Beta

Slashdot: News for Nerds

×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Canonical Ports Chromium To The Mir Display Server

Unknown Lamer posted about 5 months ago | from the then-you-port-mir-to-chromium dept.

Ubuntu 63

An anonymous reader writes "Months after Intel ported the Chromium open-source web browser to Wayland, Chromium is now running on Ubuntu's Mir. The Mir display server port ended up being based on Wayland's Chromium code for interfacing with Google's Ozone abstraction framework. The Ubuntu developer responsible for this work makes claims that they will be trying to better collaborate with Wayland developers over this code." Grab the code hot off the press.

cancel ×

63 comments

Shuttleworth is a lunatic. (2, Interesting)

Zeio (325157) | about 5 months ago | (#46407349)

I remember at openstack portland Shuttleworth gave a live demo that failed. Ubuntu fails constantly. While Redhat tries to normalize the high rates of change in Linux, Ubuntu injects massive changes all the time while providing no stability. I have many years now working with a development team where we use Ubuntu as both product appliance and infrastructure. I have never seen a bigger mess than the trash that gets pumped out by Canonical. I used to know many Ubuntu acolytes who are converting away. Shuttleworth has spent a LOT of political capital and his promises are empty. I really dislike Canonical, I dislike Ubuntu, and I really dislike this arrogant loser Shuttleworth. Bad packages, kabi and abi changes. A preseed/install system that is pathetic, instability, bleeding edge, bad stable kernel management, horrible backporting fixes, unstable userland.

Im done with Canonical and Shuttleworth.

Re:Shuttleworth is a lunatic. (0, Insightful)

Anonymous Coward | about 5 months ago | (#46407449)

And don't forget they also steal your data and pass it Amazon.

Re:Shuttleworth is a lunatic. (1, Funny)

fsck-beta (3539217) | about 5 months ago | (#46407481)

That and Wayland is the technically superior choice...

Re:Shuttleworth is a lunatic. (5, Insightful)

squiggleslash (241428) | about 5 months ago | (#46408155)

X11 is the technically superior choice. #getoffmylawn

Still, from that point of view, the Mir thing has been a success for all those itching to replace a stable, mature, well known and tested, versatile, and powerful windowing system with a new and untested stripped down windowing system simply because they don't understand why someone would want some of the features X11 has, and are under the impression it's bloated because it's bigger than Windows 2.0 was in 1989.

Mir has helped create the illusion the decision has been made already. We are transitioning, no more debate is needed (or will be accepted) as to whether we should, and the question is what we should transition to.

Much the same mistake was made with GNOME 2 to GNOME 3, a transition that Ubuntu helped along in the same way with Unity. Users rebelled, with forks like Mint attempting to roll back the damage, but the end result was a deterioration in the perception of GNU/Linux as a potential replacement for Windows. Distributions based upon GNOME 3 and Unity got the "slick", "professional", treatment, with users finding fast that it wasn't what they actually wanted. The GNOME 2 hold-outs didn't have the resources to ensure GNOME 2's forks had the same level of support, and so ended up with systems that looked to new users dated and ugly.

We will see the same with Mir/Wayland, except worse. We'll have five to ten years of having to deal with an immature windowing system that, by the end of the process, has just as many hacks and quirks as X11 but will almost certainly still lack key features X11 offers. X11 holdouts will find themselves using an increasingly unreliable and unstable platform as newer hardware requires new device drivers, without the level of support needed within the X.org X11 community to support them.

We're all going to lose. The best free software users can hope for now is that Google continues to extend Android to eventually offer a decent desktop experience. I don't know why they would, perhaps to replace ChromeOS, but at least you're looking at something mature there. But that's not here now, and the next five years will be rough for GNU/Linux users. We'll likely be as mainstream as FreeBSD by the end of it.

Re:Shuttleworth is a lunatic. (1)

Pausanias (681077) | about 5 months ago | (#46410011)

Change is not bad. X11 needs change. Perhaps they should have worked with X11, but from what I hear the X11 authors themselves didn't want to keep it.

X11 is broken in one serious way. X11 window forwarding over network is slow. Pathetically slow. Maybe some people who only ever forward terminals from X11 and whose entire computer use case involves manipulating ASCII characters might be fine with this, but those of us who work with graphics of any kinds that third-party hacks are required to make X work over a network. Witness the abomination that is NoMachine NX, where they had to basically rewrite X11 so that it would work fast over a slow network connection.

X11 may have many plusses, but this inability to do fast networking is just stupid and needs to change---whether via updating X11, Wayland, Mir, or what have you.

Re:Shuttleworth is a lunatic. (0)

Anonymous Coward | about 5 months ago | (#46411059)

Change is not bad. X11 needs change. Perhaps they should have worked with X11, but from what I hear the X11 authors themselves didn't want to keep it.

This seems to be mostly a myth.

X11 is broken in one serious way. X11 window forwarding over network is slow. Pathetically slow.

I still use it. But yes, it is not as fast as it could be. This ist mainly because the toolkits (based on xlib) do synchronous requests. This is not a fundamental problem of X!

