GNOME: Staring Into the Abyss

New submitter zixxt writes "GTK+ Developer Benjamin Otte talks about the stagnation and decline of the Gnome Project. He describes how core developers are leaving GNOME development, how GNOME is understaffed, why GNOME is a Red Hat project and why GNOME is losing market and mind share. Is the Gnome project on its deathbed? Quoting: 'I first noticed this in 2005 when Jeff Waugh gave his 10×10 talk. Back then, the GNOME project had essentially achieved what it set out to do: a working Free desktop environment. Since then, nobody has managed to set new goals for the project. In fact, these days GNOME describes itself as a “community that makes great software”, which is as nondescript as you can get for software development. The biggest problem with having no goals is that you can’t measure yourself. Nobody can say if GNOME 3 is better or worse than GNOME 2. There is no recognized metric anywhere. This also leads to frustration in lots of places.'"

Comments

Reason? GNOME3

[jmorris42]

GNOME was a good thing until version 3. It changed everything. The warning signs were there for years before. The attitude of a few dictating what was 'best' for the users, even when the users were screaming NO! NO! NO! started with the GNOME2 rewrite. They finally listened to some of the loudest arguments that time and restored enough functionality that it could become the standard Free Desktop.

GNOME3 turned that stupid up to eleven though, putting it a whole different category. It is explicitly declared it OK if any/all existing users leave, a pure "my way to the highway" deal. It is pretty much accepted that it is unusable on a standard desktop with a mouse and this isn't debatable as an issue in need of repair. The only rational explanation is that somehow, someone in that project assumes they are going to get an OEM preload deal on tablets somewhere. But GNOME's hardware requirements are higher than Android so it won't be some low end creep into the market through the back door deal, it will have to be on somebody's mid to top end hardware. Maybe RedHat has struck the deal in secret already and we are all going to be in awe of their mad negotiating skills. But it isn't the way to bet.

Or perhaps they assume that Win8 will force everyone to accept touchscreens and everything running maximized... even on 27" displays... so they just want to be there first, like how Compiz was doing the Vista eye candy a year or so before Vista shipped. Doubtful. If Win8 doesn't quickly get a recognizable default desktop on desktop class hardware users will just insist on Win7. Everything doesn't benefit from a touchscreen, keyboards and mice still have a place and aren't likely to go the way of the dodo anytime soon.

Guess if the article is right about the number of active devs left it really doesn't matter anymore because there doesn't appear to be enough left to rewrite their way out the the GNOME Shell disaster. Several of the alternates have similar manpower except KDE which has much more. It was a good desktop, it will be missed.

I'll say it!

[00Monkey]

"Nobody can say if GNOME 3 is better or worse than GNOME 2."

GNOME 3 is *worse* than GNOME 2. By far. Plus more.

Re:Reason? GNOME3

[TheGoodNamesWereGone]

GNOME was a good thing until version 3. It changed everything.

Mod up. The purpose of a DE is to enable the user to get his work done as fast and as efficiently as possible. Not eye-candy bullshit. If you can imlement eye candy that doesn't hinder or get in the way, I'm all for that, but never forget: **Enable the user to get work done fast**

chasing the "dumb it down" crowd

[Anonymous Coward]

The big issue with many modern desktops including Gnome and Win8 is they are hell bent on chasing the "dumb it down! dumb it all down! moaarrr dumber!!" crowd. Ripping out power user functionality, removing configurability, and generally making it about as annoying to use for proficient users as possible.

There aren't many "real" desktops left. KDE is left. Some like it, some don't, but at least it hasn't dumbed itself down to placate the LCD who think computers shouldn't be any more complex than operating a toaster. Win7 is alright. Most of the others have gone off the deep end in their quest to satisfy people who need the most simplistic interface possible at the expense of power features and customization.

Re:chasing the "dumb it down" crowd

[Shikaku]

XFCE is still around and became my desktop after Gnome2 support was dropped on Arch Linux.

Linux Mint also has Cinnamon.

Re:chasing the "dumb it down" crowd

[causality]

The big issue with many modern desktops including Gnome and Win8 is they are hell bent on chasing the "dumb it down! dumb it all down! moaarrr dumber!!" crowd. Ripping out power user functionality, removing configurability, and generally making it about as annoying to use for proficient users as possible.

