File chooser grid view (ie thumbnails) - that's nice, but I'm so jaded at this point to care anymore. I've been using ubuntu on and off as my personal system for the past 15 years, it sure beats windows in many respects (typing this on MacOS). But I'm just too old to be excited by UI features that should've been here long ago and that were the standard in the software of the late 90s and early 2000s. There has also been a vocal part of the community defending the lack of thumbnails in the file picker, so this whole story has a bitter taste to it.
One more thing, it feels like open source systems and UIs are developing at a glacial pace these days, either the tools used to develop open source software are sub-par or maybe there is a lack of people working on it and effort put into it. So glacial that in the age of AI, a grid view is somehow a notable feature. There seems to be less to be excited about these days and I have given up expecting great things.
This seems like a misrepresentation of the entire file picker saga.
The “vocal parts of the community” were not against the UI change. For the most part, being the ones who were charged with putting in effort to make this improvement for free, they were pointing out the significant technical challenges it involved, and they couldn’t find someone to volunteer their own personal time for free to take on this challenge.
As Gnome has made massive strides towards simplifying its codebase, these changes become far more achievable and reasonable (click through to the developer changes page and you will see GTK4 has deprecated something like 20 possible file picker classes, in favor of 3, which provide equivalent functionality).
Of course, the same people who complain but are unwilling to either put effort or money towards fixing these supposedly deal breaking problems, have also spent the last 2 years complaining about the code and architecture cleanup effort Gnome devs have made to make it possible to fix those problems.
I'm not sure what the hell you are talking about. It took like 5 years for someone from the dev team to acknowledge the first bug report [0].
Look at the attitude shown towards the constant request for implementation [1]:
> Got patch? Share. Gotn't? Just CC and be silent.
Really facinating thread... I read it every year to remind myself that my brother was born that year and now, 18 years later he's an adult so he's allowed to have beer and use a filepicker in GNOME
>> Got patch? Share. Gotn't? Just CC and be silent.
Indeed. This hostile attitude towards criticism of lack of basic QOL features, is what's keeping Linux at ~2% market share, away from the average users, and we'll be perpetually stuck in "$CURRENT_YEAR is finally year of the Linux" meme.
Hate Windows all you want kids but if you look at independent online tech related forums where people ask questions or problems on Windows, the answer is always a solution and never "RTFM!"
KDE should be commended for acknowledging and fixing some of the usability issues highlighted by the "Linus Tech Tips Gaming on Linux challenge" fiasco form 2021, when most of the Linux community went full neckbeard on him telling him he shouldn't use Linux if he doesn't know how to use the terminal.
That's because people who answer are being paid to do so, or have other types of incentive system put in place, and they wouldn't want risk losing them by writing what they actually think.
Not so much with open source, where the one answering has (potentially) just came back home from work and is just taking the time and energy to do it from their own free time and goodwill while they sip a beer and think about what to make for dinner in a couple minutes.
>That's because people who answer are being paid to do so
No they're not. Read my comment again. I wasn't referring to proprietary forums owned by the corporation who makes the platform, but to generic public tech forums that aren't owned by any company.
Windows users tend to be way more helpful than Linux users over there.
I guess one explanation is that Windows users are immensely many more, so the small portion of users who are generous enough to contribute in those kinds of forums will also be formed of more varied attitudes and skills.
Many customisations are not possible in Windows let alone requesting any specific feature from the devs.
In cases where I had ti seek forumn help in windows three common scenario were:
1. Well known bug with a simple fix that user has to make.
2. Well known bug that has a weird work around or hacky reg edit.
3. No idea what the cause is and you just reinstall and hope it doesnt reoccur.
And the ltt video was an obsolute joke because they made little to no effort to learn anything about Linux and operated it like they do windows. of course it was such a bad experience. that's what most reaction I saw was too. if you are great at driving automatic but decided to just drive a stick without putting any effort to learn its specific details or the different, that will end horribly. Cant blame the stick car is terrible.
> And the ltt video was an obsolute joke because they made little to no effort to learn anything about Linux and operated it like they do windows.
The point of the video was to approach desktop Linux in a way similar to how your average layman might (which when considering the dominance of Windows probably means using it like it’s Windows) to see if it’s ready for the less-technical masses.
It might sound silly but this is the exact sort of test that desktop Linux needs to pass with flying colors if it’s to have any hope of gaining traction beyond the extremes of the technically capable minority and grandparent who only needs a box with a browser capable of checking their Yahoo email.
Yeah, no, he read what he was doing, but the "packages to be removed: Gnome, X.org server,..." warning doesn't mean anything to anyone who just wants to use a computer to play games instead of compiling his own kernel.
Get out of your techie bubble and go see how actual non-GNU humans use computers to get a sense of how poor some SW can be at communicating information to clueless users and how little interest clueless users have to learn the nitty gritty things about their machines. They just want to play games or whatever.
In fairness, the literal next line of text up on his screen read "You are about to do something potentially harmful", verbatim. That's why people memed on him, not because he doesn't know what Xorg or GNOME is.
The non-techie folks are fun but they mostly just complain about their iPhones changing overnight and a bad Windows driver update. It's nice to just sorta be done with that rat race and focus on free software, if it's all the same to you.
>"You are about to do something potentially harmful"
That doesn't mean anything anymore. Such omenous warnings pop up everywhere, even when installing legitimate drivers and applications or updating your software, that at one point you desanitize yourself and just click Yes, Next and Accept to everything to finish the installation and start pwning some n00bs in your vidja games.
Blaming the user here is stupid, especially since Linus was following the instructions he found online, which is exactly what most users would do, even if one of the steps throws an omenous warning.
As a clueless user coming from Windows you don't expect that a Steam installation should lead to the uninstalling of your desktop environment and display server, since on Windows these are part of the core OS and no amount of fuckery can remove them. This is a massive barrier to entry for new users on Linux.
That's why Valve threw in the towel on convincing people to install Linux on their PCs and instead sold them a PC based console with Linux preinstalled and already configured to run games. Linux on the mainstream home PC is a dream that even Linus Torvalds said won't happen.
> Such omenous warnings pop up everywhere, even when installing legitimate drivers and applications or updating your software
Case in point, Microsoft Excel throws a big scary warning every time you open a file and enable editing mode. For completely basic usage, Microsoft throws in the towel and claims that anything bad happens is on you. Why should anyone expect users to take these messages seriously?
I'm not strongmanning the idea that everyone should abandon MacOS and Windows. It would be nice for those companies to lose their control over computing, but I can't move the needle either way. I use Linux on the server though, and it works fine for my desktop too. It's a better enthusiast platform than either MacOS or Windows, and it lets me avoid messy workarounds at work like Colima and WSL.
I'll happily blame Linus and his vidya-addled brain if that's what you want to blame it on. I suppose that's why kids ate the Tide pods, after all.
>the ltt video was an obsolute joke because they made little to no effort to learn anything about Linux and operated it like they do windows. of course it was such a bad experience.
If you need to grow a neckbeard to operate the computer, that is a terrible operating system.
Linux is, and perhaps always will be, an engineering OS. It's much better for development and IT than Windows, and Windows is worse than Linux for development or IT, precisely because the trade off of designing for normal people vs. designing for techies.