Maybe some people who only ever forward terminals from X11 and whose entire computer use case involves manipulating ASCII characters might be fine with this, but those of us who work with graphics of any kinds that third-party hacks are required to make X work over a network.

I use it just fine. But yes, you need a fast network with todays bloated desktops.

Witness the abomination that is NoMachine NX, where they had to basically rewrite X11 so that it would work fast over a slow network connection.

Yes, it is crap. But NX is a mess for historical reasons, not because one "basically has to rewrite X11" to make it fast. Xpra is much better, although it does not use X as a transport - which is a mistake IMHO.

X11 may have many plusses, but this inability to do fast networking is just stupid and needs to change---whether via updating X11, Wayland, Mir, or what have you.

With Wayland and Mir, you do not get any networking. The advantage of X as a protocol would be compatiblity, and that the core protocol is very flexible and can be used to implement many new techniques and interesting things for client-to-client interaction (some people think copying pixels is all which is needed - they have no clue). It would be a really cool base for many distributed applications. Throwing this away is immensily stupid. Fixing toolkits on X to make it fast would be much more worthwhile (and to properly support disconnect/reconnect - why does this still not work?). But the money is now in mobile adventures of Canonical or Intel, where all this is seen as a distraction. This is why Mir and Wayland gets funded - not to make your desktop better.

+1 Insightful (1)

turgid (580780) | about 5 months ago | (#46412033)

Tell it like it is!

Re:Shuttleworth is a lunatic. (1)

Pausanias (681077) | about 5 months ago | (#46412529)

If "fixing toolkits" is all that is needed, why hasn't anybody done it? How about yourself?

As far as I can tell, everyone who has tried to speed up X over the network has taken a different route. Maybe once you actually look at the toolkit code, it's not that easy---otherwise it would have been done ages ago.

[quote]But yes, you need a fast network with todays bloated desktops.
[/quote]
"Bloated desktop"? Hardly. Try putting up a 10-year old tcl/tk based GUI which allows you to manipulate 2D graphics by live-changing brightness/contrast. Try doing this over a 12Mbps down connection. This is my use case. And it is ridiculously slow---unusable.

ssh -X (1)

Tenebrousedge (1226584) | about 5 months ago | (#46410277)

You must not be an X11 developer. You are also, if you can talk about Linux becoming somehow comparable to FreeBSD, not very aware of Linux usage. Linux sees far more use as a server than a desktop, and while it's possible that some other project starts eating up marketshare in that segment, I would be extremely surprised if that happened, and then we have the supercomputer, embedded, and mobile markets.

Your post is devoid of technical arguments for X11, because X11 is technically a clusterfuck. [youtube.com] It does far too many things, and none of them well. The print server is probably the worst offense there, but by no means the only one. How many rendering interfaces are there, again? Tell me about why exactly I can't adjust my computer's volume while the screensaver is active, and how that is a good idea.

The only approximation of a coherent argument against Wayland is a repetitive drone of "network transparency means my ssh -X works as expected". Cool, but that's not what that means and that's not going away, and certainly not any time soon. Wayland is not ready for production, and not expected to be anytime soon, and it's blindingly obvious to everyone that it will not be considered ready for production until ssh -X works correctly.

I have to ask if you're even a developer, because my limited programming experiences have led me to believe that if an experienced dev says that the entire approach is crap and needs to be scrapped, more than likely this is true. It is unreasonable to expect that the display solutions of 30 years ago are remotely appropriate for today. It's also a deeply held Unix principle that programs should have a limited scope; why anyone would complain about a "stripped down windowing system" is beyond me. Especially with all of the obsolete, unused crap that even X11's adherents cannot deny that it contains.

So one the one side we have technical and philosophical arguments, and on the other -- please, please have something better than "ssh -X won't work". If you haven't seen the video, it's amusing and informative. Also if you haven't seen the video, you don't know what you're arguing.

Re:ssh -X (0)

Anonymous Coward | about 5 months ago | (#46411983)

You must not be an X11 developer.

That all X11 developers decided that X is crap and Wayland is the way to go seems largely a myth. Wayland has good propaganda, but let's talk about technical arguments.

You are also, if you can talk about Linux becoming somehow comparable to FreeBSD, not very aware of Linux usage. Linux sees far more use as a server than a desktop, and while it's possible that some other project starts eating up marketshare in that segment, I would be extremely surprised if that happened, and then we have the supercomputer, embedded, and mobile markets.

Quality of the Linux desktop is getting worse in recent years. I was happy for a decade, before people started to fuck it up. Wayland just seems to mean more regressions.

Your post is devoid of technical arguments for X11, because X11 is technically a clusterfuck. [youtube.com]

A link to a youtune video is not an argument. Please quote something here, so that it can be discussed.

It does far too many things, and none of them well. The print server is probably the worst offense there, but by no means the only one.

This is not a problem. X is a very flexible protocol, so some used it to define a protocol for a print server. I don't think any modern X server has this. So who cares?

How many rendering interfaces are there, again?

XRender is 10 years old and defines a modern rendering API. The older API is needed for backwards compatibility and are almost no burden in terms of code and resources. Why do you think this is a problem? How many interfaces for file I/O are there in the kernel?

