Jansson: On the Graying of GNOME [LWN.net]

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 17:21 UTC (Wed) by scientes (guest, #83068) [Link]

Nice to see more of this on different projects, in addition to the yearly analysis of Linux contributions!

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 17:38 UTC (Wed) by pbonzini (subscriber, #60935) [Link] (20 responses)

It seems like there was a peak of activity around 2009-2011 of both corporate and drive-by contributors. So the question is not just what happened in 2012 to cause a reduction of contributors, but also what happened in 2009 to cause such a sudden increase.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 17:41 UTC (Wed) by jazzy (subscriber, #132608) [Link] (15 responses)

Could it be the mobile revolution that moved many people over from desktops to phones and tablets for their daily work?

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 18:14 UTC (Wed) by alspnost (guest, #2763) [Link] (13 responses)

Do people really try to use phones for "daily work"? I'd rather use pen and paper, or a stone tablet, than trying do anything like "work" on a phone.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 18:38 UTC (Wed) by ttuttle (subscriber, #51118) [Link]

For me at least: Extended work? No. Brief chat and email messages to coordinate extended work? Absolutely.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 23:21 UTC (Wed) by NYKevin (subscriber, #129325) [Link] (11 responses)

Quite a few people cannot afford laptops, and actually do use their smartphones for everything.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 13:09 UTC (Thu) by LtWorf (subscriber, #124958) [Link] (10 responses)

Introduce me to 1 person that actually works a full time job sitting on a desk with a phone in their hand…

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 13:32 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link] (3 responses)

> Introduce me to 1 person that actually works a full time job sitting on a desk with a phone in their hand…

Lots of sales people do that including a category of people who travel frequently for work aka road warriors and they make heavy user of their phones. Is it really that hard to imagine that some percentage of people rely on their phones or an occasional tablet lot more than their computers?

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 6:51 UTC (Fri) by LtWorf (subscriber, #124958) [Link] (2 responses)

That is shifting the goal.

It has to be someone who can't afford to own a computer for their work, but do work with a different device out of poverty.

Jansson: On the Graying of GNOME

Posted Dec 19, 2020 17:38 UTC (Sat) by gnu_lorien (subscriber, #44036) [Link] (1 responses)

Lots of poor people have to choose their nice, expensive things carefully. In the modern world, if one could only afford either a computer or a phone the choice is almost always going to be phone.

In my personal experience there's also a huge demographic of users that do everything on their phones: Kids. Things that don't work on their phones basically don't exist to them technologically. If we're talking about not having a new generation of programmers this is a critical demographic and use case to consider.

Jansson: On the Graying of GNOME

Posted Aug 5, 2021 16:03 UTC (Thu) by LtWorf (subscriber, #124958) [Link]

Kids use the computer as soon as they want to play a decent game or have to write something for school. Because nobody finds touch screens more convenient for writing.

Jansson: On the Graying of GNOME

Posted Dec 21, 2020 17:25 UTC (Mon) by NYKevin (subscriber, #129325) [Link] (4 responses)

I find it interesting that you immediately jump to white collar workers. Most white collar workers probably can afford a laptop, although it might be of poor quality. But there are plenty of people who do not work white collar jobs, and many of them have significantly lower incomes.

Jansson: On the Graying of GNOME

Posted Dec 21, 2020 19:52 UTC (Mon) by zdzichu (subscriber, #17118) [Link] (2 responses)

We are talking about job; it doesn't matter if one can afford laptop, as it needs to be provided by employer.

Jansson: On the Graying of GNOME

Posted Dec 22, 2020 0:23 UTC (Tue) by Wol (subscriber, #4433) [Link] (1 responses)

And if you're self-employed ... ?

Cheers,
Wol

Jansson: On the Graying of GNOME

Posted Dec 22, 2020 6:27 UTC (Tue) by zdzichu (subscriber, #17118) [Link]

Deciding to be self-employed when you can't afford the tools? Madness.
Anyway please answer original question in your context: what kind of self-employment it is when you sit whole day at a desk with a phone only?

Jansson: On the Graying of GNOME

Posted Dec 22, 2020 0:03 UTC (Tue) by LtWorf (subscriber, #124958) [Link]

> I find it interesting that you immediately jump to white collar workers

Could that be related to the fact that those are the ones that do use a computer for work, as we were discussing on this thread?