The funny part comes out in the rare instances that Linux beats Windows in normal person terms: KDE is way better than Windows's DE, and Dolphin (the file manager made for free by like five Dutch guys) makes Explorer look like a joke. (That's because Microsoft doesn't prioritize those normal end user facets, ironically.)
But, in general, as a techie (neckbeard haver?) who prioritizes techie things more than non-techie consumer things, you'll be way happier by learning Linux (especially truly learning it -- we're talkin' Arch). Then, as frosting, you have the abstract, philosophical stuff like "I'm not slaved to a callous all-devouring corporation", "I'm using free software wow" etc.
No one makes this distinction because the conversation is de facto shifted over into the reductive premise that "Linux is worse than Windows" vs. the somewhat fantastical premise that you can dump enough GUIs on Linux to make it as good as Windows for normal end users. But we're in a thread titled "GNOME 44", so I guess that tracks. :D
>its desire to neckbeardify its users is its undoing.
Uh, no, because Linux isn't currently, nor was it ever in the past, successful or appealing because of how much it caters to normal users.
It's been successful and appealing based on how much it caters to techies (my earlier post).
That Steam decided to support gaming on Linux is why its market share among Steam users has fluctuated around 1% and why Linus Tech Tips did an exploration video about it, but if you're already measuring Linux's success based on how many "non-neckbeards" use it, it's already undone, and it was always undone (and, if anything, arguably getting slightly less undone, as its ease of use for and popularity among normal people has grown significantly over the last 10 years). :p
Except Linus wasn't trying to run steam.exe on Linux, but instead trying to install the official (borked) Linux Steam package shipped with his Pop_OS ISO, which happened to completely nuke the desktop environment and display server, landing him in a tty.
And of course the Pop_OS team was quick to blame Linus for being an idiot and not knowing he should first open a terminal and go `sudo apt-update && sudo apt-upgrade` before installing anything, instead of expecting packages shipped with the OS to just work(TM) and not nuke your system.
Linus was warned, that something wrong is going to happen and whether he really, really wants to proceed. He did. Without even reading what apt is complaining about.
The bug in Windows updater that wiped out users profile dirs with all documents inside didn't warn in a similar way. Yet, we are not harping about it all the time. We all understand that it was a bug, just like that borked steam package.
The number of new OSX users with OSes that had slowed down to a crawl, because the users came from Windows and thought closing the Window closed the app, I have seen…
In my teens all I wanted was customization. Now that I want to have a social life and I actually have things I want to get done other than staying up until 4am customizing my Gentoo install, I’m starting to believe that customizations are overrated. Not saying we shouldn’t have them, but we should stop bringing that up to as an excuse of why things are ok being left alone with all the wiring exposed.
Just a personal view that I feel I can support with Apple’s sale numbers.
So yeah, people who are volunteering their time for free are asking if this is so important to you do you have a solution other than demanding that they volunteer their time for free for your priority.
Yes, I agree. Some devs do volunteer for free. Some others do it for some other form of incentive. There's a cynical comment down below that points to this fact. Like seriously implying that most of the Windows-based support comes from paid engineers, guys, get a grip. A lot of people in the Windows community offer support for "free", thanks to the MVP badge incentive, it helps them demonstate they have reached some substancial proficiency in some Microsoft stack technology... In my view, that's not free... you're getting something for your time, something that might enhance your chances of getting a job. A lot of people in core-teams do get paid through donations. Donations, I may add, many times come from the same bastards they so profusely seek to brush off in bug reports such as this infamous one.
Actually I talked to them on IRC and Gtk3 is frozen and even if you have a patch that fixes it they will not apply it. Gtk4 is probably different but no one except GNOME itself uses Gtk4.
This comment is full of misinformation. GTK3 is not frozen. There are PRs for it open right now. They just dropped the Autotools build recently, in fact. Elementary is currently using GTK4. I can only assume Cinnamon is on their way to using GTK4, but I don't keep up. XFCE and MATE move much slower than the other popular desktop environments.
XFCE and MATE are going to be slow on the uptick since GTK4 isn't a desktop toolkit at the moment. From what I understand, it has the potential to be, but the maintained haven't opened up certain APIs.
So the people from Redhat on #gtk were lying when they told me Gtk3 was frozen? I guess getting gaslit is slightly better than gtk3 actually being frozen broken. Thanks! I'll go back to trying to get the defaults for gtk3 filechooserwidget.c changed.
As for the desktop environments, okay, so Elementary uses Gtk4 too. That's GNOME and Elementary. 2 of them. And we're ignoring all the applications that still use gtk3 (or 2) and the very few of them that have ported to Gtk4 (since the docs aren't even finished for 4).
There's a fun game that toxic internet communities like to play, where the elites of the community say one thing in one breath, and then another thing in another breath. Then their legions of weak-minded followers will cite that one thing or the other thing whenever their community identity is threatened. Either or, the community will gladly lie on the behalf of the elites. The elites pretend not to see the lies and the low-ranking community members pretend that they have no responsibility.
Silly tribal nonsense that applies to more than a few maintainers and software.
Why is it GNOME's responsibility to port non-GNOME applications or desktop environments? How are the GTK4 docs not done? I use them at least once a week.
Was this a Red Hat priority? Then Red Hat can complain about how their priorities are not being met despite them funding this project.
I mean, this is a worse argument. It’s saying “How dare you not volunteer your free time or paid time to meet my priorities even though it’s not your priority or the priority of the people who are paying you”.
That is unsubstantiated. How would we know that the same people complained about code cleanups and not having a file picker? Do you have any proof?
Users like me who got annoyed by never being able to easily select the wanted image (when in a gtk app), for many many years now, had no insight into any architectural changes or code cleanups of gtk4. So there weren't any complains from this group there.
This is the typical "everyone is against us" you often hear from the gnome side. It's a shame to read this under such a nice release, which has only positive changes as far as I can see, it really seems to address a lot of the criticism in a good way.
> How would we know that the same people complained about code cleanups and not having a file picker?
I also saw some complaints of this form. If I recall correctly they were non-contributors complaining that Gnome devs were dedicating too much time to refactoring and deprecating APIs and not spending enough time implementing features that they wanted.
I'll give you that Gnome developer communication often comes off as defensive, but as an outsider I'm also shocked at the volume and tone of complaints they get -- I doubt I would be able to put up with such toxic users for very long.
I wanted to make some minor changes and the kind of toxicity being directed at developers was scary, even as someone who was only looking to help with some translations, etc.
Ever since that experience I’ve become highly sympathetic to open source devs of all stripes. The only partial exception are blatantly corporate projects making blatantly corporate decisions (eg React/Angular or VS Code telemetry)
Im not on the Gnome side. I haven’t contributed beyond very basic bug fixes and would hardly consider my contributions to Gnome meaningful.
But as part of searching for some contribution to make I went through a whole list of bug and feature requests.
And while it’s almost impossible to prove that the people are the same, what was very evident was the sense of entitlement displayed by both groups.
Frankly, even as someone who isn’t part of the Gnome community and was simply looking to maybe fix some language translations, the toxicity from the people who didn’t contribute anything was overwhelming. It’s fair enough to complain about something not being up to par. But that very quickly transforms into claims that the developers have an evil agenda.