Tell me about why exactly I can't adjust my computer's volume while the screensaver is active, and how that is a good idea.

Why do you think this is a fundamental problem with X?

The only approximation of a coherent argument against Wayland is a repetitive drone of "network transparency means my ssh -X works as expected".

This is a damn good argument. Also the main argument is not network transparency per se, but network transparency using a backward and forwards compatible protocol.

Cool, but that's not what that means and that's not going away, and certainly not any time soon.

It will go away, because new Wayland clients will not support it. Also it is not clear if XWayland will be installed everywhere by default. For example, the Jolla phone seems to ship without it.

Wayland is not ready for production, and not expected to be anytime soon, and it's blindingly obvious to everyone that it will not be considered ready for production until ssh -X works correctly.

Haha. Upgrades in recent years broke so much stuff on the desktop that I somehow doubt this.

I have to ask if you're even a developer, because my limited programming experiences have led me to believe that if an experienced dev says that the entire approach is crap and needs to be scrapped, more than likely this is true.

Well, others say something else. This is just an opinion.

It is unreasonable to expect that the display solutions of 30 years ago are remotely appropriate for today.

What - exactly - is wrong with the X core protocol? A modern rendering API (XRender) was defined a while ago,

It's also a deeply held Unix principle that programs should have a limited scope; why anyone would complain about a "stripped down windowing system" is beyond me.

I am not opposed to refactoring the X server, but X is fundamentally more stripped down the Wayland. Wayland combines display server, window manager, and compositer into one program. This is not the UNIX principle.

Especially with all of the obsolete, unused crap that even X11's adherents cannot deny that it contains.

There is some stuff which is not used by modern clients, but this is needed for compatibility. Backwards compatibility is a good thing. And more importantly: This code does not cause much bloat or affect modern clients in any negative way.

So one the one side we have technical and philosophical arguments,

There are basically no good arguments *for* Wayland and breaking compatibility except "let's get rid of some minor amount of old code"

and on the other -- please, please have something better than "ssh -X won't work". If you haven't seen the video, it's amusing and informative. Also if you haven't seen the video, you don't know what you're arguing.

A youtube link is not a valid replacement for an argument. If you think there is a good argument, bring it up so that it can be discussed.

Re:ssh -X (1)

Tenebrousedge (1226584) | about 5 months ago | (#46413747)

The slides are here. [freedesktop.org] While I respect the idea that video is an inefficient means to convey information but, since this is an issue that you seem to care about, you may want to take the time to educate yourself. I believe it is possible to speed up playback of youtube videos. This article [phoronix.com] conveys some similar points, but not much depth. Here is an architectural overview. [freedesktop.org]

Why don't you go over some of this information, and we can take this conversation over again from the top. You should find most of your concerns answered. I'm not particularly interested in spoon-feeding it to you, however -- please accept my apologies on that score.

Re:ssh -X (0)

Anonymous Coward | about 5 months ago | (#46418799)

You didn't even bother to answer the specific problems addressed in my reply. But let's talk about the slides. Even at first glance, the slides are full of propaganda and insults ("internet peanut gallery" etc). and very light on technical details. Let's discuss some specific points:

- That X has API for rendering text or drawing rectangles which are not used by modern clients is well known. Is is worthwhile to keep them for compatibility reasons. A valid argument whould be if those API must be used by modern clients. But this is not the case. There are optional and provided for backwards compatibility.

- The next topic discussed is that the implementation of the X server was a mess before X.org took over. This is irrelevant. Also is it irrelevant that there are five different display management extensions, and several buffer management extensions. This simply shows that X was able to evolve in time. How is this bad? It tells me that the core framework is very good and flexible.

- The argument that SHM or direct rendering does not work over the network is just laughable. Ofcourse it does not, it is an optimization for local clients to get pixels faster to the server/graphics cards. In no way does this mean that network transparency is broken. 'ssh -X' works just fine (not as fast as it should), but it works.

- He then says that more stuff is not done client-side nowadays. Again, this only shows how the core protocol if very flexible and let's you do things today that was not anticipated in the 1980s. If the design would force you to do things on the server side you would rather do on the client side, this would be an argument against X. But again, this is not the case: The protocol is very flexible and let's you do it the old or the new way.

- The first real technical argument is on slide 99. That X11 is a bad IPC mechnism. The argument is that there are a lot of blocking calls (InternAtom) on program startup. But also this argument is wrong. All these calls are asynchronous in the X protocol. It is only synchronous if you use Xlib, which is a known problem with his library. When using the more modern XCB all this latency can be avoided. So how is X11 a bad IPC mechanism? Because some library uses it in a stupid way?

- The next slides are valid critisicm ('every frame perfect'), as far as I can see. But they don't provide any justification why this can not be fixed in X. In fact, this seems to be fixed already with the new Present extension and DRI3. (Also compare the Wayland FAQ. They completely admit that this can be done in X.)

- The next slides are about the screensaver issue. I don't get it (maybe I have to listen to the talk again). The point seems to be that in X it is done with the generic protocol, while Wayland does have specific support for screensavers which allows to do gets some details right (volume keys). I don't get why this is unfixable in X though. The worst case could be that one has to add a screensaver specific hack to X. But this would be no different than hard-coded support in Wayland, wouldn't it?