No of course it must be because I'm extremely prejudiced and closed minded -_-'

Jansson: On the Graying of GNOME

Posted Dec 28, 2020 17:05 UTC (Mon) by MarkVandenBorre (subscriber, #26071) [Link]

Doesn't qualify entirely, but my car mechanic uses only his smartphone for his administration. His garage doesn't have an office space at all.

He orders his spare parts from the thing. He emails customers and suppliers from the thing. It seems to work for him, better than a laptop would.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 20:25 UTC (Thu) by federico (guest, #109225) [Link]

That could be true for some Windows or Mac users, but not for Linux/BSD users. Having said this, it could be that companies moved their developers on more profitable businesses: cloud and friends. Could it be that the 2010 spike is in part due to MeeGo (https://en.wikipedia.org/wiki/MeeGo)? See Intel contributions and MeeGo's lifespan (and moblin)

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 17:54 UTC (Wed) by pizza (subscriber, #46) [Link]

The pre-GNOME 3.0 peak was due to the "all-hands-on-deck" approach to get the underlying libraries, desktop, and basic applications "done" for the 3.0 release.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 2:44 UTC (Thu) by liam (guest, #84133) [Link]

Unity, followed by Ubuntu's subsequent sharp split from the GNOME community?

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 12:02 UTC (Thu) by halla (subscriber, #14185) [Link]

The reason is pretty simple: Nokia. Nokia tried to create an ecosystem around Linux and free software, which drove a lot of development, both around GTK/GNOME and Qt/KDE.

Recently someone made a similar graph for KDE repositories and it shows exactly the same pattern.

At one point there were more than twenty people paid to work on Calligra/KOffice, just for the document viewer app for Maemo. Nokia-internal people, people from my company, KO GmbH, third party developers. And that's just one example.

And then in 2012 Nokia switched to Windows and all that work stopped.

Jansson: On the Graying of GNOME

Posted Dec 19, 2020 20:20 UTC (Sat) by Trammael (guest, #101173) [Link]

Part of it is attributed by the author to the switch to Git and more direct attribution; with CVS and SVN contributions may have been committed and attributed to someone else.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 17:42 UTC (Wed) by ngraham (guest, #142633) [Link] (10 responses)

I can't find the link right now, but I'm pretty sure I've seen a similar graph showing a peak for KDE at about that time too, following by a reduction over time. For KDE, the numbers have been growing again since about 2018 IIRC. So I'm wondering what happened around 2010-20128 to cause this mass-die off of FOSS activity.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 17:55 UTC (Wed) by pizza (subscriber, #46) [Link]

> So I'm wondering what happened around 2010-20128 to cause this mass-die off of FOSS activity.

The meteoric rise of iOS, Android, and mobile-centric computing?

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 17:58 UTC (Wed) by atnot (guest, #124910) [Link]

I don't think it's fair to call it that. It's a decline of desktop FOSS activity. I wouldn't be surprised to see an corresponding growth in activity in web and server technologies as more and more tasks shifted to browsers.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 0:39 UTC (Thu) by rodgerd (guest, #58896) [Link]

Windows 7 improved Windows reliability dramatically; Win 10 adds WSL; MacOS moves to Intel and gives a stable Unix experience that can easily run Linux/x86 and Windows VMs. The rise of non-traditional devices. The dramatic churn of desktop Linux re-inventing key, painful bits of the software stack (UI, audio, etc).

Free desktops have chosen to re-invent wheels, while competing with commercial offerings (smartphones, tablets, desktops) that have been adding new stuff.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 1:20 UTC (Thu) by Baughn (subscriber, #124425) [Link] (6 responses)

That's about the time at which I gave up running Linux on the desktop myself.

Windows got better, WSL appeared, and the Linux desktop seems to be degrading with time. It probably isn't, strictly speaking—if I put it on the hardware I had back in 2016, I'm sure it would do fine, but nowadays I use a triple-monitor HDR, HiDPI setup where the DPIs differ massively between monitors.

I have never found a desktop environment or distribution which can handle this, and I spent over a year looking. Also, trust me; that wasn't the only problem. It was only the most in-my-face one. The second most troublesome would be–

Wayland isn't ready yet. X11 is apparently deprecated. Some applications are buggy on Wayland, some are becoming buggy on X11, and my GPU doesn't work with Wayland anyway.