I'm also engaged in a (way smaller ofc) FOSS project and I can relate a bit, but it's just not proper to mix the different kind of criticisms into one, and to create the image of one enemy. GNOME is big enough that there is a lot of feedback, but that only makes it more important to properly evaluate that feedback.
The file picker bug situation and the bug report handling was just horrible (the quotes in the thread above show that quite well), the project has only itself to blame for the negative attitude it got from that. Let there be praise that this situation might improve now (will take ages until GTK4 is everywhere, probably never), not use this situation to bash the users that took the time and effort to give their feedback.
It reminds me of the Firefox users who to this day view Mozilla with contempt for having deprecated old add-ons or some barely used feature (highly recommend reading your average pale moon user's reaction when pale moon does the same)
Right? How dare those users be upset at their workflow being broken? Everyone knows that if you can't write code, you should have no expectation of software ever working for you! After all, the only other way to ever have any legitimate opinion about any product ever is if you're paying for it, but we conveniently fix that problem by making it literally impossible to pay for Firefox! Why can't the users just shut up and be grateful for the features that haven't yet been ripped out?
I don't feel any contempt for Mozilla but at the same time Firefox was never as enjoyable to use once it began chasing Chrome. I would put the high water mark at Firefox 3.6.n.
Generally speaking, FireFox still retains most of the same features, it's just harder to configure it to be well-behaved (tab bar on the side was a native Firefox preference originally- now it is a real learning experience to set up). There is also a lot of stuff bundled and enabled with Firefox by default that ought not be included at all (WTF is a "pocket"?).
I use Firefox since it was Phoenix, back on my Windows 2003 and except there was time when it was noticeable slow so I was forced to use Chrome/IE I hardly can find things to complain about regarding changes.
I see other people complain a lot, which makes me think I'm missing some big part/have some unusual usage for Firefox. Any hints what I may be missing/doing strange way?
Firefox decline was way before add-ons deprecation, it more or less started with Chrome being a few versions in due to simple configuration, streamlined ui, increased security and most importantly performance.
This also seems like misrepresentation of the entire file picker saga. First of all, it is absolutely laughable that something as essential as a thumbnail file picker is missing for that long, literally decades. Several patches have been submitted by people "willing to put muh effort and money in it" and got rejected because superficial idiotic reasons. On the other hand vast amount of resources are put into widely unpopular changes like client side window decoration, removing configuration options, turning everything into a tablet UI or breaking font rendering. And the "massive strides towards simplifying its codebase" means nothing more than dropping backwards compatibility on a regular basis while making everything more bloated and resource demanding.
Maybe Gnome is a project to sabotage the FOSS desktop. Stranger conspiracies turned out to be true.
Personally I'd recommend: Don't use their software. Don't contribute to their ecosystem. Don't go to their conferences. Don't ever give them any money.
Thank Icaza and RH for that. Gnome when it was born it was amazing. They shipped a window manager with LISP capabilities, Sawfish. Totally the opposite of Gnome3. I wish the FSF forked over Gnome in order to bring a Guile interface to Mutter instead of GJS. It would be much better. I memember when I discovered emacs when I was 15 or so, I was astounded. Self-documenting, with geek-nerd culture in every bit of it, a text editor with a chatbot, tetris and a text adventure bundled with it and it also was a LISP interpreter. You could do mail, news, IRC and with proper plugins, play IF text adventures with Malyon, allowing you to copy and paste from everything to everything. Now with Gnome 3 thanks to ReadHatware/XMLpest/SystemD everything wants to be a bad NT clone from Micros~1 with SystemD instead of svchost.exe.
>This seems like a misrepresentation of the entire file picker saga.
This is false, GNOME devs were promoting the idea that you do not need the feature. So the GNOME devs/fans were telling everyone that they are doing thingd wrong, for example if you want to upload a image you should not use the Upload button and get the File Selector dialog, the GNOME way is to Drag and Drop the image from the file manager.
Question is how you explain other DEs have this feature? What makes it so hard to do it in GNOME, is the C language,t he tool, some missing fundamentals in the gnome core stuff ?
> One more thing, it feels like open source systems and UIs are developing at a glacial pace these days, either the tools used to develop open source software are sub-par or maybe there is a lack of people working on it and effort put into it. So glacial that in the age of AI, a grid view is somehow a notable feature. There seems to be less to be excited about these days and I have given up expecting great things.
The glacial pace of the whole Wayland vs. X11 situation has nearly put me off desktop Linux completely in the last few years.
It's not just thumbnails. In ~2015 Matthias Clasen of Redhat removed the ability to chose between filename-entry and path-bar mode in the gsettings org.gtk.Settings.FileChooser location-mode. This meant it was no longer possible to open a File->Open dialog in a Gtk3 using program and paste or type a file path (like, /home/user/somefile.txt) into the File->Open window. Instead only the mouse-using bubble buttons 'path-bar' mode is available. When you attempt to paste a file path into the file->open dialog you get an error, "The folder contents could not be displayed.\nOperation not suported". (ref: http://erewhon.superkuh.com/gtkfilechooserwidget-paste-fail....)
When hundreds of people asked about this bug he said,
and all attempts to get them to fix the bug since have failed. I continue to ask in #gtk on IRC at least once every year. Gtk3 is permanently frozen broken and the same bug actually still exists in Gtk4. If you open a file->open dialog and just paste a file path you get an error. I have attempted to re-write the relevant sections of gtkfilechooserwidget.c myself, just for my machines, but I have only succeeded in allowing typing/pasting in File->Open dialogs the first time they're activated in a program. Never the subsequent activation.
I've talked to Gtk4 users on #gtk IRC and apparently this bug still exists in Gtk4. When they refuse to fix things like this you have to wonder who they're making Gtk for. Business users of IBM/Redhat's who's employees use a mouse for everything, I guess. And yes, I know gtkfilechooserwidget.c is spaghetti and it's hard to change anything without breaking something someone else wants. But this is core functionality. It should not be ignored. Especially not for an entire decade.
>maybe there is a lack of people working on it and effort put into it.
This has always been the case with FOSS GUIs.
When everyone is happy CLIing everything and telling those who aren't to get bent, GUIs move at a snail's pace. I think it's amazing we even have stuff like GNOME, KDE, and xfce in the state they are in today.
Same old same old. I don't want a tiling WM, thanks. I'm a big boy, I've been using WIMPs for 35 years, and I can manage my own windows very nicely with a few keystrokes. I don't need a tool to do it for me and fail to read my mind.
I also don't want a bloody great horizontal panel across the top of the screen wasting 1cm of vertical space on a widescreen display. No it does not help if it's visually broken up into separate islands: it's dead to me. Wasted space. Either use it, and you'd better find a remarkably good set of uses for it, or give it back.
> Same old same old. I don't want a tiling WM, thanks.
That's a very naive take of tiling. For traditional tiling, maybe.
But unlike WM that do tiling only, hyprland does more: it supports my keystrokes and lets me configure how much of space each window get, while also letting me "float" the windows like a regular window manager
> I'm a big boy, I've been using WIMPs for 35 years,
Have you considered the possibility you are so set in your ways that you are neglecting new and useful tool? The tone of your comment reminds me of the summary dismissals of fzy.
