IgnorantGuru's Blog

Linux software, news, and tips

A Crack In The Monolith

It would seem that Red Hat’s GNOME devs have had a meeting and a change of heart:

while we certainly hope that many users will find the new ways comfortable and refreshing after a short learning phase, we should not fault people who prefer the old way. After all, these features were a selling point of GNOME 2 for ten years!

Why, these people are just so darn heart-warming, aren’t they? I love it when they call alt-tab “the old way”. :) Not standard or even different or alternate, but “old”. Why do I feel that in his mind he’s making some minor concessions to senior citizens?

we’ve decided that we will compile a list of supported gnome-shell extensions. This will be a small list, focused on just bringing back some central ‘classic’ UX elements: classic alt tab, task bar, min/max buttons, main menu…

We haven’t made a final decision yet on how to let users turn on this ‘classic mode’ – it may be a switch in gnome-tweak-tool or something else.

As in, “oops, we forgot that we have users and they like to actually do stuff, so now we have to figure out how to hack flexibility into our rigidly designed system”. This is sure to be done well.

Yet the good news is they finally responded on this one issue in some form, at least in theory. Perhaps.

Earlier reading: GNOME (et al): Rotting In Threes

November 22, 2012 - Posted by | News


  1. As you previously pointed out in your Rotting In Threes post the Gnome developers don’t like extensions and break the APIs for them intentionally. At least it seems so.
    But their users are using extensions anyways, so this may be an attempt to get control over the most used extensions/most demanded features, not a concession to their users. Now they can provide working extensions with every new version, breaking the APIs for the non-official extensions, without loosing control of their holy brand.

    Either that, or at least one developer realized that they are loosing their users with their weird marketing thinking and that they actually have to do something

    Comment by TobiSGD | November 22, 2012

    • That’s an interesting thought. Whatever their motivation, they don’t seem very excited about it.

      > Either that, or at least one developer realized that they are loosing their users with their weird marketing thinking and that they actually have to do something

      I don’t think they take the desktop market seriously. They have visions of tablets bringing them billions. And tablets use these simpler interfaces.

      Funny how they spell UI “UX”. It’s not an interface, it’s a media presentation. You’re merely an observer and consumer of it. All you require is an X-Box controller, and that’s being generous. And now they’re adding a few cheats – yes, beg us for just a little power to control your own computer. Maybe we’ll give you a small list of extensions. ‘Classic Mode’ – for people who still remember that computers can do *stuff*, not just play media.

      Comment by IgnorantGuru | November 22, 2012

    • There mostly doing this because they are getting rid of Fallback mode in 3.8, and are afraid they will lose users and volunteers who have been using Fallback mode, not because of graphics issues, but because they like the fact that it preserves some gnome 2 features. Of course forcing everyone to use Composting graphics is also silly. I think they faced a second wave of backlash when they announced no more Fallback mode.

      Comment by pataphysician | November 24, 2012

  2. Don’t expect too much: “We still believe that there should be a single, well-defined UX for
    GNOME 3”.

    I’m speechless to see how breaking software seems to be normal for Gnome devs: “The downsides of extensions are that (a) there is no guarantee that they will work with a new shell release – you often have to wait for your favourite extension to be ported”.

    This is not a downside of extensions, this is because Gnome devs don’t bother to break APIs, and they write it in a so casual way…

    About the classic mode switch: “We haven’t made a final decision yet on how to let users turn on this ‘classic mode’ – it may be a switch in gnome-tweak-tool or something else.”

    Note that gnome-tweak-tool is a third-party software, not installed by default…

    Anyway, I would not lose my time to write extensions for Gnome Shell. I suppose that it’s as “pleasant” as to develop GTK3 themes when you see your work to be broken every 6 months.

    And do I really want to use a Shell forcing me to install extensions just to do so basic actions like moving the clock? I don’t think so. Just for fun, I searched about that and I found 2 extensions: “Frippery Move Clock”, not compatible with Gnome 3.6, and “Move-Clock”, a fork of “Frippery Move Clock” to make it Gnome 3.6 compatible. Both only move the clock to left of status menu button, so I can’t even move the clock exactly where I want (without learning to write extensions and wasting time to update them).

    Comment by jpfleury | November 22, 2012

    • Tweak Tool is installed by default in openSUSE, Ubuntu GNOME Remix, and Debian Wheezy.

      Comment by Jeremy Bicha | November 25, 2012

  3. This seems like a weak move so they can claim “But we DID listen to user feedback! See? We have these old features hidden away in a secluded part of GNOME!” But I’ve been in a foul mood for the past day, so it could just be me.