...

I could go on. I love Linux, but it became an act of pure masochism to keep running it.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 17:02 UTC (Fri) by walex (guest, #69836) [Link] (5 responses)

«a triple-monitor HDR, HiDPI setup where the DPIs differ massively between monitors. I have never found a desktop environment or distribution which can handle this, and I spent over a year looking.»

Now that is one of my pet peeves, and I can name the guilty party: Keith Packard, who completely messed up, and repeatedly, the X11 way of dealing with multiple monitors, by introducing and repeatedly messing up XRANDR, which is ridiculously based on the model of a single virtual display into which all monitors are mapped as different regions, thus making it very hard to represent regions with different colour depths and different DPIs (plus some of the details were even more moronic, like layout being embedded in the "Monitor" section instead of the "ServerLayout" section.

The original model of a "display" made of multiple independent "screens" did and still does allow having "screens" with very different bit depths and DPIs (e.g. monitors and analog televisions).

It just requires some static configuration and a device driver that still supports the original model. I have some sample "old style" configuration files at http://www.sabi.co.uk/Cfg/X11/ that illustrate the fairly simple syntax.

Note: The main reason why the XRANDR buffoonery was adopted is that around the same time most monitors and TVs were 96DPI 18bit or 24-bit 22-24in so broadly similar.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 18:09 UTC (Fri) by raven667 (guest, #5198) [Link] (4 responses)

Am I remembering wrong, I vaguely remember the old model of multiple screens in X11 and I thought that it didn't allow for applications to span screens or move between screens, only the pointer could move. You needed to specify which screen an app spawned into like DISPLAY=0.0, 0.1, etc. and that couldn't be changed after the app started.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 18:41 UTC (Fri) by zlynx (guest, #2285) [Link]

That's right. I used to run dual monitors on Linux that way, with two video cards and each one was an independent X screen.

Xinerama was the new thing that let you use both displays as if they were one screen. There were also some fancy VGA drivers for Matrox I think, which could do two outputs off the same card (super rare!) as if they were one screen.

The XRANDR stuff the GP was complaining about is required in X because applications cannot just change their color and pixel format settings. Everything in X does have to appear exactly the same or it won't work. The only thing that works for X is to map everything into the highest supported format. For example, to combine a 4K and 1080p display you have to make both appear 4K and half-scale the 1080p. Then applications can render at 4K everywhere. If you plan to attach a 4K external monitor to a 1080p laptop, you need to prepare ahead of time so the 1080p appears as 4K.

Wayland is more flexible about this and the compositor system can use EGL to scale windows to outputs or adjust their color rendering.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 20:39 UTC (Fri) by walex (guest, #69836) [Link] (2 responses)

«the old model of multiple screens in X11 and I thought that it didn't allow for applications to span screens or move between screens, only the pointer could move.»

Both yes and no, depends on whether XINERAMA is enabled or not, and XINERAMA is sort of RANDR "statically". But things are more complicated:

There are two distinct cases: multiple cards or multiple outputs on a card
It is quite hard for applications and the X server to span applications across cards, and across monitors with very different characteristics (e.g. potentially rendering half of a glyph at one DPI, another half at another DPI, and much worse).
The original X model sort-of had that applications could probe the "screen" once at startup, and could assume that each window was entirely on one "screen".
It would have been possible to insist that applications should probe the "screen" on every repaint, and this would have handled windows moving between "screens".
However applications where supposed to be able to create windows on arbitrary screens, at user request, and this was really good enough. Down to EMACS for example being able to create windows on both text and X11 outputs.
For the simple case of monitors of very similar characteristics attached to a single card, merging the two outputs into a single screen with XINERAMA was quite adequate.

What RANDR added which was not in the original model was the ability not just to change resolutions dynamically, but to *add* them dynamically, and to add "screens" dynamically, which I think were fairly worthless, given that switching dynamically among a static list of resolutions was already possible; but it was convenient.

What RANDR took away was a simple way for applications to probe the "screen", and as a rule to support more than one card; I think that no current driver that supports RANDR supports it across multiple cards.