For some reason stagnation is expected in computing, even when this is rare elsewhere. I call them "permanant noobs".

When you got your first bicycle and used it with training wheels, no one expected that you would still use those training wheels years later. When you got your learner's permit, it was expected that this was a stepping stone you would use to ultimatley gain enough skill to get your own proper driver's license. No one actually expected that these early learning stages would or should be permanent.

The "dumb it down" mentality with computing is the assumption that the early learning stages should be sanctified and made permanent, that they are some kind of perfect ideal, that it's not reasonable to ever expect a user's skill to grow with time. Sure, some users have more aptitude than others, some learn faster than others, but the "dumb it down" idea throws all of that out and assumes no one should ever learn anything.

It's like anything else. It grows if you feed it. It shrinks if you starve it. The constant feeding of it in mainstream thought has led to users who can operate a computer for 5+ years and still know nothing more about it than when they started. They get frustrated at the same problems that frustrated them five years ago because they have not learned anything. They demand overly-simplified interfaces and balk at the slightest investment of learning (and even then, nothing major, just paying attention and picking up facts here and there with experience would make a big difference).

It's standard penny-wise, dollar-stupid thinking. It's saving a slight effort in the short term in order to screw oneself in the long term. An intermediate user with an interface that presents the available options in an intelligent way has a much better experience than any user with an interface built on the assumption that you're an idiot. But the concept of making an investment is alien to this mentality. It's by no means limited to computing. You see it in corporations all the time, where everything is all about this quarter's earnings even when this leads to long-term sustainability problems.

Re:chasing the "dumb it down" crowd

[bertok]

Well put.

I have a theory for this based on my observations of older computer users, especially those that started in the DOS era.

Back in those days, two things were significantly different to now:

1) Software came with printed manuals written in a "tutorial" style. These days, most software comes with electronic help files at best, usually written in a "reference" style with no theory or explanations.
2) A few very popular products at the time like Norton Disk Doctor had a radically different UI style that actually explained things, and this helped people learn as they went.

I remember my father reading through the Corel Draw manual end-to-end, and he ended up learning how to use it completely. He's not a graphic artist by any means, but I've seen him develop fantastically complex multi-layer vector art for use in embedding in documents back when DOS 6 was new. These days, I'm shocked when I see vector art in a Word document. It just doesn't happen because it's "too complex" for most users, even though vector drawing programs have gotten better and easier to use!

It's the second one that I'd like to see make a come-back the most. Norton at the time was a fanstastic product, because its author realized that everyone else was doing UI design wrong. Nobody has picked up on his insight, and everybody still does it wrong.

Ask yourself this: How many times have you seen a dialox box pop up on the screen demanding an immediate response to a scary question with no explanation? Things like:

This could damage your system! Are you sure? Yes or No?

Think about it for a second. How is the poor user expected to respond to this? What the fuck is "this"? What kind of "damage"? Should he press "yes"? Or "no"? Why? Why not? On what basis should he decide?

Practically all software is like this. Operating systems like Windows literally barrage users with prompts that are exactly like that, dozens of times a day. The prompts never give any useful information, even for Administrators, let alone a non-technical user. Users learn only to click "OK" to everything and pray. No understanding is gained.

For comparison, Norton Disk Doctor had full screen dialog boxes with paragraphs of text explaining things like:
- What triggered this message
- A detailed explanation of what the question means
- What will happen if you press 'yes'
- What will happen if you press 'no'
- The risk to your data for both cases

I saw users who were still at the stage where they could only type with one finger confidently making complex technical decisions because they were informed. The explanations thought them something, and they learned, and got better at using computers.

I haven't seen a product like that since, by any vendor. Coupled with the combination of manuals becoming a rarity, it's no surprise that users aren't learning anything.

Desktop Environment Fad is finally ending.

[Technomancer]

I never quite understood the attraction. What exactly is this whole GNOME or KDE package for? Granted, there are some decent programs that came with them, but do they really require all the extra baggage of 10 layers of crappy libraries with fancy names? Both KDE and GNOME are just pointless empire building exercises by bunch of people who want to force their way of computer interaction on everyone else.

