Super cool people maintain and expand things like KDE 3 and Gnome 2 (Trinity, Mate)!
However personally I don't feel the need anymore. Gnome 3 and KDE 4 are both quite usable. Which is a stark contrast from when Gnome 3 and KDE 4 came out. They were both nothing more then glorified beta's. The first release of Gnome 3 was basically like the first Windows 10 release with the Metro UI; Totally unusable. Not that surprising since both Gnome 3 and Win 10 were created on a "tablet first" (or at least "tablet prominent") mindset.
I use Mate and I think it's misleading to characterize it as a reheated Gnome 2. That's how it started, but the project has come a long way since then. Mate now uses GTK3 instead of GTK2. It supports GPU-accelerated compositing and Hi-DPI displays. Mate is a well-maintained modern desktop environment, not a historical curiosity.
It also plays nicely with tiling WMs; you can replace the default with i3 or Awesome or whatever and enjoy mate-settingsd and having a nice GTK3 panel that looks nice, but also tiling windows.
Meh, depends on what you define as usable. KDE 3 ran fluently on my 2010 HP EliteBook 8540w but KDE 4 runs like a dog and you get nothing extra for it.
> Not that surprising since both Gnome 3 and Win 10 were created on a "tablet first" (or at least "tablet prominent") mindset.
For Gnome I hear this claim often made (mostly by people who don't like Gnome), but is that actually backed by facts?
Unlike Microsoft or Canonical (with their Unity), Redhat never seemed to have any particular mobile/touch aspirations, and so far I haven't heard any mobile adaptations of Gnome anywhere, like we have Plasma Mobile from the KDE camp.
phosh is not a first-party Gnome project, but it's a shell which essentially runs near-unmodified Gnome 3 apps full-screen on a phone display. I've tried it on a PinePhone, and the apps mostly work (but with some issues related to touch, small screen sizes, and orientation).
Look at GNOME, notice the giant paddings, title bards see look at the checkboxes,scrollbars etc.
Also is not a secret that GNOME developers considered at least touch screens if not tablet when designing the GUI.
IMo an informative video on GTK and Qt , it explains what issues you can have with GTK, is about the subsurface application(created by Linus Torwalds that moved to Qt because GTK is bad and even Linus could not get good answers from Red Hat guys) https://www.youtube.com/watch?v=ON0A1dsQOV0
KDE (Plasma) 5 (i guess you meant that) is usable but i still think KDE 3 was the high point of KDE. After all the flashy stuff was introduced in KDE 4 things went downhill fast. Similarly with Gnome, Gnome 2 was the best series, Gnome 3 was awful.
It seems somewhere around the late 2000s both projects got very nice and usable desktop environments, couldn't figure out how to "evolve" next (because apparently software just can't not evolve) and started throwing shit at the wall to see what sticks. And unsurprisingly, a lot of it did stick.
To its credit, KDE 5 does seem to try to clean up the mess a bit though in that over the years it is the only between the two that doesn't compulsively make me want to remove it from any Linux machine i get my hands on (and can remove stuff :-P).
I still find it extremely buggy. I've been hearing these "KDE 5 is fine now" for the past few months in comments, podcasts, blog posts, etc. Yesterday I gave in and decided to try the latest release. After logging in, KDE greeted me with a plasma crash. In the next few hours of light usage, some more crashes followed.
It's working as it always has been for me, since the early days of 4. It's pretty much dead for me now.
For the record, this has been happening on many machines, and I don't use nvigia GPUs and similar stuff.
Also, I try to report bugs in software that I use, but KDE has the appearance of software which receives no QA at all. It's one thing to report a bug in sway or some XFCE application which crashes maybe once in a few years, and other to use something which crashes a dozen times a day. I'd be filling bugs all day long.
> Yesterday I gave in and decided to try the latest release. After logging in, KDE greeted me with a plasma crash. In the next few hours of light usage, some more crashes followed.
What distro? New or installed on top of something else?
I have run KDE Neon since a few years ago om two (admittedly high end) laptops and it has been very smooth.
