Ubuntu 11.10 To Switch From GDM To LightDM(digitizor.com)
digitizor.com
Ubuntu 11.10 To Switch From GDM To LightDM
http://digitizor.com/2011/05/12/ubuntu-11-10-lightdm/
14 comments
It looks as if Ubuntu is making good progress on the user interface, despite the current apparent loss of functionality with Unity. Having well designed and user friendly features is only one part of the equation though. To really gain popularity they will need to be able to persuade hardware manufacturers to ship with Ubuntu installed by default, or at least as a dual boot option. So far this has proven difficult to achieve, and it's hard to buy such a system via any common distribution channel, such as supermarkets.
Part of this has to do with the cost of Ubuntu. I'm not talking paying developers to make sure it works on the machine, but the actual cost that Canonical charges for the ability to use the Ubuntu logo and becoming "certified"; It's not a trivial amount of money by any stretch.
Anyone can preinstall Ubuntu on a machine and release it, they just can't say it is Ubuntu specifically or use the logos.
And I don't advocate not paying it or not becoming certified, we are considering it currently but I can say that management has balked a few times when we've brought it up has been due to the amount of money required up front.
Anyone can preinstall Ubuntu on a machine and release it, they just can't say it is Ubuntu specifically or use the logos.
And I don't advocate not paying it or not becoming certified, we are considering it currently but I can say that management has balked a few times when we've brought it up has been due to the amount of money required up front.
Who are considering becoming certified?
The company I work for. We are a fairly small ODM, and while I don't want to say our name I'm not hard to find on the Internet, and I know I bragged about getting my job (I get paid to work on Linux, still can't believe my luck)
I imagine the concern on Canonical's part is that they will field a significant part of the support flood from new users. Of course, the OEM/ ODM will probably get the bulk of support traffic (with some falling on the retailer), the "pilot" from the netbook/ubuntu probably gave them insight. If they structure the deal correctly, and have an auditable way for the ODM and Canonical to determine the support costs/ unit sold that could be a way to get the deal done. Tell those business coworkers of yours to keep iterating on the deal they are presenting Canonical. I can't believe there's not an equitable way to get that deal done.
Doesn't Microsoft pay hardware manufacturers to ship computers with Windows? I thought I heard of this happening with Dell. That seems like a major impediment to Ubuntu's shipping pre-installed/by default and its more widespread adoption.
No, the (just-expired) antitrust settlement would not allow Microsoft to make Windows free or "less than free". However, PC vendors ship crapware that pays more than the Windows license cost.
I love the Ubuntu team's willingness to move the platform forward. Even if that means making controversial decisions.
I'm not quite sure how changing the login manager counts as a controversial decision unless you were meaning just the unity issue. The login manager has one main job and really shouldn't be a huge convoluted mess of code (from experience GDM is). I'm excited about this change although I've never heard of lightdm until this story.
Reading through the other comments on this post tells me that there is definitely some controversy over this.
I'm wondering if this means I'll have to jump through hoops again to not have everybody know what usernames are available on my computer.
Now would be a great time to file that feature request. LightDM folks have about 6 months to get it fixed (that is, if it's broken)
Does that mean it doesn't support hiding usernames yet? I'm still trying to figure that out.
Did you try setting load-users=false under [UserManager] in /etc/lightdm.conf?
I didn't install anything yet. I just googled for it to find out if it will be possible, since that was a rather large hassle during the last switch, and just recently got fixed and is now available as a checkbox option in the preferences.
I found lots of articles and developer resources, but haven't found the documentation for it on the web yet. I assume that will change as exposure increases.
I found lots of articles and developer resources, but haven't found the documentation for it on the web yet. I assume that will change as exposure increases.
You can check. I installed lightdm in a couple minutes. Auto-login didn't work (it was on a shared desktop), there was some terminal keyboard weirdness and I reverted the change. Running "dpkg-reconfigure lightdm" allowed me to easily revert back to gdm.
I think this is the path that Ubuntu should follow, replace current pieces of software with lighter and smaller solutions.
You can always replace them with more advances packages through apt-get.
You can always replace them with more advances packages through apt-get.
Unity doesn't strikes me as particularly light ...
I am not sure pissing off core GNOME developers with its ever changing new directions is going to help them in the long run.