The OS should do one thing, provide services to programs. On UI level that includes managing windows and provide some way of task switching. Widget library is nice too since it saves some time for programmers, but it doesn't really have to be part of OS. On Windows this functionality is pretty much built in to the point of being (almost) non replaceable. Thankfully on Unix one has a choice of window manager, task switchers/panels, widget libraries etc. The users should be able to mix and match them to fulfill their needs. Some distros like Ubuntu may make these choices for the users that do not care much what they use. Where do mega projects "we gonna takeover your computer and make you do things our way" like GNOME and KDE fit? Nowhere, and finally people realize that.

The only thing that can be done with these projects is to salvage any good apps they have created and make them into independent projects. There is less and less to salvage though because GNOME managed to create dumber and dumber versions of the same things (like image viewers or browsers or file managers etc).

For instance, why would anyone ever use web browsers that GNOME has created (is the latest one Epiphany or something?) when there is Firefox, Chrome or Seamonkey made by people who know what they are doing?

There are some nice projects like LXDE, and to lesser degree XFCE which are actually helpful, they put together bunch of tools, most of them optional, and give you quite a lot of choice. Although XFCE is getting fatter and fatter.

Re:Reason? GNOME3

[hazem]

The one thing I like on gnome 3 is pushing the mouse cursor up in the upper left and getting a choice of windows. But other than that, it makes things harder.

I tried for a while to find a way to have a CPU and Network monitor like you could have it docked on a panel in gnome 2 but finally gave up.

I also often use more than one terminal window, but when you click on the terminal icon in the apps list, it just takes you back to the terminal you already have open.

For vitual desktops, I personally prefer a fixed layout... email and web browser in upper left, work vitrual computer in lower left, etc. The ever-changing dynamic list doesn't work well for me.

The worst is that I can't get it to behave right with my laptop and external monitor. Laptops today come with shitty short screens, so when I work at home, I keep the lid closed and just use my external monitor. Gnome3 can't seem to grasp this and always assumes the laptop's monitor is the primary monitor, so I can't reach the widgets, menus, etc. Sure, I can muck with the display settings to fix it during a session, but I have to do it all over again if I reboot or need to open the lid for some reason.

For me, it just has an illogical way of doing things and completely breaks my work flow.

I've used a lot of linux variants over the years, but I don't really enjoy having to keep figuring out all the obscure ways to get it work right again... over and over.

Re:Reason? GNOME3

[aztektum]

You know, I felt the same way a year ago. I still wish it was more tweakable, but the extensions are helping. I am much faster at getting around my desktop in Gnome-Shell than I am with Gnome2. Reason being is I can do more with just the keyboard.

I was a big Gnome-Do user. That's pretty much built-in now. I don't have to touch my mouse to move around apps. Their Alt+Tab feature is pretty slick. It shows Chromium and Alt+~ moves through the multiple instances I have open (OK so I don't usually have more than one thanks to tabs, but as an example...)

It's a bigger resource hog, but I have 12GB of RAM in the box I run it on. It doesn't feel that polished, but I really have few serious problems.

What they should be doing is focusing on the extensions paradigm. Let people create extensions to turn it into whatever type of system they want. If you want a traditional taskbar, get an extension. Distros could apply whatever extensions they want to create varying types of "Gnome". That would give them some direction that they say the project has lost.

I finally mostly like Gnome 3.4

[amorsen]

Gnome 3.0 had me trying out various tiling window managers to get rid of the horrible Shell.

Gnome 3.2 came out and I went back to the Shell. I needed a ton of extensions to get a usable desktop.

Now, with 3.4, all I need to add is a direct shortcut to each desktop. Alas, the GUI offers me shortcuts only for the first four desktops, but at least it is possible to set shortcuts for all of them on the command line. I no longer have any extensions installed. Super + typing part of the application name is wonderful.

All in all, 3.4 is IMHO nicer than Gnome 2. The road to get there has been horrendous and it may have cost too many users and developers for Gnome to be viable in the future. I hope Gnome will survive, because it is the best desktop I have tried so far.

