackages, active maintenance, and legacy.
Anyway, this is off-topic and we should get back to display manager

My point was that the official repository of Debian/Ubuntu included more software than Mandrake or Red Hat. If the official repository includes more software, then users are less likely to need to add third party repositories, and therefore less likely to end up with breakage. I ran Red Hat (pre-Fedora), Mandrake, CentOS, Debian and Ubuntu, and in my personal experience there was more package breakage under the RPM based distributions, even though I used up2date/yum/smart/urpmi/apt-rpm. I remember when it was normal to use rpmfind.net (and others) to find packages and manually fix broken dependencies. In my personal experience, under Debian/Ubuntu it was easier to find packages and satisfy dependencies, and dependencies got broken less often. Sites like rpmfind and freshrpms weren't necessary, perhaps because there were so many packages included in universe, or because apt-get worked well to make sure the installed system was coherent.

This is just my personal experience: I'm not saying this is absolutely the way it was, because maybe other people had a different experience. I'm not saying that Debian/Ubuntu were perfect either, far from it - stuff did get broken - but everything was fixable from within the system - I never had to resort to downloading binary packages from a third party web site (and that was common, the rpm guide even advised users to use third party sites for packages).

[finalzone]

My point was that the official repository of Debian/Ubuntu included more software than Mandrake or Red Hat. If the official repository includes more software, then users are less likely to need to add third party repositories, and therefore less likely to end up with breakage. I ran Red Hat (pre-Fedora), Mandrake, CentOS, Debian and Ubuntu, and in my personal experience there was more package breakage under the RPM based distributions, even though I used up2date/yum/smart/urpmi/apt-rpm. I remember when it was normal to use rpmfind.net (and others) to find packages and manually fix broken dependencies. In my personal experience, under Debian/Ubuntu it was easier to find packages and satisfy dependencies, and dependencies got broken less often. Sites like rpmfind and freshrpms weren't necessary, perhaps because there were so many packages included in universe, or because apt-get worked well to make sure the installed system was coherent.

Time has changed since then. freshrpms is legacy as the maintainer is in rpmfusion.org, a result of merging livna.org, freshrpms and dribble.
rpm dependencies issue was related to lack of maintenance and a nearly absence of policies compared to dpkg due to Debian very well documented package policies. OpenSuse, Fedora and Mageia worked altogether to improve the base package manager rpm.

I never had to resort to downloading binary packages from a third party web site (and that was common, the rpm guide even advised users to use third party sites for packages).

The reason about third party website like rpmfusion.org is the result of policies in Fedora case due to no allowed closed or no proprietary software. Because of its US origin, Fedora cannot afford to take a legal risk.

[Awesomeness]

Mir may be able to compete with Wayland, not because of any technical merit, but because of Canonical's existing user base and loyal fans

People often forget that even in FOSS many technologies are developed by corporations that need to make money. To this day Canonical loses money: http://www.internetnews.com/blog/ske...rofitable.html
Wayland's backers like Red Hat and Intel do make money. Lots of money.
I doubt many people will continue to develop Mir after Canonical goes bankrupt.

there will be apps that require Wayland, and apps that require Mir, and then these distros wouldn't be able to run them

No applications should require a specific display server. At least in Qt and AFAIK also GTK talking to the display server is handled by toolkit plugins,

without MIR noone would care about Wayland atleast for the next century and we would be using X anyway. After Ubuntu announced MIR suddenly there is some rushing for Wayland and, of course, KDE and RedHat keeping the shitstorm against Ubuntu/Canonical going.

KDE is preparing for Wayland since at least 2010: http://blog.martin-graesslin.com/blo...-on-opengl-es/
Quote: ?it?s the first part of the KWin port to Wayland?
Back then Wayland adoption was slow for two very simple reasons: Wayland did not have a stable protocol (this changed with the 1.0 release) and KWin?s main developer was not yet employed to work on KWin full time. That changed in late 2012 with the clear goal to finalize the KWin Wayland port: http://blog.martin-graesslin.com/blo.../kwin-hacking/
All that happened before the announcement of Mir.