In conclusion, this does not convince me at all. X seems is fastly superior protocol. Wayland is a bit simpler, but much less flexible. And this comes at a very high price: backwards compatibility is lost. The usual arguments that X is bloated, unfixable broken, or forces some outdated way of doing things on modern clients do not seem true, and I cannot see any argument in this talk which convince me otherwise.

Re:Shuttleworth is a lunatic. (1)

BlackPignouf (1017012) | about 5 months ago | (#46411111)

I'm happy with Mint-MATE, thank you very much, and sure as hell doesn't look dated and ugly.

Re:Shuttleworth is a lunatic. (1)

Burz (138833) | about 5 months ago | (#46413077)

X11 should be dustbinned just for the lack of multi-target network transparency. You know, the limitation that says while OSX and Windows users can efficiently share apps and desktops in a teleconference, Linux systems have to use VNC to toss around bitmap deltas instead. Its like getting a shot of Novocaine in the mouth everytime you head out to a party.

Oooooh, wait! Did I just attack X11 on its hallowed territory... Network transparency?! Well, indeed I have and its true that X11 has not gotten any overhauls to support this very important and common use case.

Security also stinks to high heaven on X11, and it took an OS like Qubes completely re-worked around a VM security model to address that architectural flaw (regular hypervisors like VMware won't even protect you). The priestly developers of X11 implementations do not appear to give a rats ass.

This stack (and its anachronistic neckbeard clique) has run its course and should have been on its way out 10 years ago. I think you're wrong about developing replacements for X11; Apple users never regretted it for an instant.

BTW, I don't know about you but I'm tired of my Linux UI's being interspersed with character-mode upchuck, screens flickering and popping momentarily in an out of existence whenever something different happens in the runlevel or login status or number of displays.

Re:Shuttleworth is a lunatic. (1)

lister king of smeg (2481612) | about 5 months ago | (#46416711)

X11 should be dustbinned just for the lack of multi-target network transparency. You know, the limitation that says while OSX and Windows users can efficiently share apps and desktops in a teleconference, Linux systems have to use VNC to toss around bitmap deltas instead. Its like getting a shot of Novocaine in the mouth everytime you head out to a party.

Oooooh, wait! Did I just attack X11 on its hallowed territory... Network transparency?! Well, indeed I have and its true that X11 has not gotten any overhauls to support this very important and common use case.

Security also stinks to high heaven on X11, and it took an OS like Qubes completely re-worked around a VM security model to address that architectural flaw (regular hypervisors like VMware won't even protect you). The priestly developers of X11 implementations do not appear to give a rats ass.

This stack (and its anachronistic neckbeard clique) has run its course and should have been on its way out 10 years ago. I think you're wrong about developing replacements for X11; Apple users never regretted it for an instant.

BTW, I don't know about you but I'm tired of my Linux UI's being interspersed with character-mode upchuck, screens flickering and popping momentarily in an out of existence whenever something different happens in the runlevel or login status or number of displays.

X11s security problems the ones patched resently by OpenBSD? One of the many operating systems besides Linux that uses X. One of the operatting systems that wayland and Mir have stated they have no plans for supporting?

Re:Shuttleworth is a lunatic. (1)

thegarbz (1787294) | about 5 months ago | (#46415563)

X11 is the technically superior choice. #getoffmylawn

The get off my lawn comment is quite appropriate in this case, because in many cases X11 vs Wayland is the equivalent of an old person using modern technology vs some young kid. The old person has a wealth of knowledge and experience behind them, and along with it an entire lifetime of prejudices on how things are supposed to be, and as such they often have a problem understanding workflows on a computer system (how may times have you seen people hit the X in the top corner of the dialogue box asking them to make a decision).

The new kids are young and immature. They don't know everything but they claim they do. They lack experience, and they can't actually do everything, but the things they can do they do far better and more efficiently than the old timers.

To understand why I make this analogy did you realise you can output the X11 framebuffer to a printer and actually print GLXGEARS frame by frame? No one actually wants to do that but it's a feature available in X11. The protocol was designed in a way that made it insanely powerful, yet insanely complex and inefficient, much like grandpa hitting the X button on the dialogue box and then wondering why his printer didn't print. For all the power of X it has created some asinine limitations which mean many modern workflows don't work any more. You got volume soft keys on your laptop? Shame they don't work when the screen saver is running, unless of course the screen saver was written to work around X11's peculiar shortcomings.

X11 is NOT technically more superior, it's just more mature.

Re:Shuttleworth is a lunatic. (0)

Anonymous Coward | about 5 months ago | (#46417503)

The protocol was designed in a way that made it insanely powerful, yet insanely complex and inefficient, much like grandpa hitting the X button on the dialogue box and then wondering why his printer didn't print. For all the power of X it has created some asinine limitations which mean many modern workflows don't work any more.

Could you point to a specific problem with the core X protocol, which makes in insanely complex and inefficient. I don't see it. Or to a specific "asinine limitations which mean mean many modern workflows don't work any more." I also don't see it.

You got volume soft keys on your laptop? Shame they don't work when the screen saver is running, unless of course the screen saver was written to work around X11's peculiar shortcomings.