> I have run KDE Neon since a few years ago om two (admittedly high end) laptops and it has been very smooth.
I have been using always-latest KDE5 with Manjaro on a very old (Core 2 Duo, Intel graphics) laptop for some years already and it's perfect. Less than 5 crashes a year.
KDE4 would crash as soon as I attached an extra display. KDE5 also crashed often when it was immature. Now it works like a breeze. As somebody who has watched (and tried regularly) KDE since the inception of KDE3, I can say KDE has finally reached maturity (if not perfection) about 1.5 years ago.
The KDE bug reporting process is a huge pain, too. I discovered that Konsole crashes when unlimited scrollback is enabled and /tmp becomes full. Submitting the crash report requires answering a pop quiz, and if any of your answers are unsatisfactory then you get a message that your report is not useful. Navigate that gauntlet and then you are asked to create an account and write a bug report, and your crash dump becomes public.
Good software needs a really easy way to report bugs and get back to work. And people to triage the reports. I do not work for KDE, but the bug reporting process treats me like I do.
> After logging in, KDE greeted me with a plasma crash. In the next few hours of light usage, some more crashes followed.
This has been my experience since KDE 4 too pretty much 4 out of 5 attempts i made to check out KDE, including trying KDE Neon (which included weird visual glitches that i never saw before in any desktop environment).
Though FWIW my latest attempt is openSUSE Tumbleweed running on my GPD Win 1 (which isn't exactly a mainstream device, though it does have an Intel integrated GPU which has open source drivers, which helps a bit) and it seems to work fine there. I do very light use of it however, mainly launching games and such - normally i'd be using Window Maker, but GPD Win 1 needed a very recent kernel to boot and also RandR to be configured to rotate the screen (the 'native' screen is actually in portrait mode but the computer uses it in landscape) and i didn't want to bother with that at the moment i installed it so i grabbed the first non-Gnome-based rolling distribution that had official support for some form of desktop environment that i expected to have a GUI application to setup rotation... which basically meant KDE :-P.
Note that i was never a big fan of KDE in general, i just happen to like KDE 3 more than the other versions (though TBH i haven't used KDE 2 at all and while i do like how KDE 1 looked and behaved - it was the first Linux environment i ever used - it was a bit too rough, though that was also because of Qt1 too) and i used it as a second environment in some computers.
But the two main reasons i think KDE 5 is worse than KDE 3 is that as a UI it follows the same pattern as Gnome 3 (and Windows 10's "modern" apps, sadly) for wasting screen space everywhere and it does feel like it has a lot of moving parts everywhere which aren't exactly fitting perfectly with each other even though they are made for each other (i know that KDE 3 was probably also similar, but it always felt solid and stable - i wonder if breaking kdelibs or whatever it was called to a myriad of tiny little libraries that all applications and each other needing to be there anyway made things more brittle, though that might be my bias from whenever i try to install a KDE application in a non-KDE environment and i get a dependency vomit from the package installer).
I might sound too hard on KDE 5, but TBH if i absolutely had to choose between it and Gnome and had no other option (or the other option was a tiling WM - which i never liked), then i'd use KDE. It isn't bad, i'd still install it on a computer i do not want to fiddle around with much.
I would say if you like KDE 3, you liked KDE 2 as well. KDE 2 was the tick of the tick-tock of KDE 2/3. KDE 5 is the tock of the tick-tock of KDE 4/5. At least that is how I'd describe it (Intel been tock-tock-tock'ing the past years notwithstanding).
I have fond memories of KDE 2/3 and GNOME 2. I also ran KDE 1 and GNOME 1 back in the days (initially GNOME cause of RedHat Linux and the license issues with KDE). If you compare them to Windows 9x and early NT they were quite good.
I think that with the exception of the kernel being a layer on top of MS-DOS (but that would be comparing with Linux not with KDE), Win9x was better technically than KDE1.