That is just because no desktop has bothered to integrate an Android app compatibility layer. Canonical were considering it at one point, but apparently Google said they weren't interested. It was demonstrated at an Ubuntu conference running real apps though, and the developer who implemented it said it took only 1 day to do, so it can't be that much work.

It does seem a little odd that the new Samsung laptops are going to ship with the capability to run Android apps on Windows 8 when Linux desktops can't; I had anticipated a Linux compatibility layer would have been done within 6 months of the release of the Android source code.

What you're referring to is Alien Dalvik, a commercial project that runs on many platforms. To properly run all Android apps, more than just compiling the modified Java runtime is needed. Simple apps may run after only compiling Davlik. Anything else depending on special Android features such as permissions have problems.

Way back noone thought you will need a GUI Installer that could be used by your grandma. Ubuntu solved that.

The best Linux installer I ever saw was Caldera OpenLinux 2.3. Still beats many of today's installers. First it asked a few straight forward questions, then it copied the packages and while doing that offered to play Tetris.

Increased public awareness regarding alternatives to windows and mac on the desktop.

Funny that despite alleged increased public awareness GNU/Linux is still only used by 1% ? the same number as 10 years ago.

[erendorn]

Funny that despite alleged increased public awareness GNU/Linux is still only used by 1% ? the same number as 10 years ago.

Well, it's not like Apple and Microsoft are standing still and doing nothing in the meantime.
When you don't have as much cash and marketing power as the other players, maintaining your market share (especially with the market still growing, albeit slower) can somehow be considered as a success.

At least part of that success can be attributed to Ubuntu in my opinion.

[dee.]

People often forget that even in FOSS many technologies are developed by corporations that need to make money. To this day Canonical loses money: http://www.internetnews.com/blog/ske...rofitable.html
Wayland's backers like Red Hat and Intel do make money. Lots of money.
I doubt many people will continue to develop Mir after Canonical goes bankrupt.

Well, that remains to be seen. They could still turn that around and make a profit. It's not like Ubuntu is the worst operating system in the world - after all, microsoft still manages to make money and their product is total shit... Canonical would just have to take their heads out of their asses and stop flirting with microsoft - no point in attaching yourself to a sinking ship.

No applications should require a specific display server. At least in Qt and AFAIK also GTK talking to the display server is handled by toolkit plugins,

"Should"? Maybe not. But some still do. Not everyone follows best programming practices. Especially when it comes to things like games, proprietary software, things ported from other OS's. There are even now apps that depend directly on X, and there will probably be apps that depend directly on Wayland or Mir.

Funny that despite alleged increased public awareness GNU/Linux is still only used by 1% ? the same number as 10 years ago.

That's actually debatable - we don't have any accurate statistics on the real usage figures of Linux/GNU on the desktop. All we have are things like NetApplications and such, which are nowhere near accurate representations of real usage statistics, due to their methodology being total crap. The real figures could be anywhere between 2 and 10 percent, depending also on how you count dual boot systems and virtual machines...

[chrisb]

Funny that despite alleged increased public awareness GNU/Linux is still only used by 1% ? the same number as 10 years ago.

What market share are you talking about? U.S.? Global? And how is it being measured? By the user-agent of web browsers visiting predominantly English-language web sites owned by Western companies? I mean, w3schools puts Linux usage at 4.9% for May 2013, w3counter puts Linux at 2.73%, Wikimedia at 1.46%, there is quite a lot of variance, and that's with the same methodology of counting user agents. Are we only measuring desktops? If we include Android as Linux, as the Wikimedia stats do, then "Linux" is at 7.63%. What about the desktops/laptops/hybrids that are shipping with Android - "Android invades the desktop" - do we include those? It was only a few years ago when Microsoft estimated that Linux global market share was bigger than OS X (and that didn't include Android). How accurate are these figures?