This gets repeated a lot. Is this the reason to scrap decades of backwards compatibility? Because the the volume key doesn't work when the screen saver is running?

Re:Shuttleworth is a lunatic. (2)

JoeMerchant (803320) | about 5 months ago | (#46407523)

So, don't hold back, tell us how you really feel....

What's your idea of an easy migration path out of Ubuntu, preferably something still Debian based, but really, anything that serves the ideals that Ubuntu used to address 6 to 10 years ago and has gotten away from?

I see lots of hard-core Arch users around, has anybody made something like "Stable KDE on Arch" into a supported distro yet?

Re:Shuttleworth is a lunatic. (3, Interesting)

zodmaner (3564157) | about 5 months ago | (#46407667)

Just use Debian, seriously. I'm running KDE 4.11.5 on Debian Sid for about a year now, and it has been a very pleseant experience. Also, Sid works more or less like a rolling release distro, so you will constantly get the latest version of the softwares you use. It's not as fast or as bleeding edge as some distros, but it's more way, way, _way_ more stable. For example, Sid gets the latest version of Firefox about a week after the official release, sometimes sooner than that. PS. I'm an ex-Arch users, having switched away from Arch due to many reasons, and couldn't be happier.

Advice (1)

morgauxo (974071) | about 5 months ago | (#46409055)

A long time ago I was a big Debian fan. I had two problems with it, eventually I became pissed off and moved on to Gentoo. I still love Gentoo on my desktop but would like something less labor intensive for any and every other computer I might have to maintain. I am wondering if Debian has improved and if I might want to use it for this.

My first problem was packages asking for user input during installation. I wanted to start a dist-upgrade and walk away. Often I would do so, coming back hours later expecting all 100s of packages to be updated only to find that it only got a few down it's list, started asking some stupid question about something I didn't even care about and stopped there, waiting for me to press a key.

Then there were the repos. Debian's package repos were pretty conservative, prefering stable over bleeding edge. Overall that is a good thing. But it meant there were always specific packages, especially when using Debian as a desktop where I just had to get a more recent version from somewhere else.

If you wanted to keep everything neat, tidy and managed by apt-get and dpkg the best way was to add third party repos. Again, not a problem. I don't mind doing that. My problem was that the third party repos didn't always stay in sync very well with the official Debian ones. The result was that often, once a few third party repos are added and the install is no longer pure official Debian, when trying to update one package Debian would suddenly want to un-install many many othes. It was easy to miss a message, press the 'Y' key and before you know it half your system is gone! This was especially bad given my first Debian problem, looking for an unattended upgrade. I was pretty prone to just pushing 'Y'.

So, anyway, my point isn't to complain about Debian. That was over 10 years ago! What I want to know is "has it improved?".

Thanks!

Re:Advice (1)

JoeMerchant (803320) | about 5 months ago | (#46409505)

Just rambling here, but I did Gentoo about 10 years ago, back when Gentoo was the most practical way to get a 64 bit Linux OS... it took about a half day to compile and install, which isn't the end of the world, but it left me in mostly "static" maintenance mode - I wasn't inclined to try anything because of the potential cost. I was kinda jazzed about all my packages being recompiled with optimization switches tailored to my AMD64 system, but I don't think it did anything noticeable for me.

I did one full system re-compile-install of Gentoo after about 2 or 3 years, and when that didn't address my problem du-jour, I went with Ubuntu and stayed with it until the rig died of dust accumulation. Ubuntu was definitely easier to "try things" in, but it, too, suffered the classic two rednecks driving in a car fate - what are their last words? "Buckle up, I'm gonna try somethin' here..." At least a total Ubuntu system re-install only takes about an hour...

"static maintenance mode" (1)

morgauxo (974071) | about 5 months ago | (#46428475)

Yeah, I can definitely see how waiting for compiles could discourage people from trying things. I've adopted a mode of ssh to home first thing in the morning from work, start a screen session, start something big emerging". Then I can sign out, focus on work and when I get home I have something new to play with. If I think the compile will take longer than that, and I think I will want to use my computer that night I just set it to some low priority nice level. That way I can go on using my computer for every day stuff while it continues to build. It can finish whatever day it finishes.

This routine works for me but I can imagine it isn't for everyone. For me it has been better than the issues I had back when I used binary packages. But.. It's also not something I want to do on multiple machines.

Re:Advice (1)

lordcorusa (591938) | about 5 months ago | (#46409945)

Apt does feature an unattended-upgrades mode. It's not the default, which is annoying, but it's pretty easy to configure. It's one of the first things I configure on a new Debian box.

As for outdated packages: Debian unstable and experimental usually contain cutting and bleeding-edge versions of most open source software that is packaged in Debian. Unlike the grandparent poster, I would not recommend running sid (AKA unstable) as my main repository, because doing an apt-get upgrade has occasionally wrecked my system on sid. I use Linux Mint Debian Edition, which is based on regular tested snapshots of testing, but I do occasionally need cutting-edge packages. I would recommend looking into apt preferences [debian.org] , which provides a nice way to grab specific cutting-edge packages from unstable and experimental while dragging in a minimal number of unstable dependencies. So far I have had almost no problems doing this, and certainly far fewer problems than using 3rd party repositories.