> I also don't want a bloody great horizontal panel across the top of the screen wasting 1cm of vertical space on a widescreen display
Then make the panel vertical instead of horizontal, or just close it lol - all of which can be done with a few keystrokes
You know that there is basically a standard set, imposed by Windows in about 1986 or something and also supported in GNOME 2, MATE, Xfce, LXDE, etc etc.? I am more interested in if it supports them. I mean, I don't know what your set are, and I am not for a moment saying there's anything wrong with them, but there are standards for this stuff, used heavily by millions of blind computer users for example.
> Have you considered the possibility you are so set in your ways that you are neglecting new and useful tool?
Could be. I am a professional assessor of, and commentator on, this stuff, though.
I mainly use a desktop I switched to in 2011. :-) Before that, I changed in 2004, after a change in 2001, after a change in 1995, after a change in 1992, after one in 1989, etc. etc.
I mean I am an old pharte, fair call, but I am a reasonably adaptable one, I think. :-D
> Then make the panel vertical instead of horizontal
Why don't any of the screenshots show that, then?
I see 6 horizontal panels in the screenies on the homepage and Github, and one with none. From that, I don't think it's unreasonable to conclude this is not a core feature or something.
> You know that there is basically a standard set, imposed by Windows in about 1986 or something and also supported in GNOME 2, MATE, Xfce, LXDE, etc etc.? I am more interested in if it supports them. I mean, I don't know what your set are, and I am not for a moment saying there's anything wrong with them, but there are standards for this stuff, used heavily by millions of blind computer users for example
Yes, and the terminal not copy pasting with ctrl-c infuriates me: it's long be remapped to ctrl-x (right next!) on mine to achieve consistency.
As for the general shortcuts, it seems to, and the extra ones I wanted because I'm used to them (ex: Win+Shift+S = snipping tool) I could add easily. Now it's effortless to do the same thing I did before (Win+L locks etc)
> > Then make the panel vertical instead of horizontal
> Why don't any of the screenshots show that, then?
A lot of the information is implicit to a domain: a video of bash not showing handling of say SIGTERM shouldn't be taken to mean it doesn't handle them, the good old absence of evidence is not evidence of absence
> From that, I don't think it's unreasonable to conclude this is not a core feature or something.
Core features are more like what most people tastes converge to. You are to pick app to achieve your desired end result: waybar for example is not part of hyprland, it's often used with it, because it's complementory like say bash and vi, even if there are other possible choices
Anyway, seriously give it a try- to the likes of fzf too. We have a lot of cool/better tech than existed in 2011, about 12 years ago
Your previous upgrade cycles were 3 years, so I think you're long overdue - might as well take the best there's now!
If you do upgrade, I suggest not trying to take any of your old tools if they are not wayland native it's often not the effort: there's a replacement somewhere, you just may not know about it, but it will certainly do a better job once you find it
I say all that super seriously: I really love my new stack, it gives me leves of productivity I had only been able to reach in Windows (the UI is wonderful if you know how to use it with the right keystrokes)
> But unlike WM that do tiling only, hyprland does more: it supports my keystrokes and lets me configure how much of space each window get, while also letting me "float" the windows like a regular window manager
All dynamic tiling WMs do those things. Hyprland is not special in that way.
Linux desktop has not had this much velocity basically ever. We went from it being unusable for most work to useable for nearly everything in the last half decade.
Gnome 2 or XFCE4 with the autoconfiguring X.Org instead of X11 was good enough since the 2.6 Linux release, FreeBSD 4.3 or since OpenBSD 4.6. OpenBSD with toadd/hotplugd brings the same capabilities on device automounting as Linux without a dependency nightmare. If you simlink /vol to /run/media/$USER, you magically get UDEV-like mounts. Add tray-app from packages and you'll almost get the same usability as Ubuntu on device automounting. From that everything went downhill on desktops, with tons of bloat to do basic things.
Then Microsoft threatened to sue, everyone panicked, all the big players did weird phone-like UIs or Mac-like UIs and the little independents who weren't worth suing went off and made half-assed Win9x clones.
Microsoft got what it wanted. It sowed discord, broke up a harmonic community, and drove it to the four winds.
If you've only seen it since 2010 or so, all you've seen is people trying to reassemble the wreckage.
No, this is pure, fabricated nonsense. People involved in those projects, e.g. Miguel de Icaza, have told you numerous times, as have a lot of users who lived through that time.
But the MS threats are a matter of widespread legal record. This is not some kooky opinion of mine. This happened, it's documented, and the responses are too.
For instance, Novell (owners of SUSE then) and Red Hat were specifically named:
I remain amazed this is totally forgotten. Even for the people in their mid-twenties now, this was THE news when most of those folks were probably first discovering Linux and computers.
I mean, yes, it got lots of press coverage, but nobody involved in the industry thought it really had any merit. In other words, nobody I ever heard of thought that there was any AT&T code in Linux.
For those outside of the industry, well, it may have seemed possible, and to those who don't know anything about computer program code, it would not be apparent that:
• it's hard to check;
• the volume of code involved is vast;
• it's hard to prove because different people often implement the same things the same way.
But to those in the trade, it was manifestly obvious it was false, but it would be a lot of slow, expensive work to prove it was false.
Whereas the MS thing is different.
There are only so many GUI designs in the world, and to those of us who used this stuff in the 1980s, paid some attention, and remember, it's easy.
Early 1980s GUIs were wildly diverse. Some had menus at the top of the screen, some at the side, some inside windows, some only appearing when you clicked. Of the ones that were on demand, some were a list, some were pie-shaped, etc.
Some OSes put drives on the desktop, some in a folder, some didn't show them.
Some had app launchers, some didn't and you used the file manager to find binaries and run them directly.
Some had a clock, some didn't, some it was optional. Some it was on the desktop, some it was in a window, some it was in a menu somewhere.
Some multitasked, some didn't. Of the ones that did, some had app switching tools of some form, such as a window of app icons, or a list, or a menu. Some had no graphical switcher at all: if you want a window, then click on it. Can't see it? That's your problem. Go find it.
Some kept files on the desktop. Some kept minimised windows on the desktop. Some kept a mixture. Some didn't use the desktop directly. Some didn't have any desktop metaphor at all. Some you couldn't see it.
The point being that there was almost no standardisation and nobody much felt that was missing or wanted. It was just how things were. For some this was a reason to prefer one vendor's offerings over another.
Then OS/2 1.1 and later set a look and feel that others copied: Microsoft in Windows 3, and Motif on Unix. CDE and other Unix "desktops" took some parts of it. OS/2 3 took some of CDE, and so on.
Then Windows 95 came along, and it was pretty good. It had a visual app launcher, as standard. Nothing else had one front and centre like that. It had a visual app switcher, again front and centre. It had a click and a standardised way to show system status icons.
It was for its time extremely good. Amazingly good, and it hurt OS/2 and classic MacOS.
MacOS 8 and 9 picked up some features. Small stuff: arrows on the corner of "aliases" (its version of "shortcuts", which were Microsoft's visible version of Unix symlinks -- something which Unix didn't distinguish from any other files), and an alt-tab style switcher, and stuff.