For example I do not think to this day there is anything tech-wise like COM with all of its related tech like OLE (which allows documents to embed objects exposed by other applications which themselves can run in 'embedded mode' to edit said embedded objects), ActiveX (which in a similar way allows applications to make use of reusable controls that can describe themselves), Automation (which allows applications to expose scriptable interfaces that can be called by any other application), Active Scripting (which allows applications to embed scripting languages which themselves are registered with the system so that, e.g., you can install a scripting engine for Python which then can be used to script any application that uses Active Scripting even if that application has no idea about Python) and a bunch of other things.
(note that DBus does provide the 'expose interfaces' part, but it is only about that whereas COM is a more generic technology that was used for all of the above)
Yeah i know about KParts, however KParts is more limited and the tech is only about KParts whereas COM is used for a lot of things, including even stuff unrelated to the GUI (e.g. Active Scripting). In addition COM defines a binary interface for accessing objects that can be supported by any programming language that can support structs and pointers whereas KParts is only for C++.
Windows 9x was a single user operating system; Linux and KDE/GNOME were and are not. Windows 9x is an absolute security nightmare, including ActiveX. Therefore, the comparison is moot (though ironically enough, X is also a security nightmare).
ActiveX is still available on Windows 10, it has nothing to do with security aside from being shoved into IE4 as a way for Microsoft to lock the web into Wintel back when Java's crossplatform-ness was seen as threatening to their desktop dominance - and thus gaining a bad name because it really wasn't designed for such a use.
ActiveX is just a standardized way to embed controls, in practice it isn't any difference than any desktop application that has support for plugins nor any less (or more) safer than that. COM in general is just a standardized form of a plugin system with a bunch of predefined interfaces (or binary protocols if you want) for applications that weren't written with each other in mind to still be able to work together and share functionality.
I see the Plasma team decided to keep the old ugly designs. Look at that Trinity Control Center window. More than a decade after and everything still feels the same.
Things like that make me thing this can be arguably the biggest waste of effort of all FOSS development, all because "fear of the new" - Trying to keep the huge monstrosity of Qt3 which EOL was more than 10-15 years ago, the whole DE that was built on it and all of their apps, even developing new ones... Just for the appeal of few people with the nostalgia for that kind of GUIs. One could argue KDE5 is as stable or even more than KDE3 ever was, and less buggy - not even mentioning the ability to replicate that interface with a few clicks and some fiddling.
That being said, Cinnamon and MATE were produced because of the "fear of the new". While it has been a while since I've used MATE, Cinnamon is doing fairly well.
It's also worth noting that people have different workflows and tastes. One of the perks of FLOSS is the ability for people to choose what works for them. I would hardly call it a waste of effort because of that.
I'll take predicable over shiny any day. I've not seen an interface that has increased my productivity in decades. I'm not afraid of new things but the best "new" interfaces I've seen are just minor tweaks to what's been working since win95. I'm waiting for something truly unique that's an actual improvement. Until then, I'm still using xfce with all the candy turned off.
Fear of new UIs is justified. The last time anyone did serious usability engineering and published results and methods was Windows 95 (or at least, that was the last round I've seen).
I'm afraid of wasting my time learning totally new interactions that are less useful and get changed again in 1-4 years. I would rather cling to my 'legacy' interactions that keep working for as long as I can, than to keep churning.
I was never a KDE person, but GNOME 2 was getting close to having everything right before GNOME 3 won the race to ruin desktops.
Gnome 2 was nice, but it also lacked serious usability features that were present elsewhere. Windows exposé (overview), a global menu, and window tiling/snapping were just a few.
They could have continued to extend the desktop with these features, but chose to reevaluate the desktop instead. Part of the reevaluation was GTK 3, and part of it was CSS for Shell, I think. I agree it could have been better. I don't agree that the rich choice between DEs (and richness within each) is worse today than it was 15 years ago.
> Gnome 2 was nice, but it also lacked serious usability features that were present elsewhere. Windows exposé (overview), a global menu, and window tiling/snapping were just a few.
I'm fairly certain these were easily available as plugins on Gnome 2 with Compiz.
Gnome 2 had "expose", though you had to run a compositor to use it.
There was a patch for a global menu but IIRC the Gtk developers blocked that feature so it never went anywhere. Does Gnome support a global menu nowadays?