Since it is difficult or expensive to support more than two outputs with a single card, most RANDR configurations don't.

If RANDR has been designed better, and without large changes from one version to the next, and not as a complete departure from the "screens" model, a lot of trouble and limitations could have been avoided.

NB: It is possible to use RANDR to change the DPI reported for an area of screen connected to a specific output, with an obscure syntax (which is IIRC wrongly documented in the manual), but that seems ineffective, perhaps many applications ignore that. The xrandr command itself (a poorly misdesigned thingie itself) does not even report that DPI setting...

Jansson: On the Graying of GNOME

Posted Dec 19, 2020 22:07 UTC (Sat) by jafd (subscriber, #129642) [Link] (1 responses)

> What RANDR took away was a simple way for applications to probe the "screen", and as a rule to support more than one card; I think that no current driver that supports RANDR supports it across multiple cards.

My dual amdgpu setup, as well as i915+amdgpu laptop (where the external displays are driven directly by the AMD GPU) all beg to disagree.

> Since it is difficult or expensive to support more than two outputs with a single card, most RANDR configurations don't.

Both my RX580 (4 outputs) and WX4100 (4 outputs) beg to disagree.

Sometimes I use a 4-screen setup where two outputs are used from one card, and two from another, or three from one and one from another. RANDR supports all of them just fine.

Jansson: On the Graying of GNOME

Posted Jan 1, 2021 14:17 UTC (Fri) by daenzer (subscriber, #7050) [Link]

Yeah, limitations are mostly due to the architecture of the Xorg DDX, not due to the RANDR extension (which I'd argue is one of the better products of Keith's, and a big reason for X remaining relevant in a world where users expect monitor hotplugging to work seamlessly).

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 18:42 UTC (Wed) by remicardona (guest, #99141) [Link]

Canonical ditching Unity thus contributing more to core Gnome projects?

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 19:28 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (20 responses)

Perhaps GNOME has to look at itself? It spent years on trying to reinvent itself as a touch-based UI and alienated a lot of its users. I don't blame GNOME developers too much, after all, it's interesting to work on shiny things.

And then all the touch stuff basically failed utterly, with not a single device shipping with it in any quantity. This probably killed quite a lot of interest.

Out of curiosity, I checked xfce development and it's holding steady, with developers coming and going: https://github.com/xfce-mirror/xfce4-panel/graphs/code-fr... , https://github.com/xfce-mirror/libxfce4ui/graphs/code-fre...

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 20:54 UTC (Wed) by atai (subscriber, #10977) [Link] (1 responses)

How is GNOME failed in this regard compare to KDE? GNOME does not seem to stand out in this regard to any other contender in this space (FOSS UI for mobile)

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 21:01 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]

KDE _also_ had its own GNOME3 moment with Plasma and the whole "semantic desktop" nonsense. It's actually amazing that these two large projects managed to shoot themselves in the foot at pretty much the same time and in a similar fashion.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 23:08 UTC (Wed) by joshl (guest, #91369) [Link] (17 responses)

> It spent years on trying to reinvent itself as a touch-based UI

I don't think this is true; it's just repeated a lot. Evidently, most of GNOME's attention since the early 3.x releases has been on keyboard-and-pointer input. Try using it on a tablet and dispute this.

> And then all the touch stuff basically failed utterly, with not a single device shipping with it in any quantity.

In the same vein, GNOME spent years trying to invent itself as a desktop-based UI, with not a single device shipping with it in any quantity. I guess that's actually what killed quite a lot of interest, right?

GNOME has to be usable on the hardware people buy. Now, for example, that includes a great selection of hardware that uses both traditional input devices and touchscreens. Any desktop that cripples my hardware is not desirable.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 23:36 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (15 responses)

> I don't think this is true; it's just repeated a lot. Evidently, most of GNOME's attention since the early 3.x releases has been on keyboard-and-pointer input. Try using it on a tablet and dispute this.

Oh, they just failed. Yet it had been a goal that warped a lot of design into a total mess: https://wiki.gnome.org/Projects/GnomeShell/Design/
> Effectively works on contemporary hardware: the Shell will provide an excellent experience on touch-based devices and will scale down to small screen sizes. It has also been designed with wide-screen in mind

I remember bugs in GNOME being WONTFIXed with justification "won't work on touch".