Every new desktop that came along after Win95 copied some parts of it. On Linux, KDE, GNOME 1, Xfce. QNX's Neutrino. OS/2 4. BeOS.
And everything else died, or went away to some tiny unimportant niche. OS/2 basically died, AmigaOS died, DR GEM died, Irix died. The handful of surviving graphical Unixes had CDE, based on Motif, which was licensed from MS anyway, and Solaris adopted GNOME 2.
Only two pre-Win95 GUIs survived: NeXTstep became Mac OS X, and Acorn's RISC OS survived Acorn and was updated by other companies. Now it's FOSS.
And you know what? Those are the only 2 that had visual app switchers, and one of the people that wrote RISC OS went to work for NeXT right before NeXTstep came out and the rest of the team swear blind he took his Archimedes with him to California and NeXT nicked the idea.
Microsoft knew about NeXT -- everyone did -- and took the germ of the idea from NeXT.
If you were there, as I was, as a lot of people still in this industry that are not yet retired or dead, it's visible and obvious and clear that the Linux desktops did copy wholesale from Microsoft, specifically from Win95, and Microsoft being what it is, yes it would have patented all this, because it's highly litigious, it's got sued a lot, it often lost (e.g. against Apple, and against STAC, and against DR/Novell/Caldera).
So MS did have a case.
And lots of companies settled with MS, and those that didn't promptly redesigned their desktops and put something new out with a different design.
And they have to deny this because otherwise they would be admitting guilt.
But also, it's not directors and lawyers that design and implement GUIs, and it's not programmers that litigate against other corporations, so it's entirely possible that some of these people have no idea why management suddenly funded their project and they got to ship it.
I don't know. I feel like 2002-2008ish was the most exciting time for the Linux desktop. It was going from strength to strength and really holding its own. Around 2009-2010 something happened and Linux really fell behind on the desktop front.
To me, GNOME is what makes Linux a desktop OS as pleasing to use, if not more, than macOS. The apps are neat and focused, do one thing and do it well, and everything “just works”.
Using macOS for ~5 months now after decades of Windows and Linux and it's... subpar. Which is surprising since it has so many devoted evangelists. I'd honestly rather use Gnome than macOS on the desktop.
I've been using macOS for quite some time and I have the same feeling. I think it is a great system in general for your "average Joe" but for development and power user stuff it is inferior to Linux.
Don't even get me started on Docker and how it works on it.
The Mac OS X of around 15 years ago had an elegant simplicity to it that I really appreciated. Everything just seemed to work well.
That seems to have been lost. New and often unwanted features just add to the complexity. A lot of first party applications seem subpar. And iCloud often seems more like a nuisance than a feature.
I still like Apple hardware but I’m looking forward to trying Asahi Linux soon.
I'm a Windows wizard first and foremost, but I have to give credit where credit is due.
MacOS is very clear about what it will and won't do. So long as you subscribe to the One Apple Way(tm) and adhere to what MacOS will do, you're golden; and a common man's use case almost never goes beyond what MacOS will do.
Combine that with a consistent design language and I, sincerely, would recommend a Macbook Air of some description to most people. MacOS Just Works(tm).
Windows and Linux are for the tinkerers, for the masses MacOS is the best option out there.
It's possible that a thing is simply inferior to another. Not everything has a place in the world. The amount of tinkering and preferences available on macOS is staggering, and makes Windows an also ran. Remember that Microsoft straight-up removed the ability to have the taskbar on the sides of the display. It's just gone. Microsoft can be commended for its software back compat, but its UI has deprecated so many uses through the years that Apple has diligently kept, hidden to varying degrees.
To this day, if you want, you can run Finder with no chrome, as a strictly spatial file navigator. There is a vibrant, reactive, and profitable market for third-party software to change the capabilities of the OS. Windows has a market too, but it is far from what the mac community manages to make. macOS is still a tinkerer's dream to a greater extent than Windows.
I would say that Linux is better only for a particular type of power user/dev.
In my case, as a computer user of 27 years who’s been writing code in some capacity for around 20 years, macOS is preferable because its stock config is the closest to how I want things to work. It takes very little config prodding to get to my productive “happy place”.
With Linux on the other hand, reaching that same “happy place” feels like a Sisyphean task… no matter how lunch time I spend on tinkering it still isn’t “right” and that further detracts from productivity because it’s distracting. I don’t think I could make Linux my primary OS without developing my own DE, which rather than Sisyphean is Herculean.
Unless you're shipping an iOS/MacOS app though, getting to that "happy place" is usually the same number of tweaks away. At most places I've worked, the Linux build environment is just the one we use for production, and the MacOS one is ported to Darwin so devs can work on it. Like the parent said too, if you're working in a Docker environment it becomes almost a no-brainer. MacOS has it's uses and you can argue for it in places like design and video editing. For development though, MacOS feels increasingly less-relevant as time goes on. It's not a bad thing, but the majority of Mac devs I've met in the past half-decade do their work over SSH or in a VM.
For Docker purposes I’d rather keep a headless Linux box around that I remote into with a Mac than have to try to live with the Linux desktop as my daily driver.
Same experience here, coming up on a year however. I was assuming at first that I was just used to Windows since that was my primary OS my entire life, and tolerated any finickiness/configuring to my needs on Linux DEs. But it still hasn't clicked with me.
Don't get me wrong, it's perfectly fine and probably works great for most people. But there's been a plethora of tiny issues and problems that make it juuust annoying enough to use for me that I'd rather not.
Window management is one example. Needing a third party program running to fix scrolling with a mouse is obnoxious. Had an update that failed because it couldn't close a program, I clicked the cancel option on the dialog, and when I later closed that application then the machine restarted and applied the update. I haven't had even Windows force an update on me like that in years. The VPN client will store a password, then not use it (although maybe that's related to the host?). There have been other issues for me, those are just the ones I remember.
It's a shame too, I love the hardware of the MacBook: screen, speakers, track pad, even the keyboard are all great. If this was a personal device I'd probably be getting into Asahi despite its very much in development state (and yes I know a lot of what I like about the hardware has underlying software making it great which doesn't translate to Asahi right now, but I'm confident it'll get there one day).
I can't believe how bad Finder is despite having to use it daily on work machine.
That one view where icons don't snap to any grid but move like pixels in a raster editor... I am baffled as to why did it ever got made, and how is it staying in every new version.
Not a GNOME user, but i3. The fact that on MacOS I can't put more than 2 windows side by side to tile the screen space drives me nuts. Sadly, it's a work laptop and can't install 3rd parties to have a similar UX as i3.
You can use BetterTouchTool and configure it to tile an arbitrary amount of windows. As a bonus those windows don't have to be in full-screen mode, so you can cmd-tab quickly.
I agree that the windowing system of MacOS is an unusable mess.
Except for devs, that need to enjoy doing XML by hand for GUI layouts, as the wannabe replacement for Glade is yet to deliver, and even then it is based on Web stack instead of Gtk.
Ah those JavaScript extensions all over the place.