Window snapping was supported as well - many window managers support that for years.
Also note that Gnome 2 allowed you to use any window manager you wanted, so even if Metacity (the default one) didn't do the trick for you, you could use some other window manager (and it wasn't uncommon for people to do that).
My car doesn't require me to learn new pedal placements every year, my steering wheel doesn't randomly turn into a pair of levers, the speedometer doesn't one day decide it's too old school and rearrange itself. I would never put up with random unrequested changes to my car. Or a hammer, or a screwdriver. Why should I put up with random changes to my computer? These are all just tools I use to do my real work.
To be fair, it is very difficult to find car without touch screen today. Or with power switch that physically cuts power to the engine, fuel pump and ignition. Or without GPS tracker...
The underlying tech might be different but the interface is mostly the same. After all you can have a Windows 95-like environment without actually using MS-DOS as your kernel.
Maybe not quite 1995, I’m seeing more of an unholy alchemical mixture of Aqua and Windows 98. Windows 98 was an iteration of Windows 95 which was heavily inspired by NeXTSTEP. Aqua was the theme Apple eventually slapped on the WindowServer that replaced Display PostScript when they renamed Rhapsody née OPENSTEP née NeXTSTEP to Mac OS X.
The long and the short of it is, this is why incest is bad.
My thoughts exactly, I understand why people long for the old text UIs, but this sort of thing was never pleasant nor cutting edge. It's like they're aiming at maintaining the first terrible skins people were creating after seeing the Aqua demo and missing the point of it.
Except the screenshot is nothing like a KDE 2 default installation looked back in the days. Different colours in taskbar? Nope, was homogeneous in KDE 1/2/3/x (the way it is now is unnecessarily distracting). Taskbar on left and bottom? Nope. There was one the bottom, still is. Dock like macOS? Nope, wasn't there by default.
Also, back in the early days of KDE and GNOME you had applications using all kind of toolkits. A DE providing applications with one sole toolkit (be it Qt or Gtk) was a novel direction in the Linux world.
There seems to be an irrational fear of shipping old toolkits (Qt3)
Heck Ubuntu even removed Qt4 from their repos, when I wanted to run an older third party app that hadn't been updated to Qt5 yet, I had to install a PPA with Qt4.
Not shipping a large amount of complex software with few users that went officially EoL long ago and is maintained by a skeleton crew, is hardly irrational.
I think I remember looking at the project page a while ago and thinking, huh! What is the relevance of this project? While it is always great to see Linux DE's/projects living on, this is just another project which sadly fragments the Linux desktop ecosystem even further. I agree, we need more Linux apps (Before electron apps take over the entire Linux DE), not distro's or DE's (checkout https://makealinux.app/). And if you are still adamant on creating a new DE, make it interoperable between other DE's. Remember https://www.freedesktop.org/wiki/ ?
Also, things have changed:
- KDE 5 plasma is lighter than GNOME.
- GNOME became a project with zero regards to the community. But is breaking less since the last few releases.
- You have so many "traditional DE's". XFCE, Mate, Solus DE, Cinnamon, Deepin DE.
I agree with someone who have said that none of the new GUI have made him/her/they productive. My progression for productivity through DE's has been GNOME for 10ish years, KDE 5 for 1.5, Another 1.5 years in XFCE and now I use i3wm. Looks like you focus more on functionality as you become a mature user and less on the bells and whistles.
- Moved from GNOME after using from 3.0 till 3.22 because of crashing. It is better now.
- Moved from KDE because every update on my rolling release distro was a panic attack due to the bizzilion dependencies each software depends on. Not to mention it could be a security disaster. just install Okular or Korganizer to see what I am talking about. Krita has been the only exception I have noticed.
- Window managing has been the bottleneck in XFCE, it is the best for a rolling release distribution. So now I use i3wm. I only see a gradual transition to Swaywm now since Xorg is breathing it's last few, so I will have to move from i3wm.