Re:Advice (1)

zodmaner (3564157) | about 5 months ago | (#46410501)

This is an excellent answer and match my experience exactly.

Just wan to add that I have been doing an aptitude full-upgrade/aptitude upgrade almost weekly, and have only encountered 2 packages that require user interactions so far: glibc (warning about restarting some services) and OpenJDK (licensing stuffs). Other packages just install "silently" during an upgrade.

Re:Shuttleworth is a lunatic. (1)

lordcorusa (591938) | about 5 months ago | (#46407735)

I'm quite happy with Linux Mint Debian Edition. It provides a nice middle ground between stable-but-out-of-date-and-highly-political Debian and unstable-but-nice-features Ubuntu. LMDE is basically a tested snapshot of Debian testing (like CUT was supposed to be, but actually regularly maintained) with a lot of the ease-of-use features (proprietary drivers, etc) of Ubuntu. Unlike Ubuntu or regular Linux Mint (which is based on Ubuntu), LMDE is completely compatible with Debian. I can use the apt preferences mechanism to pull in some more recent packages from Debian unstable and experimental, more or less seamlessly.

Both the MATE and Cinnamon editions of LMDE are good, depending on your GUI preferences. MATE seems a bit more mature, but it's definitely old technology. Cinnamon provides modern compositing coolness, and has been making some huge improvements, but it's still rough around the edges with respect to some of the GUI configuration screens.

They just cut the new ISOs a few days ago, so the packages on them are fresh.

Re:Shuttleworth is a lunatic. (3, Informative)

Anonymous Coward | about 5 months ago | (#46407799)

This comment is a fabrication, I was working on the demo for OpenStack Portland and it worked just fine. The rest of your post is just a rant with no technical details in it whatsoever.

Re:Shuttleworth is a lunatic. (0)

Anonymous Coward | about 5 months ago | (#46461071)

Liar. I was there. You all scoured the videos from Youtube too. Ill be posting the video of your failure when I find it.

I found the rant to be completely true. And Shuttleliar doesnt count centos+redhat in his lies.

Re:Shuttleworth is a lunatic. (5, Insightful)

Anonymous Coward | about 5 months ago | (#46407803)

I really love Ubuntu. Everything Just Works, TM. You just download it, usage and installation is a breeze. Hardware compatibility is off the charts, even compared with Windows. The amount of online information if you run into problems is amazing, as well. I don't pay Ubuntu for support, yet each problem I've encountered is already documented somewhere and a workaround/fix available. The usability is on par or better that Windows, and close to Mac. All programs that have Linux versions come neatly prepackaged for Ubuntu.

I get why purists hate it, because it takes away the nice hours and days of tinkering for something to work. Yep, it's not as customizable in the sense that is difficult to change something and get it to work better. But if you want get something done WITH computers and not FOR computers, it's amazing.

AC for corporate reasons.

Re:Shuttleworth is a lunatic. (0)

TyFoN (12980) | about 5 months ago | (#46409189)

Posting to remove moderation.
I realize this was not meant to be funny, but serious.

Re:Shuttleworth is a lunatic. (1)

Wootery (1087023) | about 5 months ago | (#46410401)

I get why purists hate it, because it takes away the nice hours and days of tinkering for something to work.

No, you quite clearly don't get it. The clue, regarding the objection of the purists, is right there in the word purist. Relevant FSF page [gnu.org] .

I'm not a purist myself (I run Mint Linux in VMware on Windows 7), but you are confusing Free Software proponents with OS hobbyists. The former advocate gNewSense. The latter play with Syllable, MenuetOS, Haiku, etc.

Re:Shuttleworth is a lunatic. (0)

Anonymous Coward | about 5 months ago | (#46415643)

I wanted to love Ubuntu, and did at first. Then stuff started breaking. I would post and post to forums, and never get any ROI for the 'ubuntu' spirit I put in to make Shuttleworth's project better.

Shuttleworth is the reason many use Linux Desktop (4, Insightful)

kervin (64171) | about 5 months ago | (#46407875)

Hate him all you want. But when I looked around for a Workstation preinstalled with Linux, Ubuntu was the only serious choice I got. Redhat didn't even have a preinstalled system they would sell me. That's right, they haven't even paid enough attention to Linux Desktop to have a partner provide a well-spec'ed, modern, supported Linux laptop.

After a lot of digging I found a list of Windows laptops Redhat swore would also run their OS. But asking users to buy one OS ( Windows ) and reinstall another is an automatic fail for the vast majority of desktop buyers. Not that I can't do install an OS, but not having a supported OS is just not worth my time anymore. I'm no longer in college with lots of time to tweak and troubleshoot.

I wish I could go to Redhat.com, enter my credit card and have a partner laptop shipped to me in a few weeks. Complete with modern specs and OS support direct from Redhat. But that's not possible even if I'd happily pay a premium. At least Ubuntu has System76 [system76] .