I am currently working on kotlin native bindings for the whole gtk/gobject/glib stack and with a small DSL/fluent api layer on top, building UI's by composing Widgets without XML is not that complicated anymore.
Furthermore documentation to learn to do XML by hand sucks a great deal. It’s admirable how beautiful the Gnome desktop have evolved which such a subpar developer experience. QML is light years ahead.
In a sense, yes. HTML and CSS (with JS?) are the best UI creation combo anyone could ever design, because so much effort has gone into making websites look nice over the decades. Millions of people and thousands of companies trying to create websites has pushed the standards really far ahead of any other UI kit.
I think eventually they'll become integrated with operating systems natively, rather than requiring webview. But that means desktop managers will become much more integrated with the web as well.
Basically I see a future where your OS is your internet browser as well as window manager, process manager, etc.
HTML and CSS are not the answer. If anything, you should be scared off by the fact that webpages require at least three languages to implement anything useful (HTML, CSS, and JS), and that doesn't count whatever you could use for a backend (PHP? I'm not a web developer, so I don't know what the cool kids use, other than maybe Node.js.)
In my book, the fewer langauges you use, the better: I will even extend this to Qt and QML (and it should be pointed out that I am passionate about using Qt and QML!). Multiple languages necessitate (often fragile) bridges to work together; plus, they take much more time and effort to learn. If you could learn one language and use it to make an app, wouldn't you jump at that instead of learning 3 or 4 languages before you could make an app?
Others have already added some great comments about the wisdom (or rather the lack thereof) of integrating a web browser into the OS. Myself? I don't wish anybody a future of bloated webapps for everything.
Fear of bloated webapps taking over the world is maybe paranoia, but well placed paranoia at that. It's not the fault of the tools though, it's the developers who have a terrible mindset. You can argue that the "system" encourages this, but I think it could just be growing pains and incompetence. We're in the teenager phase of the web, in my opinion.
I also have an argument against the "too many languages" point: What's happening in the modern web is that HTML, CSS, and JS are slowly combining. A lot of projects are moving towards TSX components (Typescript and React, Vue, Svelte, etc.) with Tailwind CSS. So: Typing, layout, scripting, and styling all in one cohesive package that can easily be abstracted into components or classes.
I think eventually there will be a singularity of sorts where we have a standard toolkit/language.
From my own experience building websites and native apps, I can move around in web tech like a god damn ninja but I hate electron. I hate the inefficiency of using a browser for the UI. Qt and Gtk feel clunky and limited compared to web, but are native. But now there's Tauri, and I think it's a large step in the right direction.
> you should be scared off... the fewer langauges you use, the better
As an engineer, I see this differently. A "language" is not simply (or always) syntax. Generally speaking, any API is already a language, and it's normal to use several APIs in one application; I'd even say, the more the merrier. Further, it is OK for an API to be represented as a DSL (domain-specific language) with a syntax of its own, so it's all good.
HTML and CSS are maybe the best UI creation combo if you only ever care about visual aesthetic design.
Any other aspects of what UI design is supposed to be about, ergonomics, usability, consistency, accessibility, internationalization is total crap in HTML/CSS. What is the taborder of your custom HTML/CSS datepicker? What will happen when I press <Esc> on your custom css popup? What is the keyboard shortcut to highlight that form field or to open that hamburger menu? How will my screenreader pronounce that hamburger symbol? Will it be the same for the next great HTML/CSS creation? Will the date format follow my locale?
HTML/CSS is great only if you are the strictly visual aesthetics kind of "designer" who only care about reproducing their napkin drawings and photoshop mockups. For all the other things one would ask of a proper UI, HTML/CSS is either useless or actively detrimental.
There were far, far better UI kits in the 90s, which is why programs were a lot more usable back then. NeXT comes to mind. It doesn't help that Linux has wasted the past decade struggling to solve the difficult problem of (checks notes) who is logged in to a graphical seession.
> Basically I see a future where your OS is your internet browser as well as window manager, process manager, etc.
I sure hope you're wrong. It's definitely happening in some respect and to me it feels like we are undoing the shift from corporate controlled, centralized computing to personal computing where users are in control that happened from the late 70's through most of the 90's.
I've posted this quote here before, but I still think about it all the time. In Ellen Ullman's excellent book Life in Code she quotes Whitfield Diffie:
> We were slaves to the mainframe! he said. Dumb terminals! That's all we had. We were powerless under the big machine's unyielding central control. Then we escaped to the personal computer, autonomous, powerful. Then networks. The PC was soon rendered to be nothing but a "thin client," just a browser with very little software residing on our personal machines, the code being on network servers, which are under the control of administrators. Now to the web, nothing but a thin, thin browser for us. All the intelligence out there, on the net, our machines having become dumb terminals again.
Windows Active Desktop and ChromeOS have shown how "great" it happens to be.
Even WebOS has support for Qt based applications, and ChromeOS had to add support for Android and GNU/Linux containers to run alongside the Chrome stack.
> In a sense, yes. HTML and CSS (with JS?) are the best UI creation combo anyone could ever design,
Maybe for a very specific sort of aesthetic optimization, but for everything else it is my opinion that modern web stack remains inferior to Visual Basic and HyperCard, both of which were easier to use for creating interfaces, and in my opinion both of which looked better.
> In a sense, yes. HTML and CSS (with JS?) are the best UI creation combo anyone could ever design, because so much effort has gone into making websites look nice over the decades
Even with all that effort it has just recently surpassed the era of WinForms. How much effort is it to customize a date picker in web tech? Layouting was completely broken for decades, only now is it semi-okay. WebComponents are also a recentish development, but they are absolutely essential.
Also, how come WYSIWYG editors are not common?! They were everywhere.
Mostly, yes. But... can you configure three-finger touchpad swiping to perform back/forward actions yet (like on just about every other platform)? Last I tried it was essentially impossible (well, short of getting into the code and re-compiling your own). The built in browser (web/epiphany) was the only app that implements three-finger swipe, but lacking extension support until very recently, wasn't really an option.
I could adapt if I exclusively used Gnome, but I regularly use a few different computers with different OSes, so some consistency would be nice. Would also make it easier for people who are switching to Gnome/Linux.
So many UIs these days try to shove some random idea of a PM into ones face recently E.g. "OneDrive this", "Adobe Cloud that" on just saving a document.
GNOME apps would simply need to show up to become a superior experience in the long term.
Great to see that they do so much more than 'just show up'.
> GNOME Circle is a fantastic collection of apps that are developed as part of the GNOME Project. Since GNOME 43 was released, ten new apps have joined:
I think I probably understand just what three of those do (Eyedropper, Lorem and Chess Clock), the rest seems indecipherable by name only. Is it really that impossible to come up with creative yet understandable names?
I actually like Gnome for it's no-nonsense names. "Web", "Email", "Calendar" etc. Often the "under the hood name" is different, e.g. Image Viewer is "internally" called "Eye of Gnome or eog" or "Videos" is internally called Totem. This is for Core, though.
Circle has - indeed- many apps with funky names. But quite often easy to find using terms you expect "edit markdown -> Apostrophe"
Agreed. Alternately renaming it to something like "Overview" or "Home" might better match new user expectations about what will happen when it is clicked.