> In the same vein, GNOME spent years trying to invent itself as a desktop-based UI, with not a single device shipping with it in any quantity. I guess that's actually what killed quite a lot of interest, right?
GNOME actually did ship on a lot of devices pre-installed (on desktop computers). I personally had a laptop that came with GNOME2 around 2009. But GNOME Shell has never been shipped on a tablet - a clear failure.

> GNOME has to be usable on the hardware people buy. Now, for example, that includes a great selection of hardware that uses both traditional input devices and touchscreens. Any desktop that cripples my hardware is not desirable.
The fact is, GNOME essentially crippled the most used devices (keyboard/touchpad/mouse) in favor of non-existent tablets.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 0:34 UTC (Thu) by joshl (guest, #91369) [Link] (10 responses)

>the Shell will provide an excellent experience on touch-based devices and will scale down to small screen sizes

This doesn't support your assertion. Working excellently on touch-based devices and small screen sizes != reinvented as a touch-based UI.

>I remember bugs in GNOME being WONTFIXed with justification "won't work on touch".

Again, GNOME must work on hardware people buy. If GNOME adds a feature that works when I use a keyboard and pointer, but doesn't when I want to use my touchscreen, that's simply broken and should be reported as a bug. In fact, existing situations of this type is why GNOME is clearly a keyboard-and-pointer-first desktop.

>GNOME actually did ship on a lot of devices pre-installed (on desktop computers)

Define "a lot". It was certainly a small fraction of Linux's tiny presence on desktop computers. If that company is now defunct perhaps that's due to GNOME 2 being a "total mess" and the market deeming it a failure.

>But GNOME Shell has never been shipped on a tablet - a clear failure.

A clear failure to ship on tablets? Sure. A clear failure to create an excellent UI that recognises the reality of the laptop I bought in 2019, not so much.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 2:44 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (9 responses)

> This doesn't support your assertion. Working excellently on touch-based devices and small screen sizes != reinvented as a touch-based UI.
Uhh... Whut? To be able to work on touch devices like iPad, you basically need to re-invent the classic desktop (circa year 2005) completely.

> Again, GNOME must work on hardware people buy. If GNOME adds a feature that works when I use a keyboard and pointer, but doesn't when I want to use my touchscreen, that's simply broken and should be reported as a bug. In fact, existing situations of this type is why GNOME is clearly a keyboard-and-pointer-first desktop.
And this ended up making GNOME hard to use for keyboard-and-mouse users. Which are basically EVERYONE, since GNOME has never shipped on a touch-based device.

> Define "a lot".
"More than 2".

> It was certainly a small fraction of Linux's tiny presence on desktop computers.
Yet GNOME3 ended up DECREASING the amount of Linux on desktops.

> A clear failure to ship on tablets? Sure. A clear failure to create an excellent UI that recognises the reality of the laptop I bought in 2019, not so much.
Your laptop has no touchpad? I pity you.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 3:56 UTC (Thu) by joshl (guest, #91369) [Link] (8 responses)

> To be able to work on touch devices like iPad, you basically need to re-invent the classic desktop (circa year 2005) completely.

Not really. There's basic iterative things you can do such as:

* Banish GNOME2-style pop-out nested menus wherever possible (just good for everybody...)
* Enlarge UI elements (this also makes it easier for pointer users)
* Ensure desktop/laptop-only interactions like right-click can be accessed in other ways (e.g. long press)
* Implement a useful on-screen keyboard (which will never be seen while running on traditional PCs)
* Implement gestures (e.g. pinch-to-zoom; can also be useful for touchpads)
* Implement screen rotation

And so on.

You must internalise that touch-based devices are not just tablets. There's a plethora of laptops with
touchscreens and 2-in-1 convertibles that we can't just pretend don't exist. If users want to use these
devices while running Linux, they should be able to.

>And this ended up making GNOME hard to use for keyboard-and-mouse users.

Have you ran tests on this? I certainly don't think so. Counterpoint: I interact with GNOME primarily with my keyboard and it's fluid and fantastic. Launching apps, opening files and folders, managing windows and multiple desktops, all can be done efficiently with just a keyboard with shortcuts that make sense.

>Which are basically EVERYONE, since GNOME has never shipped on a touch-based device.