Doesnt mean he's not a moron (0)

Anonymous Coward | about 5 months ago | (#46408621)

Most people eat mcdonalds hamburgers, but they certainly arent the best burgers around

Usually... (1)

Dcnjoe60 (682885) | about 5 months ago | (#46409759)

Usually I buy a computer (laptop or desktop) based on its hardware features and then the OS. Yes, in theory, it sounds great to be able to have linux preinstalled, but most of those systems are pretty pricey compared to the equivalent hardware with Windows preinstalled. Since I tend to reconfigure a default install anyway, having to actually do the install is minor inconvenience, particularly if I'm saving $200-$300 over having it preinstalled. While my time is valuable, so is my money.

To each their own. However, I admit that if there were a competitively priced laptop with linux on it, I would definitely consider it. Then again, that doesn't mean I wouldn't install my distro of choice over the top of it.

Re:Shuttleworth is the reason many use Linux Deskt (1)

Burz (138833) | about 5 months ago | (#46413185)

RedHat/Fedora is way outclassed by Ubuntu in terms of supported hardware. Just check out their respective HCL pages.... I dare any RedHat "workstation" lover to find out if they can stomach the difference and RedHat's obvious neglect.

The RedHat ken only makes *noises* about supporting desktops. There is no commitment or vision. Fedora is a only testbed distro for haphazardly plopping misc desktop components onto a base server OS.

Re:Shuttleworth is a lunatic. (0)

Anonymous Coward | about 5 months ago | (#46407953)

I've used Ubuntu and Red Hat for many years. I can assure you, the door swings both ways. Each one has their shining moments as well as the "oh my gosh what on earth are these morons thinking" type of moments. Red Hat by no stretch of the imagination is problem free. I mean, they're great, but they have problems too. Ignoring to mention that only suggests a level of bias. Sorry, blunt 2c.

Re:Shuttleworth is a lunatic. (0)

Anonymous Coward | about 5 months ago | (#46407959)

Yea, and Ubuntu raped my daughter and Mark Shuttleworth eats little babys for breakfast.

Re:Shuttleworth is a lunatic. (1)

wiredlogic (135348) | about 5 months ago | (#46408775)

I have many years now working with a development team where we use Ubuntu as both product appliance and infrastructure.

Backend bits are better implemented on Debian. The slow release schedule and more rigorous release testing provide more stability.

Re:Shuttleworth is a lunatic. (1)

jones_supa (887896) | about 5 months ago | (#46409341)

I remember at openstack portland Shuttleworth gave a live demo that failed. Ubuntu fails constantly. While Redhat tries to normalize the high rates of change in Linux, Ubuntu injects massive changes all the time while providing no stability.

Well yeah. I believed in Ubuntu for a long time, because I wanted it to become at least a "de facto" distro instead of all the fragmentation. But indeed the quality assurance is really bad in Ubuntu. If you want a slow and buggy Linux, install Ubuntu.

Shuttleworth is playing Batman (1)

tiago.bonetti (1995614) | about 5 months ago | (#46409453)

Shuttleworth is not what Linux users want, is what they need. Competition makes things better look at IE after Chrome. Wayland needed competition, it maybe better as many said, but it was going slowly. Mir maybe going nowhere but it's going fast, and now everyone talking about them. Shuttleworth may just throw Mir away once Wayland is ready.

Re:Shuttleworth is a lunatic. (1)

Pausanias (681077) | about 5 months ago | (#46409799)

You sound like someone who's gotten burnt by installing non-long-term-support Ubuntu in a production environment. That was your error, really; the non-LTS releases get minimal support, so installing them in an infrastructure-critical environment is pure silliness.

I've never had an issue with LTS releases... I have a machine that's been continuously updated since Ubuntu 8.10 (non-LTS), and the thing has miraculously upgraded with zero hitches via 8.10->9.10->10.04->12.04 and soon to be 14.04. This was a 2008 Mac Pro, to boot, so I've enjoyed incredible support in terms of mac-specific drivers and even the Broadcom wifi working out of the box.

Shuttleworth has inspired many people, and while he's made mistakes he's not afraid to push the boundaries and make disruptive changes because that's what keeps things going. The formula of disrputive changes in non-LTS with stability in LTS works great. Just stick to 0.1 releases of LTS Ubuntu and you'll have a stable system.

Re:Shuttleworth is a lunatic. (0)

Anonymous Coward | about 5 months ago | (#46461099)

bzzt. never used non LTS. You are a guy with a single laptop. You must know a lot about how Ubuntu works on a myriad of computers. Snicker.

I haven't watched the video (0)

Anonymous Coward | about 5 months ago | (#46407351)

but from the still image the guy looks like a monkey

Re:I haven't watched the video (2)

Chrisq (894406) | about 5 months ago | (#46407479)

but from the still image the guy looks like a monkey

he's a code monkey!

fuck beta (-1)

Anonymous Coward | about 5 months ago | (#46408797)

He's white, not black

Re:I haven't watched the video (1)

Zaiff Urgulbunger (591514) | about 5 months ago | (#46411371)

but from the still image the guy looks like a monkey

Ummmm, I'm gonna need you to go ahead and have those TPS reports for us this afternoon.

Posting to undo bad mod. (0)

Anonymous Coward | about 5 months ago | (#46407357)

Posting to undo bad mod, sorry..