On the other hand I might be proved wrong if while increase its user base Ubuntu manage to convert some of them in core durable developer.
IMHO, then it's not GNOME anymore. It's a new desktop akin to XFCE as they diverge more and more. It was foreseeable, especially in light of previous two year that's what was going to happen.
Time will tell.
I am not sure pissing off core GNOME developers with its ever changing new directions is going to help them in the long run.
On the other hand I might be proved wrong if while increase its user base Ubuntu manage to convert some of them in core durable developer.
IMHO, then it's not GNOME anymore. It's a new desktop akin to XFCE as they diverge more and more. It was foreseeable, especially in light of previous two year that's what was going to happen.
Time will tell.
I'm glad they are deviating from core GNOME as they seem to have more focused art direction and better usability in general, for me Ubuntu is a much smoother and more pleasant Linux experience than any other distro for this reason.
Good for them, as long as they don't FUD on GNOME.
Better usability ? I don't think so, I remember being orce fed the first version of message indactor with Gwibber, Pidgin and Evolution were 3 entries of the same menu behaving totally differently and unconsistently.
Unity feels quite raw and bulky. I don't think my dad will search it's application by their exact names. INMHO, the Gnome-Shell menu is much better in that it let you see you apps a people familiar with any recent smartphone won't by lost. I can't say the same for Unity main menu.
Better usability ? I don't think so, I remember being orce fed the first version of message indactor with Gwibber, Pidgin and Evolution were 3 entries of the same menu behaving totally differently and unconsistently.
Unity feels quite raw and bulky. I don't think my dad will search it's application by their exact names. INMHO, the Gnome-Shell menu is much better in that it let you see you apps a people familiar with any recent smartphone won't by lost. I can't say the same for Unity main menu.
Quite frankly, who cares? I'm tired of the GNOME release plans: remove features, make GTK+ controls bulkier than ever. It's like they're trying to ensure that the same amount of stuff fits on a screen despite increased screen resolution.
I'm happy to see the role Compiz is taking on. A recent rewrite that cleaned things up, replacing more and more window management tasks, support for non-accelerated backends. I'd love to see Compiz's wallpaper plugin mature and replace nautilus for the task. (It'd also allow for the global menu to not be awkward when the wallpaper has focus and for multiple wallpapers on different workspaces). Though frankly I'd love to see nautilus die in a fire anyway.
I'm happy to see the role Compiz is taking on. A recent rewrite that cleaned things up, replacing more and more window management tasks, support for non-accelerated backends. I'd love to see Compiz's wallpaper plugin mature and replace nautilus for the task. (It'd also allow for the global menu to not be awkward when the wallpaper has focus and for multiple wallpapers on different workspaces). Though frankly I'd love to see nautilus die in a fire anyway.
I don't care about them drifting away. I might even understand some point they made about GNOME. What I don't appreciate it's some people in Ubuntu world (that would be community and canonical) blaming GNOME Upstream for the drift. They want to lead their own platform somewhere else, fine, just do it. Don't pretend that you've been pushed to.
The Gnome Zeitgeist project originated from the Ubuntu community and at first they weren't accepted for inclusion in the project and started to whine in a very ubuntuesque way. They weren't listening at all to the maintainer's rationale for rejection. Happily at some point they started to listen and will probably be a key piece of GNOME 3.2
Now, I can't say the same for some others projects from canonical. challenging upstream is one thing, not being open to dialogue another, pushing its agenda by FUDing another beast altogether.
As for removing features ... what are talking about ? GNOME 3.0 ? It's not like it is a finished and dead product that won't evolve anymore ! Bulkier controls ?! Come man, just change your theme. Increased screen resolution ? Like Unity wasn't made for netbooks and tablet. I'll stop feeding the troll right here ;-)
The Gnome Zeitgeist project originated from the Ubuntu community and at first they weren't accepted for inclusion in the project and started to whine in a very ubuntuesque way. They weren't listening at all to the maintainer's rationale for rejection. Happily at some point they started to listen and will probably be a key piece of GNOME 3.2
Now, I can't say the same for some others projects from canonical. challenging upstream is one thing, not being open to dialogue another, pushing its agenda by FUDing another beast altogether.