So I ask again, what is the point of Trinity? What is it bringing? We need to start seeing Linux as the future since big corporates are moving to Linux more. I would like to see Trinity resources going to KDE which is so much better now. But hey, that is just me.
People who like KDE 3 being able to use KDE 3 with fixes and stuff.
> this is just another project which sadly fragments the Linux desktop ecosystem even further
This doesn't really fragment the desktop since any desktop application that works under X11 will work on it. Of course if said application makes assumptions about what is underneath then it might face trouble if said assumptions are not met, but that is really a problem with the application (and in practice usually workaround-able).
> People who like KDE 3 being able to use KDE 3 with fixes and stuff.
Fair enough.
>This doesn't really fragment the desktop since any desktop application that works under X11 will work on it. Of course if said application makes assumptions about what is underneath then it might face trouble if said assumptions are not met, but that is really a problem with the application (and in practice usually workaround-able).
I meant fragmentation of unnecessary DE/Distro's which keeps reinventing the wheel instead of working with another project doing the same thing but very well.
In the case of Trinity DE, I agree there was a need initially when the KDE moved on leaving a huge number of users dry, but this should've been a temporary project. Or maybe evolve like the MATE project has (which still could've been avoided if they collaborated with XFCE instead of creating the MATE spinoff). This doesn't seems to be the case here. But yeah, I haven't test driven this RC of Trinity DE yet and my last test drive or Trinity DE was an year old, maybe things are better. Still not sure if this benefits Linux as a whole but I guess that is not my call.
> I meant fragmentation of unnecessary DE/Distro's which keeps reinventing the wheel instead of working with another project doing the same thing but very well.
I'm willing to bet that the people who want/use/develop this do not believe that there is another project doing the same thing - even if it might not appear like that to everyone (i mean, you read comments about people wondering why there are so many distributions and why not everyone works on a single one to make it the best :-P).
> In the case of Trinity DE, I agree there was a need initially when the KDE moved on leaving a huge number of users dry, but this should've been a temporary project.
Why temporary? The developers and users of TDE may prefer the way KDE3 does things or Qt3 or whatever that wouldn't work with anything that came after.
MATE is an interesting case - IMO MATE should have forked Gtk+ 2 instead of switching to Gtk+ 3 because Gtk+ 3 has a different "feel" to it that -again, IMO- is much worse (and feels a bit sluggier) than Gtk+ 2. This would also help the ton of other projects that at the time were still stuck in Gtk+ 2 and would waste a lot of time down the road to converting their code to Gtk+ 3 while also preserving Gtk+ 2 as a stable API for GUI applications that was still in widespread availability.
> I'm willing to bet that the people who want/use/develop this do not believe that there is another project doing the same thing - even if it might not appear like that to everyone (i mean, you read comments about people wondering why there are so many distributions and why not everyone works on a single one to make it the best :-P).
Haha. People are harder to deal with. It's just easier to fork! ;)
> Why temporary? The developers and users of TDE may prefer the way KDE3 does things or Qt3 or whatever that wouldn't work with anything that came after.
I agree. While I wouldn't be a fan of it, I can understand changes are harder to accept for people (including me sometimes.)
> MATE is an interesting case - IMO MATE should have forked Gtk+ 2 instead of switching to Gtk+ 3 because Gtk+ 3 has a different "feel" to it that -again, IMO- is much worse (and feels a bit sluggier) than Gtk+ 2. This would also help the ton of other projects that at the time were still stuck in Gtk+ 2 and would waste a lot of time down the road to converting their code to Gtk+ 3 while also preserving Gtk+ 2 as a stable API for GUI applications that was still in widespread availability.
> Of course it is too late now.
Never thought about it this way. I totally agree with you. It was GTK2 which should've been forked and maintained.
55 comments
[ 4.1 ms ] story [ 107 ms ] threadHowever personally I don't feel the need anymore. Gnome 3 and KDE 4 are both quite usable. Which is a stark contrast from when Gnome 3 and KDE 4 came out. They were both nothing more then glorified beta's. The first release of Gnome 3 was basically like the first Windows 10 release with the Metro UI; Totally unusable. Not that surprising since both Gnome 3 and Win 10 were created on a "tablet first" (or at least "tablet prominent") mindset.