Why? Linux, overwhelmingly, isn't running on desktop computers because it shipped on them.

>Yet GNOME3 ended up DECREASING the amount of Linux on desktops.

Is there causation? Is there even correlation?

>Your laptop has no touchpad? I pity you.

If I pay for a laptop with a touch-enabled screen or a 2-in-1 convertible, I'm not interested in desktop environments that patronise me and tell me I'm buying my hardware wrong.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 7:52 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (1 responses)

> And so on.
Indeed. And you end up with unusable nonsense. Look at iPad applications on macOS - they are nigh unusable because of this.

> Have you ran tests on this? I certainly don't think so. Counterpoint: I interact with GNOME primarily with my keyboard and it's fluid and fantastic. Launching apps, opening files and folders, managing windows and multiple desktops, all can be done efficiently with just a keyboard with shortcuts that make sense.
We have a real-life test: flagging popularity of Linux on desktop. GNOME3 drove me personally away to macOS.

>>Which are basically EVERYONE, since GNOME has never shipped on a touch-based device.
> Why? Linux, overwhelmingly, isn't running on desktop computers because it shipped on them.
I'm not following. GNOME2 had been partially successful, it was gaining popularity with actual first-party manufacturer support (e.g. Dell, IBM). And GNOME3 caused a lot of problems for the existing users in an attempt to capture tablet market.

To be fair, Windows 8 made the same mistake. But unlike GNOME they also supported Windows 7 during the lifetime of Win8.

> If I pay for a laptop with a touch-enabled screen or a 2-in-1 convertible, I'm not interested in desktop environments that patronise me and tell me I'm buying my hardware wrong.
Yet forcing users to GNOME3's "unique" desktop model was fine?

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 13:38 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]

> Indeed. And you end up with unusable nonsense. Look at iPad applications on macOS - they are nigh unusable because of this.

Things like laptops with touch screen, 2:1 laptops and netbooks etc was what GNOME was targeting. Not iPads.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 14:52 UTC (Thu) by sandsmark (guest, #62172) [Link] (5 responses)

> Have you ran tests on this? I certainly don't think so. Counterpoint: I interact with GNOME primarily with my keyboard and it's fluid and fantastic.

My pet peeve is that the gtk file dialog is unusable for me with a keyboard (and I have to use it because chromium only supports all functionality with the gtk3/4 dialog).

The biggest issue is that if I try to use the keyboard to jump to files starting with a letter, it instead starts recursively listing all matching files and folders, so I end up having to resort to the mouse pretty much every time.

And yes, it's marked as a WONTFIX (I don't remember if they were open to making it configurable, but it's a long time since it broke so I assume no).

There's also some minor paper cut issues with focus and whatnot, but that doesn't matter much since I can't use the keyboard in the first place.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 2:55 UTC (Fri) by motk (subscriber, #51120) [Link] (4 responses)

My current rage-out is the the gnome shell currently captures all my function keys and uses them as media keys, and there is literally no way to fix this except by fiddling around with keymaps and config files like some sort of caveman. Pressing F8 should get the results I need on the vsphere stuff I'm installing, not decide to darken my screen or whatever. Nuts.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 5:48 UTC (Fri) by pizza (subscriber, #46) [Link] (1 responses)

> My current rage-out is the the gnome shell currently captures all my function keys and uses them as media keys, and there is literally no way to fix this except by fiddling around with keymaps and config files like some sort of caveman. Pressing F8 should get the results I need on the vsphere stuff I'm installing, not decide to darken my screen or whatever. Nuts.

Are you sure this isn't the actual keyboard? I've noticed that modern laptop keyboards tend to map the Function keys to "alternative" functions (eg media controls) by default, and you have to use a Fn key to get the "traditional" function mappings. (There's typically a "Fn Lock" that will toggle the unmodified default..)

This is the case for my newer laptop, plus a couple of the other keyboards I have floating around here, but my old laptop, and the keyboard I'm typing this on, operate as one would expect.

(In other words, if the keyboard sends the "Decrease Screen Brightness" keycode when you press the physical F8 key, there's no way for GNOME to know that's not what you actually wanted...)