Re:Reason? GNOME3

[jmorris42]

Here is a hint. It is one of the strengths of Linux from a certain way of looking at it.

Attaching and detaching a display from a laptop is something no DE is ever going to make 'just work' for everyone. You use case might sometimes be just what the developer was thinking, others you will lose. On the 'other' platforms you just live with it, we have options. On my laptop the F7 key is silkscreened with display/panel in blue, meaning Fn+F7 is the approved way and what would work on the 'other' OS. So to make it easy to remember I bound CTRL-F7 to a script.

It examines the state of the dock and doesn't try to 'do the right thing' for anyone and everyone, it does exactly what [I] want for either state. With only a little more work (when I get a spare round tuit) I'll extend it to look at the VGA port and deal with the presence of a projector automagically. Yes I means I have to hit a hot key when the automatics do the wrong thing (almost every time) but it means I always get what I want and it beats filing bug reports that get closed WONTFIX when the distro goes out of support and just bitching about it being broken.

#!/bin/bash

BUSDOCK=/sys/bus/platform/devices/dock.0/
DOCKED=`cat ${BUSDOCK}/docked`

if [ "$DOCKED" == 1 ] ; then
        echo "Docked"
        xrandr --output HDMI2 --auto --mode 1024x768 --rotate normal --pos 0x0 --primary --output VGA1 --off
        sleep 2
        xrandr --output HDMI2 --auto --rotate normal --pos 0x0 --primary \
                          --output LVDS1 --auto --right-of HDMI2 --set "scaling mode" "Full aspect"
fi
if [ "$DOCKED" == 0 ] ; then
        echo "Undocked"
        xrandr --output LVDS1 --auto --mode 1024x768 --set "scaling mode" "Full aspect" \
                      --primary --output HDMI2 --off --output VGA1 --off
        sleep 2
        xrandr --output LVDS1 --auto --set "scaling mode" "Full aspect" --primary
fi

Re:Reason? GNOME3

[jmorris42]

What you seem to want really isn't posible. You can't just mix and match in one part of a major subsystem and you certainly can't mix Debian and Fedora packages beyond the very end user applications that have no connections into the different plumbing that seperates the two trees of development and that basically static link everything.

You seem enamoured with the Android package system without undertanding it. Android packages work because they are very restricted in what they can do. For example, they must be Java; that means they cannot alter any of the system level components. So replacing part of GNOME would be like replacing the native binary parts of Android, which an .apk can't attempt. They also work because there is only one Android line and it is carefully kept backward compatible. While Linux distros can upgrade from one major version to another entirely via the package system you could never upgrade from Android 2.2 to 2.3 via the Play Store. The OS components involved simply aren't part of the package manager on Android. The kernel on most devices isn't even in a file.

Every few weeks some kid shows up on a Linux forum demanding that we rebuild everything to support a binary only cross distro 'app' model. Usually with notes about how much more successful Windows or OS X is and attributing that success to this binary model. Not happening. The reason we have different distros is because they aren't all alike except for the package manager, each is trying new things. If a consensus emerges that one has really done something right the others of course adopt it but there is no central planner and we don't want one. Good luck convincing a Gentoo ricer to adopt binary packaging and a strict binary API. Systemd or sysV init? PulseAudio, ALSA, ESD, ARTS or OSS?

Feel free to create yet another distro and show us all how it should be done, that is of course where the existing ones came from. And maybe you will succeed in attracting a following and eventually some of your ideas will migrate.

Nope, we _can_ say.

[HaZardman27]

Nobody can say if GNOME 3 is better or worse than GNOME 2. There is no recognized metric anywhere

False. GNOME 2 was most certainly better than GNOME 3, and the metric used to measure quality is number of "what the fuck"s, "how the fuck"s, and "why the fuck"s per hour (lower is better).

Re:Reason? GNOME3

[M. Baranczak]

just like Torvalds did when he wrote git and replaced subversion

Git was never intended as a replacement for Subversion. From the way that Torvalds talked about Subversion, I doubt he ever even used it. Git was a replacement for BitKeeper - which worked on a distributed-repository model, just like Git.