Re:Posting to undo bad mod. (2)

oodaloop (1229816) | about 5 months ago | (#46407447)

Does posting AC undo bad mods?

Why is this on Slashdot? (0)

rmdingler (1955220) | about 5 months ago | (#46407367)

News for nerds, my ass!

Sorry, it was funny in my head.

Mir? (2)

slapout (93640) | about 5 months ago | (#46407407)

I thought Mir [wikipedia.org] crashed long ago

Re:Mir? (1)

bluefoxlucid (723572) | about 5 months ago | (#46407517)

They mistyped NIHere.

Re:Mir? (1)

Merk42 (1906718) | about 5 months ago | (#46407583)

...and in some alternative universe Ubuntu is using Wayland and people are complaining that Canonical is using someone else's work.

Damned if you do, damned if you don't.

Re:Mir? (2)

bluefoxlucid (723572) | about 5 months ago | (#46408227)

Pretty much. In this case, it's doubtful Wayland is at critical bodies and so adding the programming resources from Mir onto Wayland would produce a product with better quality and a richer feature set in a shorter time. That's how you make projects successful; in this case, of course, projects have a much more flexible budget of "anyone who has time and desire"--with the double-edge that budgeted resources aren't assigned by a higher power, but rather volunteered.

Canonical has volunteers in a pool which are budgeted across multiple projects, and so budgeting these resources to take advantage of an existing project's resources gives more ROI than building and maintaining their own. This is more traditional budgeting, and means that Canonical actually has decision making power over budgets: they don't just start doing something and hope people join in, but rather they influence how resources they already possess are spent. In this case it's not entirely traditional: many resources are volunteer, and there is a probability that more volunteers will appear or some will leave; but they do have some influence to direct resources at efforts.

They are using that influence inefficiently. It's the same as if a person decided to get a car, and so bought metal and fuel and made fire and machined all the parts for a $30,000 Ford Mustang, spending $350,000 in the process. They may say: "I wanted a larger engine". They may say many things. In the end, however, if their explanation is rational and utility--if it is that a Ford Mustang is lacking in some way, rather than that they would like the experience of building such a thing or that they just "want to"--then the judgment of their decision stems from three objective assessments. The first: is the Ford Mustang essentially lacking in that way? The second: Is this important? And the final: Can we more effectively correct this deficiency in any other way--by finding a better car, by modifying a Ford Mustang, or by convincing Ford to change next year's model?

There are two criticisms here. Canonical asserts that Wayland is inadequate, hence Mir; criticism comes in the form that it would be more efficient and effective for Canonical to devote the resources spent on Mir instead to improve Wayland. The second: we assume Canonical behaves in an irrational manner or to other rational considerations (such as vendor lock-in to Mir, some future ability to implement DRM, etc.), which we criticize directly.

They did, both of them (0)

Anonymous Coward | about 5 months ago | (#46407525)

Ubuntu could have been the one Linux to rule them all.

Re:Mir? (1)

Anonymous Coward | about 5 months ago | (#46407535)

In Soviet Russia, display server crashes you!

Re:Mir? (1)

rossdee (243626) | about 5 months ago | (#46407727)

Putin wants it back

Re: Mir? (1)

morgauxo (974071) | about 5 months ago | (#46409061)

It did! So did Ubuntu. What is your point?

Wait, what? (3, Interesting)

jones_supa (887896) | about 5 months ago | (#46407697)

My logic says that the toolkit that Chrome uses should be ported to have a Mir backend, rather than Chromium itself? I guess Google uses so much in-house stuff that it makes this necessary.

Not that I would be interested in Ubuntu anyway. The Unity desktop is laggy and, I'm not a big fan of having a custom display server (Mir) instead of the widely-adopted Wayland.

Re:Wait, what? (2, Insightful)

Anonymous Coward | about 5 months ago | (#46407883)

None of the serious distros use Wayland yet. I would not call it widely-adopted.

Re:Wait, what? (1)

jones_supa (887896) | about 5 months ago | (#46407999)

Yeah, I guess I have said "widely-planned".

Re:Wait, what? (1)

Burz (138833) | about 5 months ago | (#46413229)

None of the serious distros use Wayland yet. I would not call it widely-adopted.

And the chances of that changing are poor, given that Ubuntu and its spinoffs are the only popular distros that are even capable of handling multi-monitor setups correctly.

Apple (actually, NeXT) taught us long ago that if there is one area where you should second-guess and buck the Unix herd, its in graphics architecture. IMO, Canonical are trying to copy some of Jobs'/Apple's engineering decisions.

Re:Wait, what? (0)

Anonymous Coward | about 5 months ago | (#46416947)

My logic says that the toolkit that Chrome uses should be ported to have a Mir backend, rather than Chromium itself?

What toolkit? Chrome is a platform unto itself and has it's own libs for drawing and fonts.

Re:Wait, what? (1)

jones_supa (887896) | about 5 months ago | (#46417979)

Yeah, that was kind of my suspicion in the next sentence in my message.
Check for New Comments
Slashdot Account

Need an Account?

Forgot your password?

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>
Create a Slashdot Account

Loading...