Jansson: On the Graying of GNOME

Posted Dec 23, 2020 1:51 UTC (Wed) by motk (subscriber, #51120) [Link]

It's a keychron k1, but I've noticed similar things on a dell lattitude where it would ignore Fn-whatever completely and just insist it was using a media keyboard. I'll play some more.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 9:09 UTC (Fri) by mjg59 (subscriber, #23239) [Link]

As far as I can tell this isn't the default behaviour of gnome (I'm running unmodified Fedora here, and all my F keys get passed to the application). As mentioned elsewhere many modern keyboards default to having the F keys perform media actions unless the Fn key is held with them, and the mechanisms to revert this vary depending on the keyboard in question.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 15:00 UTC (Fri) by zlynx (guest, #2285) [Link]

Yeah media keys are not something Gnome does. That's your keyboard.

My Dell has a control for that in the BIOS to set the startup default. You can also press Fn+Escape (I think) to switch the mode. There's a little lock icon on the key. Or press Fn+F8 (or whatever) to get either F8 or Keyboard Backlight, depending on the default Fn mode.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 19:11 UTC (Thu) by mcatanzaro (subscriber, #93033) [Link] (3 responses)

> The fact is, GNOME essentially crippled the most used devices (keyboard/touchpad/mouse) in favor of non-existent tablets.

This is a strange statement. Keyboard, touchpad, and mouse all work fine in GNOME. If they didn't, nobody would be using GNOME today.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 19:47 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (2 responses)

> This is a strange statement. Keyboard, touchpad, and mouse all work fine in GNOME.
Some functionality that works best with classic input devices was damaged in favor of touch-based interfaces.

> nobody would be using GNOME today.
Which is a good approximation of the reality.

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 0:18 UTC (Fri) by adam820 (subscriber, #101353) [Link] (1 responses)

>> Some functionality that works best with classic input devices was damaged in favor of touch-based interfaces.
Such as? I daily-drive a 2-in-1 that works clean for most things in either mode, except for obvious standouts where one or the other make the most sense.

>> Which is a good approximation of the reality.
...your reality, maybe? I have several machines that run GNOME, and my Linux-using friends/co-workers also all use GNOME. I think you're discounting the whole enterprise Linux market.

It feels like all the stuff surrounding the rough edges of the early GNOME 3.x days nearly a decade ago are still informing your opinion (which is fair, you're free to do so, I suppose).

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 1:48 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link]

I haven't been using GNOME for the last 8 years or so. The things that actually drove me away were:
- Missing right-click menus.
- "Hot corners" that couldn't be disabled.
- No bottom/top panels like in GNOME2.
- Broken menus in some apps (Eclipse).

I remember spending several days with gnome-tweak-tool, before giving up and moving to xfce.

Jansson: On the Graying of GNOME

Posted Dec 16, 2020 23:53 UTC (Wed) by nedrichards (subscriber, #23295) [Link]

> I don't think this is true; it's just repeated a lot. Evidently, most of GNOME's attention since the early 3.x releases has been on keyboard-and-pointer input. Try using it on a tablet and dispute this.

You are entirely correct. I really wish people would stop repeating this because it's just so obviously not true.

Jansson: On the Graying of GNOME

Posted Dec 17, 2020 12:07 UTC (Thu) by sam.thursfield (subscriber, #94496) [Link]

An impressive piece of research!

I got involved in GNOME years ago because I wanted to develop apps and started to see room for improvement in the platform. As the platform gets better, this avenue starts to close off. Luckily there is still plenty of work to do in the core GNOME apps :D

Jansson: On the Graying of GNOME

Posted Dec 18, 2020 9:42 UTC (Fri) by lag (guest, #143477) [Link]

I relied heavily on Gnome2, but once they took the decision to copy Apple in a way worse than Canonical's Unity did it was too much and made me take a complete turn to KDE. I really can't work comfortably with global menus and single window paradigms. It is probably ok if you worked on photoshop/illustrator/aftereffects all day, or if your only need was to alt-tab between an ide and a browser.
The rest is history, and Gnome developers have been really patronising in the early days, dismissing clearly broken stuff and regressions as "useless" and obsolete needs often taking the blame on its users' bugfix requests.
Not saying that other WMs developments don't suffer of similar problems, but the way GnomeShell alienated part of its users when it came out has been quite unique.