For Gnome I hear this claim often made (mostly by people who don't like Gnome), but is that actually backed by facts?
Unlike Microsoft or Canonical (with their Unity), Redhat never seemed to have any particular mobile/touch aspirations, and so far I haven't heard any mobile adaptations of Gnome anywhere, like we have Plasma Mobile from the KDE camp.
Also is not a secret that GNOME developers considered at least touch screens if not tablet when designing the GUI.
IMo an informative video on GTK and Qt , it explains what issues you can have with GTK, is about the subsurface application(created by Linus Torwalds that moved to Qt because GTK is bad and even Linus could not get good answers from Red Hat guys) https://www.youtube.com/watch?v=ON0A1dsQOV0
It seems somewhere around the late 2000s both projects got very nice and usable desktop environments, couldn't figure out how to "evolve" next (because apparently software just can't not evolve) and started throwing shit at the wall to see what sticks. And unsurprisingly, a lot of it did stick.
To its credit, KDE 5 does seem to try to clean up the mess a bit though in that over the years it is the only between the two that doesn't compulsively make me want to remove it from any Linux machine i get my hands on (and can remove stuff :-P).
Sadly XFCE seems to be heading down the same way.
It's working as it always has been for me, since the early days of 4. It's pretty much dead for me now.
For the record, this has been happening on many machines, and I don't use nvigia GPUs and similar stuff.
Also, I try to report bugs in software that I use, but KDE has the appearance of software which receives no QA at all. It's one thing to report a bug in sway or some XFCE application which crashes maybe once in a few years, and other to use something which crashes a dozen times a day. I'd be filling bugs all day long.
What distro? New or installed on top of something else?
I have run KDE Neon since a few years ago om two (admittedly high end) laptops and it has been very smooth.
I have been using always-latest KDE5 with Manjaro on a very old (Core 2 Duo, Intel graphics) laptop for some years already and it's perfect. Less than 5 crashes a year.
KDE4 would crash as soon as I attached an extra display. KDE5 also crashed often when it was immature. Now it works like a breeze. As somebody who has watched (and tried regularly) KDE since the inception of KDE3, I can say KDE has finally reached maturity (if not perfection) about 1.5 years ago.
Good software needs a really easy way to report bugs and get back to work. And people to triage the reports. I do not work for KDE, but the bug reporting process treats me like I do.
This has been my experience since KDE 4 too pretty much 4 out of 5 attempts i made to check out KDE, including trying KDE Neon (which included weird visual glitches that i never saw before in any desktop environment).
Though FWIW my latest attempt is openSUSE Tumbleweed running on my GPD Win 1 (which isn't exactly a mainstream device, though it does have an Intel integrated GPU which has open source drivers, which helps a bit) and it seems to work fine there. I do very light use of it however, mainly launching games and such - normally i'd be using Window Maker, but GPD Win 1 needed a very recent kernel to boot and also RandR to be configured to rotate the screen (the 'native' screen is actually in portrait mode but the computer uses it in landscape) and i didn't want to bother with that at the moment i installed it so i grabbed the first non-Gnome-based rolling distribution that had official support for some form of desktop environment that i expected to have a GUI application to setup rotation... which basically meant KDE :-P.
But the two main reasons i think KDE 5 is worse than KDE 3 is that as a UI it follows the same pattern as Gnome 3 (and Windows 10's "modern" apps, sadly) for wasting screen space everywhere and it does feel like it has a lot of moving parts everywhere which aren't exactly fitting perfectly with each other even though they are made for each other (i know that KDE 3 was probably also similar, but it always felt solid and stable - i wonder if breaking kdelibs or whatever it was called to a myriad of tiny little libraries that all applications and each other needing to be there anyway made things more brittle, though that might be my bias from whenever i try to install a KDE application in a non-KDE environment and i get a dependency vomit from the package installer).