    It’d be nice if they turned around and created a mode that _is_ stable, but spartan, and left that shell to extension writers, and only update it (e.g. break the API) when absolutely necessary. However, this alternate mode won’t have the “polish” of GNOME 3’s “brand”, so it’s up to theme and extension writers to make it good.

    It could be called “Unsafe Mode” or “unsupported mode” or some other negative language (because they’ll regard it negatively anyway), and it’d give application developers a place to make things for GNOME without the development cost (e.g. time and effort) skyrocketing.

    But if something like that’s going to be made, why not just use a different DE altogether? That’s what would be created.

    Comment by sporkbox | November 23, 2012

  4. It comes down to this; they don’t make money out of gnome, so they don’t worry about losing money. And therefor do what ever they wish dispite users’ objections. We must be thankful that they consider bringing back lost features.

    quote “Well the classic version in this case is a set of GNOME Shell extensions, while Cinnamon is a GNOME Shell replacement. That said maybe these extensions address enough of peoples complaints to make MATE and Cinnamon mostly redundant.”

    So why the bloody hell cause so much disturbance? Why don’t they use Cinnamon instead? It would be less work for them since Mint is developing it! .

    Comment by Σταύρος Δαλιακόπουλος | November 23, 2012

  5. Gnome is dead for me. I do not and will not ever trust them, even if some day they truely try to abandon all this disgusting nonsense. And have ditched all the GTK apps even the brilliant SpaceFM and use openbox + razor-panel. Will migrate to Gentoo when time permits.

    Comment by tablethater | November 23, 2012

  6. “Old way”, or “old school”:

    William Jon McCann [reporter] [developer] wrote: We have a nice floating status bar now so the old school status bar isn’t needed.

    Remove the statusbar (nautilus)

    Comment by jpfleury | November 28, 2012

    • I think that for Gnome devs, the “new” way is “touch”. See these reports about removing Nautilus features:

      William Jon McCann [reporter] [developer] wrote: The combination of panes and tabs is just too much. It is inconsistent with the file chooser and doesn’t work well with touch. — Remove extra panes

      William Jon McCann [reporter] [developer] wrote: It [tree side bar] is inconsistent with the file chooser, doesn’t work well with touch, is really hard to use, and isn’t consistent with any other GNOME 3 apps. — Remove tree side bar

      See also the last GNOME Users And Developers European Conference (GUADEC) in July 2012:

      The GNOME OS plan calls for the desktop to be incorporated in the project’s own Linux distribution, including a standalone installer and update mechanism. The first steps in this direction have already been taken by adding offline update capabilities to the desktop. This goes along with a plan to have GNOME OS distributed on specific hardware, a road that could lead to an additional revenue stream aside from the GNOME Foundation’s donation income, according to López and Sánchez.

      The developers say that the GNOME project needs to shed its focus on the traditional desktop and should broaden its reach to the mobile market. To facilitate this, GNOME 4 would include a fully touch-enabled user experience for both tablets and phones. — GNOME developers set ambitious goals at GUADEC

      Comment by jpfleury | November 29, 2012

    • About removing Nautilus features, reasons given by devs are funny. Examples:

      William Jon McCann [reporter] [developer] wrote: Remove the open folders in new windows option

      This is not something we want to encourage as it basically
      breaks the browser model.

      Remove always-use-browser option

      This isn’t necessary since the removal of spatial mode. Having
      it as a hidden preference is counterproductive because it may
      have been set before and it will silently break the behavior of
      the current version. — Remove the option to open all folders in new windows

      William Jon McCann [reporter] [developer] wrote: The compact mode does horrible things like horizontal scrolling and it really isn’t significantly different from icon mode with labels on the side. — Remove compact mode

      Comment by jpfleury | November 29, 2012

      • The real solution to such GNOMEism don’t use and care GNOME at all.

        Nautilus alternatives include SpaceFM, Caja, Thunar, PCManFM, Nemo, …
        All of the alternatives may have glitches and problems, but they don’t have hostile view on features. Helping one of them will make your and other people’s life easier.

        Comment by Ma Xiaojun | November 29, 2012

  7. I don’t think the classic mode will be very good. I guess it would be something looks like GNOME 2 but have very limited configurability.

    Even if the classic mode is decent, then what about the applications?
    Since GNOME’s design vision is clearly tablet.
    I guess most people would definitely lose their favourite features at some point.
    That kind of losing can never be solved by any extensions.
    I guess most application is still familiar because GNOME has limited man power. Ground breaking design idea cannot be implemented overnight.