As for removing features ... what are talking about ? GNOME 3.0 ? It's not like it is a finished and dead product that won't evolve anymore ! Bulkier controls ?! Come man, just change your theme. Increased screen resolution ? Like Unity wasn't made for netbooks and tablet. I'll stop feeding the troll right here ;-)
What? Gnome 2.3 has been shedding features for sometime. Ability to change GDM theme? Gone. Ability to configure gnome screensavers? Gone. That's not even mentioning 3.0, but I understand that it's like KDE, in that 3.1 will be much better suited for full time use.
You miss the point on the controls. Unity WAS made for netbooks and tablets. It conserves UI space in a more concise manner (not to mention that I enjoy the search/icon paradigm. It's what Windows 7 did with dock/search-start-menu. OS X has the Dock and spotlight. The jury is in, it's popular)
Most of the themes I've seen for GNOME3 still look bulky. Maybe I haven't seen enough themes yet.
You miss the point on the controls. Unity WAS made for netbooks and tablets. It conserves UI space in a more concise manner (not to mention that I enjoy the search/icon paradigm. It's what Windows 7 did with dock/search-start-menu. OS X has the Dock and spotlight. The jury is in, it's popular)
Most of the themes I've seen for GNOME3 still look bulky. Maybe I haven't seen enough themes yet.
You're right about gdm, still I didn't noticed it until you told me. I think most people just change the wallpaper and at most apply a different control theme.
Well I don't really get you point on the controls and screen estate then. To me "bulkier" controls precisely ease their use on netbooks and tablets.
Well the search is a great feature if it is fuzzy and intelligent enough to match on related keyword. Now, for now, it doesn't quite work that way. Regular people don't dare searching because they might not even know what to search. So to me gnome-shell as a better compromise it does not force search on you. With Unity, if you don't want to search, you need extra clicks to get to the actual application list.
There are currently 9 Gtk+3 themes on gnome-look.org that's hardly enough to leverage the possibilities of the theming engine. Better theme will come and will harness the power of the new the engine.
Well I don't really get you point on the controls and screen estate then. To me "bulkier" controls precisely ease their use on netbooks and tablets.
Well the search is a great feature if it is fuzzy and intelligent enough to match on related keyword. Now, for now, it doesn't quite work that way. Regular people don't dare searching because they might not even know what to search. So to me gnome-shell as a better compromise it does not force search on you. With Unity, if you don't want to search, you need extra clicks to get to the actual application list.
There are currently 9 Gtk+3 themes on gnome-look.org that's hardly enough to leverage the possibilities of the theming engine. Better theme will come and will harness the power of the new the engine.
> I think this is the path that Ubuntu should follow, replace current pieces of software with lighter and smaller solutions.
I fully agree, but am not sure this is an example of that. The article says LightDM uses WebKit. Loading an entire html rendering engine for a login manager seems like overkill.
I guess this explains why LightDM is smaller than GDM in code size - but is it smaller in total size when including all necessary libraries?
Now, some of this might be offset if you have WebKit loaded anyhow as a shared library. This seems unlikely, though,
1. This is the login manager - it's before you load your desktop and web browser. (It might save loading it later though.) 2. The default browser in Ubuntu, Firefox, is not WebKit based. Neither is Opera for that matter (but it is not as popular). 3. The most popular WebKit based browser, Chrome, bundles its own version of WebKit - so there would be no sharing anyhow.
I fully agree, but am not sure this is an example of that. The article says LightDM uses WebKit. Loading an entire html rendering engine for a login manager seems like overkill.
I guess this explains why LightDM is smaller than GDM in code size - but is it smaller in total size when including all necessary libraries?
Now, some of this might be offset if you have WebKit loaded anyhow as a shared library. This seems unlikely, though,
1. This is the login manager - it's before you load your desktop and web browser. (It might save loading it later though.) 2. The default browser in Ubuntu, Firefox, is not WebKit based. Neither is Opera for that matter (but it is not as popular). 3. The most popular WebKit based browser, Chrome, bundles its own version of WebKit - so there would be no sharing anyhow.
Empathy (the default IM client) uses webkit to render it's themes. It is quite likely to be a shared library. On natty, webkit is fairly hefty at 24MB but it doesn't take Empathy long to load at all. And by the time your login manager has started /usr is already mounted, assuming you are not using their default partitioning which is one huge partition.