I might sound too hard on KDE 5, but TBH if i absolutely had to choose between it and Gnome and had no other option (or the other option was a tiling WM - which i never liked), then i'd use KDE. It isn't bad, i'd still install it on a computer i do not want to fiddle around with much.
I have fond memories of KDE 2/3 and GNOME 2. I also ran KDE 1 and GNOME 1 back in the days (initially GNOME cause of RedHat Linux and the license issues with KDE). If you compare them to Windows 9x and early NT they were quite good.
For example I do not think to this day there is anything tech-wise like COM with all of its related tech like OLE (which allows documents to embed objects exposed by other applications which themselves can run in 'embedded mode' to edit said embedded objects), ActiveX (which in a similar way allows applications to make use of reusable controls that can describe themselves), Automation (which allows applications to expose scriptable interfaces that can be called by any other application), Active Scripting (which allows applications to embed scripting languages which themselves are registered with the system so that, e.g., you can install a scripting engine for Python which then can be used to script any application that uses Active Scripting even if that application has no idea about Python) and a bunch of other things.
(note that DBus does provide the 'expose interfaces' part, but it is only about that whereas COM is a more generic technology that was used for all of the above)
[0] https://techbase.kde.org/Development/Tutorials/Using_KParts
[1] https://en.wikipedia.org/wiki/KDE_Platform_4#KParts
ActiveX is just a standardized way to embed controls, in practice it isn't any difference than any desktop application that has support for plugins nor any less (or more) safer than that. COM in general is just a standardized form of a plugin system with a bunch of predefined interfaces (or binary protocols if you want) for applications that weren't written with each other in mind to still be able to work together and share functionality.
...I thought that was the entire point of trinity?
https://wiki.trinitydesktop.org/images/4/44/TDE-screen1.png
Things like that make me thing this can be arguably the biggest waste of effort of all FOSS development, all because "fear of the new" - Trying to keep the huge monstrosity of Qt3 which EOL was more than 10-15 years ago, the whole DE that was built on it and all of their apps, even developing new ones... Just for the appeal of few people with the nostalgia for that kind of GUIs. One could argue KDE5 is as stable or even more than KDE3 ever was, and less buggy - not even mentioning the ability to replicate that interface with a few clicks and some fiddling.
But whatever makes them happy.
It's also worth noting that people have different workflows and tastes. One of the perks of FLOSS is the ability for people to choose what works for them. I would hardly call it a waste of effort because of that.
I'm afraid of wasting my time learning totally new interactions that are less useful and get changed again in 1-4 years. I would rather cling to my 'legacy' interactions that keep working for as long as I can, than to keep churning.
I was never a KDE person, but GNOME 2 was getting close to having everything right before GNOME 3 won the race to ruin desktops.
They could have continued to extend the desktop with these features, but chose to reevaluate the desktop instead. Part of the reevaluation was GTK 3, and part of it was CSS for Shell, I think. I agree it could have been better. I don't agree that the rich choice between DEs (and richness within each) is worse today than it was 15 years ago.
I'm fairly certain these were easily available as plugins on Gnome 2 with Compiz.
There was a patch for a global menu but IIRC the Gtk developers blocked that feature so it never went anywhere. Does Gnome support a global menu nowadays?
Window snapping was supported as well - many window managers support that for years.
Also note that Gnome 2 allowed you to use any window manager you wanted, so even if Metacity (the default one) didn't do the trick for you, you could use some other window manager (and it wasn't uncommon for people to do that).
The long and the short of it is, this is why incest is bad.
Also, back in the early days of KDE and GNOME you had applications using all kind of toolkits. A DE providing applications with one sole toolkit (be it Qt or Gtk) was a novel direction in the Linux world.
Heck Ubuntu even removed Qt4 from their repos, when I wanted to run an older third party app that hadn't been updated to Qt5 yet, I had to install a PPA with Qt4.
What’s xml?
(I have a dream - noxml jr)
Also, things have changed:
- KDE 5 plasma is lighter than GNOME.
- GNOME became a project with zero regards to the community. But is breaking less since the last few releases.
- You have so many "traditional DE's". XFCE, Mate, Solus DE, Cinnamon, Deepin DE.