I suspect the accuracy falls quite a bit when western companies are trying to estimate market share in, say, China (particularly given that a lot of the large sites where data is taken from - like Google, BBC, Wikipedia etc. are blocked in China). I would also suspect that in developing nations Linux would be a bit more popular than in the West, due to the price being a more important factor, and due to concerns over security and reliance on a western company, countries like China would be more keen to switch from Windows than the U.S. is.

Having said that, we have to be realistic: no desktop operating system has ever broken out of single digit global market share since the rise of Windows. Not one. The simple reason is that preloading is really important; once a device is sold, what percentage of buyers are going to take it home and wipe the operating system to install a new one? Probably less than 1%. That means 99% of devices are going to keep on running whatever operating system they get sold with, and Microsoft has done a very good job of ensuring that the vast majority of computers get sold with Windows. That is a really hard problem to solve, but people are working on it - Canonical reckon that about 2% of the global desktop market now ships with Ubuntu preloaded.

It seems like everyone has problems if you believe the press.. Is the Linux desktop becoming extinct? Microsoft's overall device market share has dropped from 97% to 20% in the last decade. Gartner: Microsoft Windows 15% Mobile, Desktop Market Share? iOS Drops to 17.3% Market Share.

[mrugiero]

What market share are you talking about? U.S.? Global? And how is it being measured? By the user-agent of web browsers visiting predominantly English-language web sites owned by Western companies? I mean, w3schools puts Linux usage at 4.9% for May 2013, w3counter puts Linux at 2.73%, Wikimedia at 1.46%, there is quite a lot of variance, and that's with the same methodology of counting user agents. Are we only measuring desktops? If we include Android as Linux, as the Wikimedia stats do, then "Linux" is at 7.63%. What about the desktops/laptops/hybrids that are shipping with Android - "Android invades the desktop" - do we include those? It was only a few years ago when Microsoft estimated that Linux global market share was bigger than OS X (and that didn't include Android). How accurate are these figures?

I suspect the accuracy falls quite a bit when western companies are trying to estimate market share in, say, China (particularly given that a lot of the large sites where data is taken from - like Google, BBC, Wikipedia etc. are blocked in China). I would also suspect that in developing nations Linux would be a bit more popular than in the West, due to the price being a more important factor, and due to concerns over security and reliance on a western company, countries like China would be more keen to switch from Windows than the U.S. is.

Having said that, we have to be realistic: no desktop operating system has ever broken out of single digit global market share since the rise of Windows. Not one. The simple reason is that preloading is really important; once a device is sold, what percentage of buyers are going to take it home and wipe the operating system to install a new one? Probably less than 1%. That means 99% of devices are going to keep on running whatever operating system they get sold with, and Microsoft has done a very good job of ensuring that the vast majority of computers get sold with Windows. That is a really hard problem to solve, but people are working on it - Canonical reckon that about 2% of the global desktop market now ships with Ubuntu preloaded.

It seems like everyone has problems if you believe the press.. Is the Linux desktop becoming extinct? Microsoft's overall device market share has dropped from 97% to 20% in the last decade. Gartner: Microsoft Windows 15% Mobile, Desktop Market Share? iOS Drops to 17.3% Market Share.

I don't think Android should be taken into account, since (except for its kernel) it's a completely different ecosystem. Most people don't care about the Linux part on Android, but about the Android part. A lot of people I know doesn't even know they are using Linux within their smartphones.

[You-]

I'll give you several, though really it was the combination that made Ubuntu so successful.

• Marketing. Ubuntu created a brand awareness that stretched beyond existing Linux users. I know several people who switched to Linux because they heard about how great Ubuntu was. (Yes, marketing is a non-technical problem)

• AFAIK Ubuntu generally took users away from other linux distributions. while having little impact on overall linux marketshare.
  
  

• Ease of install. I have used Linux for a long time (since Slackware days) and Ubuntu was the first desktop distribution that had a really easy to use graphical installer. And it correctly detected the hardware that didn't have open source drivers, and installed the necessary closed source drivers, so users didn't have to spend hours trawling the web trying to get their modem or graphics or sound working. You could literally install a desktop in 10 minutes and expect the usual pain points to not be a problem.