It would be interesting to know how many users use Empathy. I don't myself, and maybe I'm not representative, but I don't think it is that many - Pidgin seems more popular among people I know.
I don't follow the /usr bit - what does mounting /usr have to do with loading or not loading WebKit?
WebKit is around 24MB, which is as you say quite hefty, but it's even worse as it depends on some other libraries - sqlite, libpng, etc.
I don't follow the /usr bit - what does mounting /usr have to do with loading or not loading WebKit?
WebKit is around 24MB, which is as you say quite hefty, but it's even worse as it depends on some other libraries - sqlite, libpng, etc.
Firefox and Chrome are going to be on more equal footing in Ubuntu now, so webkit is more likely to be needed. Also, how large is webkit compared to Gnome? Maybe they want to get the login on-screen faster?
> Firefox and Chrome are going to be on more equal footing in Ubuntu now, so webkit is more likely to be needed.
Chrome bundles a completely separate copy of WebKit. There would be no savings here, it would be exactly as if Chrome wasn't there for this matter.
Chrome bundles a completely separate copy of WebKit. There would be no savings here, it would be exactly as if Chrome wasn't there for this matter.
My bad! Thanks for the correction :)
The blueprint (how features are proposed for ubuntu) is at https://blueprints.launchpad.net/ubuntu/+spec/desktop-o-ligh...
Does LightDM have all the support GDM does for non-user/password login capabilities, warning banners, etc?
PAM is the standard and it doesn't get more secure, flexible and expandable than PAM. LightDM is PAM compliant so I am not expecting any trouble.
PAM is the standard, yes, but the DM has to support different PAM features (such as smart-card login).
I think idea is not to have the most feature-full offering in existence but to have something that does the job for most people and let others change if needed to something more power-full.
Ubuntu choosing a different default display manager is hardly newsworthy, would you seriously discuss Windows or OS X swapping some code behind the login window?
I would be more interested in hearing about Canonical putting lots of thought and their weight behind an improved desktop experience developed independently of the current looks and functionality of Windows, especially since for the moment they are the GNU-Linux-on-the-desktop company.
I would be more interested in hearing about Canonical putting lots of thought and their weight behind an improved desktop experience developed independently of the current looks and functionality of Windows, especially since for the moment they are the GNU-Linux-on-the-desktop company.
It's newsworthy when that code is part of the open source community, there are political implications of the switch, and Ubuntu is part of a larger effort to get more people to use GNU/Linux.
It's fantastic that Ubuntu is pushing to improve stuff, but it worries me the amount of old stable code they are ripping out.
GDM is not a good example of "old stable code" that's being ripped out on the initiative of Ubuntu, since it has recently seen a rewrite upstream, among the reasons for which was bit rot.
I hated it when Ubuntu switched from Gnome to Unity by default, but I like this move away from GDM. If there's one thing in Ubuntu that's less customizable (in terms of look-and-feel) than Unity, it's GDM.
While it looks like Ubuntu is throwing caution in the wind in many respects, I like that this is bringing attention to overlooked projects and attempting to clean up some old messes.
Matthew Garrett opinion on the matter : http://www.advogato.org/person/mjg59/diary/296.html
I like mingetty http://ubuntuforums.org/archive/index.php/t-546544.html
Surely GDM will still be available as a package though...? If so, does the choice of default login manager really matter?
Defaults are what 99% of people use, luckily Ubuntu allows you to change the defaults as easily as "apt-get install gdm".
Ubuntu is all about the ease of use. How do you explain to your non-technical wife, for example, what a login manager is, what it does and why you'd want a better, lighter one? I like this change. I wish they'd switch away from Gnome entirely and go with something like XFCE by default. I want my DE to take up as little resources as possible out of the box.
Unfortunately, there are too many standard programs out there that need Gnome to run. I can't see Ubuntu using XFCE but still requiring Gnome installed; for newbies that would not be worth the effort. XFCE has a steeper learning curve than Gnome for people coming from Windows.
does this still apply with gnome3? it does not seem very close to a windows experience to me.
Actually the difference in performance/resources between a very light WM and gnome is not that big nowadays.
On my 4 years old netbook I run compiz stand alone as WM and I am not getting any slower than xfwm or openbox
On my 4 years old netbook I run compiz stand alone as WM and I am not getting any slower than xfwm or openbox