Still lets you turn off Bluetooth with a single click, even when all you have is a Bluetooth mouse and keyboard.
Seriously, I only use USB input devices on desktops to bootstrap the system, and then unplug them and tidy the system away.
For a few times now I've clicked that button by mistake (it's there in GNOME 43, just without the ">") and promptly turned off Bluetooth and had to go fetch a USB mouse from storage, crawl under the desk, and plug it in to turn everything back on.
They are also a major proprietary hassle, whereas with Bluetooth and a Logitech K380 keyboard, which is cheap, readily available everywhere and reliable as heck I can switch between three machines at the press of a button.
I use that and an M720 Bluetooth mouse to switch between two work machines and a personal one on the same desktop, which is something you just can't do with the Logitech Logi Bolt dongle - even if you buy multiple ones.
Also, I've been using this setup with Macs for ages (that's one of the machines above), and never ran into the same problem.
Pros and cons? With the Logitech (and others) dongles you can also just plug them into a KVM and swap that way which isn't an option with Bluetooth, the device specifically has to support multiple devices.
But to use a KVM you have to have some form of commonality between devices (one of mine only has USB-C ports), plus you have to have all the cabling. And most consumer KVMs only take 2 outputs.
I'll take Bluetooth _any day_ over having another mess of cables, another box, and what boils down to less ease of use.
I use a USB C hub with my KVM for my laptop. The hub takes the KVM USB A cable and the DP cable. One cable connects the laptop and I can switch PCs with a touch of a single button. I don't like switching inputs on my monitor manually through the OSD. You can get KVMs that do 3 or 4 computers with multiple monitors, although granted they do cost more. Again, pros and cons! I find it pretty slick though.
I added a small udev rule to send a "change input" dcc command to the monitor when my USB switch connects or disconnects. The monitor has a USB-C input, and I connect the USB ports it provides to my USB switch.
Cheaper than a KVM, though it only works if the computer that has the rules is turned on (that monitor accepts commands from inactive sources).
If somebody knows a ready-made "reverse dock" (USB-C dock connector -> USB-PD out + DP in + USB device), I'm all ears!
Entirely reasonable stance. Previous story[0] about keyboard “security”: 8/10 tested were vulnerable to sniffing or injection. A handy summary table someone put together
Protocol Sniffing Injection
Plexgear Yes Yes
Rapoo Yes Yes
Logitech No Yes
Corsair Yes Yes
iiglo Yes Yes
Exibel Yes Yes
Razer No No
And what about side channels? It should be relatively easy to link timing to each key pressed, for instance, if not outright extracting the encryption key from emitted signal power fluctuations.
These approaches are not just theoretical, they have been shown to work. And that's assuming the protocol itself is not vulnerable and has been implemented correctly. Moreover, I tend to use my keyboards a long time, making it likely that a new attack becomes possible over its lifetime.
Personal question, feel free not to answer: what is your occupation that someone would find so interesting as to drag over a truckload of RF equipment near your location to try to home in on your 2.4Ghz keyboard, which operates in one of the most saturated frequency ranges known to modern man? :)
> Still lets you turn off Bluetooth with a single click, even when all you have is a Bluetooth mouse and keyboard.
Is there any system that wouldn't act the same way?
It is a UX issue though for sure, I don't disagree. While GNOME might or might not ever fix it, you could hack together a solution yourself with udev rules and a USB stick.
Something like this: (freeform, you have to adjust for your own usage)
I think something like that should work. Put that in `/etc/udev/rules.d/99-bluetooth-usb-activation.rules` and do `sudo udevadm control --reload-rules && sudo udevadm trigger`, then after that, when you plug in the USB stick into the computer, it should start bluetooth automatically again.
You could purchase some random USB stick from your local hardware/"everything" store and use that as a bluetooth activator.
That doesn't work in practice because the machine in question is under my desk, and reaching back into the USB ports entails moving it (and sometimes unplugging other things).
Or in another way - what's the value behind having a clutter-free desk if you randomly disconnect bluetooth and need to stop everything in order to reconnect it again?
That's strange. I don't have a macOS machine in front of my right now, but on Windows 10 and 11 it seems to work the same way, it allows me to turn of Bluetooth with just one click, no confirmation or anything.
Is there something particular with the GNOME Bluetooth toggle that makes you wanna click it compared to the Windows toggle?
I just upgraded my laptop from Fedora 37 to 38 Beta and the transition was painless. I have done this since Fedora 35 and the upgrade process has been excellent, far better than on Windows and macOS in fact.
For me, Fedora has surpassed the macOS "out-of-the-box" experience, and I have seen how non-techies and older people "get" the UI much faster than both Windows and macOS.
I have written https://taoofmac.com for the last 20 years, and have spent the last year progressively using GNOME and Fedora more and more.
I use the WhiteSur GTK theme (https://github.com/vinceliuice/WhiteSur-gtk-theme) so that I don't have to decipher the nuances of the standard GNOME iconography (some of it is... well, not as pretty, and definitely not as intuitive).
I love this setup, but if GNOME decides to keep breaking user themes, I will have to look elsewhere (most likely a different Fedora spin, although definitely not KDE).
I'm excited to fresh install F38 Beta if only to provide the counterexample to this "transition is painless" statement.
Every startup, I need to "killall gnome-control-center" because it loads and hangs. (I also can't access "About" and a backtrace basically tells me I'm in a glib application.)
The default file manager gets wonky scrolling up with an MX Master 3. I replaced it with something else but now double clicking a zip file will open the terminal to my home directory.
Some apps (GTK or otherwise) need workarounds or "downgrading" to X11 because of using a 30-bit monitor: Room EQ Wizard, anything based on JetBrains, OpenSCAD, GIMP. That was a lie, Room EQ Wizard doesn't work at all after installation. No workaround to be found.
Anything related to codecs or graphics has been a headache---whether aptX, H.265, VP9, WebGL, ...
-----
If these problems could simply go away, I'd be a much more productive coder.
I did the upgrade. Bluetooth setup is as painful as before: about two minutes and thirty keystrokes. You have to advance through the modals (no Ctrl+Alt+T into terminal until setup is complete) and open the Bluetooth panel. When you type <Super> + "blue", the search bar suggests a bunch of matches from the Characters app (blue heart emoji, blue circle, whatever else) and searches from a second app. If you go into Settings but not the Bluetooth panel directly, you have to Shift+Tab about ten times to select the mouse for pairing.
(I swear there was a distro that would autoconnect to the first input device it found in pairing mode, but maybe that was a pleasant dream I had.)
gnome-control-center crashed when I disabled the (nonfunctional, for me) monitor color profiles. It then crashed again when I disabled scrollbar overlay (after disabling the overlay). When I would try to re-enable scrollbar overlay, Settings would crash without re-enabling. After going through the very lengthy backtrace process (downloading symbols presumably, "sign up", "make API key"... I got a "Processing Failed" error. Then, Software crashed when disabling Flatpak.
Most amusing to me was submitting a bug report that libreport had "Misbehavior". Right now, the next morning, my Problem Reporting dialog has 18 system failures or application crashes.
Updates in the Software app are extremely unintuitive: there's a very thin progress bar built into the bottom of the Cancel button. No "time/items remaining" or "waiting for XYZ repository"--- just hope that the thing is working.