    IBus is a framework for developing CJK and other languages’ input methods. It gives engine developer unlimited freedom since 2009 (though API is changed from time to time). Now IBus is integrated with g-c-c. Guess what? GNOME devs uses two white lists that completely destroyed the openness of IBus framework.
    The experience in above debate clearly show that GNOME developers want exclusive control. Even if they never bothered to input a single paragraph of Chinese. They still feel very bad by the fact that users and distributions can decide what input method engines to show and use.

    My hope for Linux enthusiastic people is that do not even give a shit to current GNOME. Even if you can write / use extensions to save your days, will you introduce GNOME to other people? Can you deploy GNOME in large scale? I guess MATE or XFCE is the working alternatives. GNOME does have many applications, but most of them have DE agnostic alternative, e.g., Empathy => Pidgin. The problem for MATE is that it is hard to install. It doesn’t have spin distributions like that of XFCE and LXDE. If you are already insiders (developer, trusted users, sponsors, whatever) of some major distributions, can you offer some help on MATE inclusion?
    Once we have a big, working MATE or XFCE community, we never longer need to bothered by any GNOME change, let them play their toys. And we can really promote MATE or XFCE.
    ( I haven’t used LXDE much, sorry)
    If GTK3 really sucks in theme or other stuff, just try to avoid it. I don’t think GTK2 will ever die, as long as X11 protocol is still supported.
    I really admire extension people given the fact that how little documentation about GNOME Shell extension is available. However, I wonder why wasting time on such short-term and non-elegant solution? Having a big working community on a sane desktop environment is a much better option.

    We can probably look back to GNOME when it releases 4.0, as we may also interested how Windows really looks like.

    Comment by Ma Xiaojun | November 29, 2012

  8. Interesting links, Ma Xiaojun. Limiting input methods is a touchy subject, and one that’s not totally been solved yet. I currently use SCIM with Anthy support. SCIM’s not actively developed anymore, but *it works*, and that’s what makes the difference for me. It’s depressing to see GNOME folks aiming for the latest trends in UI design instead of focusing on usability and flexibility. Those two things are what make a UI truly usable, instead of catering to the lowest common denominator and forcing power users to use a suboptimal workflow.

    I await the day when this web 2.0 UI design trend passes and we can get back to creating interfaces that grow with the user. Vim does a fantastic job of that, and yet most people would say it has a horrible interface because the garden variety idiot who doesn’t read manpages or consult the ‘vimtutor’ app can’t figure it out. An interface should be judged according to how well it performs complex tasks, not by how easily an idiot can navigate it.

    Just my two cents.

    Comment by sporkbox | November 29, 2012

    • This reminds me of some thoughts I had reading Linus’s request for help with configuring gnome 3 on Fedora. Are these interfaces really ‘easy to use’ when it’s so difficult to make any adjustments to them? Making an interface easy by dumbing it down and removing the power is ridiculous. Yes, interfaces should be as simple and easy to use as possible, while still performing the required functions. That is where GNOME 3 appears to fail. It’s not really easy at all – it makes it far more difficult to do what you want.

      Comment by IgnorantGuru | November 29, 2012

  9. Just updated to Mint 14.1 using MATE as I really liked Gnome 2, MATE 1.4 runs superb, the guys behind MATE are even patching old Gnome 2 bugs like the toggle button in the location bar.

    In another computer I run Xubuntu 12.04 with XFCE 4.10, it is absolutely excellent too.

    No, we do not need GNOME for anything now, those Projects will keep getting better.

    There is also Enlightenment E17 about to be released.

    GNOME will be more and more ostracised as time goes by, I’m really curious about what MATE will do for 1.5/1.6.

    Unity will have to improve a lot or will share Gnome 3’s fate.

    The future is bright, what Gnome did is akin to the situation with XFree86, except that Gnome can be much more destructive to the Linux ecosystem.

    Gnome earned hard a dead by a thousand forks. Their own Linux distribution, Interface for phones and tablets… I would laugh if it wasn’t so incredibly absurd and out of touch with reality.

    Comment by Sanders | December 1, 2012

  10. More wanton GTK2-3 breakage, this time in Glade, with debian hostile to backwards compatibility – surely this will affect MATE and other GTK2 projects.

    Comment by IgnorantGuru | December 3, 2012

    • Very “good” atmosphere:

      Thanks for attaching more garbage to this bug report, but we know where
      to find such data, thanks.

      BTW, what is not clear in tagging the bug “wontfix”?

      It makes you want to participate…

      It was tagged WONTFIX 7 minutes after the report was filled. After that, users are assumed to shut up?

      Comment by jpfleury | December 3, 2012

Sorry, the comment form is closed at this time.