The accessibility menu shows up even if you disable "Accessibility Menu" if you have any accessibility option set, such as Large Text---which just makes sense on a monitor over 160 pixels/in. (200% scale is too extreme to be usable)
So far? Nice visuals but unimpressive stability and functionality.
And then there's Nvidia. The install took 3 hours and that's with the experience of doing it a half dozen times before. (In fairness, this time I was using Secure Boot.) In 37, the nvidia-akmod package would work. This time, for some reason, not. Wayland is still disabled by default for the Nvidia proprietary driver, which is good because when I force Wayland, I get a segfault running nvidia-settings (which runs fine on X).
I know... I know... "It's your own fault! Nvidia sucks! No, really, Linus said so himself!" except I actually want to be able to view 10 bits per pixel on multiple 4K 60+ Hz monitors without visual glitches and without sounding like Cape Canaveral on launch day. My Thinkpad has a Radeon and OLED, and that system has had at least as many pain points as the Nvidia desktop. I'm trying to use a graphics card from the GPU market leader on a computer built by the processor market leader. Both Nvidia and Intel have their own Linux distros! The fact that my Intel computer doesn't have TOSLINK support unless I install Windows or no HDR unless I install Windows or no stupid case LED customizer unless I install Windows, etc. is extremely frustrating.
Also, odd that the default F38 Beta wallpaper isn't choosable in the Wallpapers dialog. You have to manually navigate to its path, if you know where that is...
-----
One positive: Files does proper scrolling now with the MX!
And Firefox is still a distro package, which is nice because the snap not working properly with a discrete GPU is one major reason I've stopped using Ubuntu.
I can't match the experience unfortunately. F36->37, it couldn't open the default file manager for some reason, and also, the "night light" function was broken. I had to add some color calibration thingie in the UI to get it to work. Unfortunate, because the system was practically untouched, so the upgrade should have been painless. But the reality is that I can't get attached to an installation too much.
How would that work for deprecations and new API introductions? GNOME is built on top of glib/gobject/gtk and a host of other libraries which all have to be ABI-compatible in some way. And GNOME itself is used as a base platform for applications.
Having a predictable versioning scheme helps limit the number of releases for downstream application maintenance work. You can't expect applications to have the full git tree available at build or runtime to make some features conditionally available.
I have Fedora 38 beta/G44 on my laptop. I don't use my laptop too much; mostly as a terminal or browser portal for research and chilling. I don't code or compile much, and I don't access local filesystems often.
I think the navigation and UI of Gnome 44 looks great. I love the redesigned system tray - Mac-ish but functional and informative. Capable of touch-based but not ridiculously so.
I don't have the experience with the filesystem to complain about the navigation or view there. Windows has always had the best file viewer IMO - MacOS is confusing as hell and Gnome has always had limited viewing options as well.
> We are also providing our own installer images for debugging and testing features. These images are meant for installation in a vm and require GNOME Boxes with UEFI support. We suggest getting Boxes from Flathub. os.gnome.org
What if I’m on Windows and want to try out the latest GNOME before making a decision to migrate to Linux? What makes Boxes so special? Is it really so hard to make your images usable with other common virtualization tools?
Tried booting it in Hyper-V, didn’t work. I got it to work in VirtualBox (although I had to switch to UEFI and make sure to select Linux as the OS) - and it’s not a live system, but a very dumb installer.
I think they mean that GNOME contributors can use it to debug and test features. If you want to evaluate it, you can probably best try out a distribution (e.g. Ubuntu if you want to go with popular, or Fedora if you want to get closer to the way GNOME intended it) that ships it.
Linux distros ship heavily modified versions of desktop environments, especially Ubuntu. Linux distros also typically do not ship the latest version on release day.
If you're merely contemplating whether to switch to Linux in the first place, I wouldn't worry about the latest version of Gnome - they're released every six months, so these changes probably aren't going to be the decisive factor in whether or not to switch.
I'd recommend Ubuntu if you don't really know much about Linux yet, simply because it's the most popular. It sounds like you do actually already know more and are interested in Gnome specifically, in which Fedora really doesn't modify it all that much. If you really want to try the newest Gnome release in particular, then waiting a month or so should give you a Fedora release that ships it.
I ran Fedora on an old laptop plugged into my desktop peripherals and began to realize that I wasn't using my Windows desktop anymore so I installed it on the desktop. My suggestion is to install it on another machine if you have one.
Fedora 38 (currently in beta) comes out in a month and will have Gnome 44. Fedora 37 is on Gnome 43. They tend to time their releases to include the latest Gnome.
Boxes isn't that special but it is built in, easy, and doesn't require privileges like virtmanager. I like it.
Better try out your distro of choice in a VM like VirtualBox or VMware, but even then, most distros perform way differently in VMs than they will on the bare metal your system happens to be using due to the different HW (emulated or not) and drivers.
Wtf? Can't you spin up a vm yourself and go through an installer? A desktop environment is not a distro, it's nice enough already that they provide some vm images. Btw afaik gnome boxes is simply a gui for libvirt.
You just use whichever virtualization tool you like. Boxes is one that ships with gnome on many distros, so it is easy to recommend.
Nothing hard at all, I think you just confused a recommendation for a requirement.
Its GNOME Boxes and therefore the best and nothing else matters according to GNOME logic. This is the Desktop Environment in an ecosystem where the window manager draws the title bars that decided the app needs to draw the title bars so you can really easily get two title bars on a GTK app when used on not GNOME.
I like GNOME, but every time I've tried using it, it eventually begins to suffer from one issue or another (text randomly disappearing; stutters; high RAM usage). Each time I try it, I eventually get fed up and return to XFCE.
Desktop Environments are OK, but they do not give me anything I cannot already do with fvwm (or other window managers).
I am not looking forward to the day Wayland is forced upon us by various Linux Distros. Seems Red Hat type distros are marching full speed ahead to Wayland. Thus no fvwm for me (at least on Linux).
Switch to OpenBSD, you get FVWM (and CWM) by default. Or Hyperbola, where libre licenses, a BSD-like philosophy and light desktops (no XFCE, Gnome or even KDE, just Lumina and JWMKit) do a match. OFC you get FVWM and CWM too.
I'm not really sure what your objection is. The cups web ui is fine for normies. I've certainly had much more trouble with the windows printer ui than cups.
I guess the url is not obvious, but once you do it once you'll remember it forever.
File chooser grid view (ie thumbnails) - that's nice, but I'm so jaded at this point to care anymore. I've been using ubuntu on and off as my personal system for the past 15 years, it sure beats windows in many respects (typing this on MacOS). But I'm just too old to be excited by UI features that should've been here long ago and that were the standard in the software of the late 90s and early 2000s. There has also been a vocal part of the community defending the lack of thumbnails in the file picker, so this whole story has a bitter taste to it. One more thing, it feels like open source systems and UIs are developing at a glacial pace these days, either the tools used to develop open source software are sub-par or maybe there is a lack of people working on it and effort put into it. So glacial that in the age of AI, a grid view is somehow a notable feature. There seems to be less to be excited about these days and I have given up expecting great things.