I agree with someone who have said that none of the new GUI have made him/her/they productive. My progression for productivity through DE's has been GNOME for 10ish years, KDE 5 for 1.5, Another 1.5 years in XFCE and now I use i3wm. Looks like you focus more on functionality as you become a mature user and less on the bells and whistles.
- Moved from GNOME after using from 3.0 till 3.22 because of crashing. It is better now.
- Moved from KDE because every update on my rolling release distro was a panic attack due to the bizzilion dependencies each software depends on. Not to mention it could be a security disaster. just install Okular or Korganizer to see what I am talking about. Krita has been the only exception I have noticed.
- Window managing has been the bottleneck in XFCE, it is the best for a rolling release distribution. So now I use i3wm. I only see a gradual transition to Swaywm now since Xorg is breathing it's last few, so I will have to move from i3wm.
So I ask again, what is the point of Trinity? What is it bringing? We need to start seeing Linux as the future since big corporates are moving to Linux more. I would like to see Trinity resources going to KDE which is so much better now. But hey, that is just me.
People who like KDE 3 being able to use KDE 3 with fixes and stuff.
> this is just another project which sadly fragments the Linux desktop ecosystem even further
This doesn't really fragment the desktop since any desktop application that works under X11 will work on it. Of course if said application makes assumptions about what is underneath then it might face trouble if said assumptions are not met, but that is really a problem with the application (and in practice usually workaround-able).
Fair enough.
>This doesn't really fragment the desktop since any desktop application that works under X11 will work on it. Of course if said application makes assumptions about what is underneath then it might face trouble if said assumptions are not met, but that is really a problem with the application (and in practice usually workaround-able).
I meant fragmentation of unnecessary DE/Distro's which keeps reinventing the wheel instead of working with another project doing the same thing but very well.
In the case of Trinity DE, I agree there was a need initially when the KDE moved on leaving a huge number of users dry, but this should've been a temporary project. Or maybe evolve like the MATE project has (which still could've been avoided if they collaborated with XFCE instead of creating the MATE spinoff). This doesn't seems to be the case here. But yeah, I haven't test driven this RC of Trinity DE yet and my last test drive or Trinity DE was an year old, maybe things are better. Still not sure if this benefits Linux as a whole but I guess that is not my call.
I'm willing to bet that the people who want/use/develop this do not believe that there is another project doing the same thing - even if it might not appear like that to everyone (i mean, you read comments about people wondering why there are so many distributions and why not everyone works on a single one to make it the best :-P).
> In the case of Trinity DE, I agree there was a need initially when the KDE moved on leaving a huge number of users dry, but this should've been a temporary project.
Why temporary? The developers and users of TDE may prefer the way KDE3 does things or Qt3 or whatever that wouldn't work with anything that came after.
MATE is an interesting case - IMO MATE should have forked Gtk+ 2 instead of switching to Gtk+ 3 because Gtk+ 3 has a different "feel" to it that -again, IMO- is much worse (and feels a bit sluggier) than Gtk+ 2. This would also help the ton of other projects that at the time were still stuck in Gtk+ 2 and would waste a lot of time down the road to converting their code to Gtk+ 3 while also preserving Gtk+ 2 as a stable API for GUI applications that was still in widespread availability.
Of course it is too late now.
Haha. People are harder to deal with. It's just easier to fork! ;)
> Why temporary? The developers and users of TDE may prefer the way KDE3 does things or Qt3 or whatever that wouldn't work with anything that came after.
I agree. While I wouldn't be a fan of it, I can understand changes are harder to accept for people (including me sometimes.)
> MATE is an interesting case - IMO MATE should have forked Gtk+ 2 instead of switching to Gtk+ 3 because Gtk+ 3 has a different "feel" to it that -again, IMO- is much worse (and feels a bit sluggier) than Gtk+ 2. This would also help the ton of other projects that at the time were still stuck in Gtk+ 2 and would waste a lot of time down the road to converting their code to Gtk+ 3 while also preserving Gtk+ 2 as a stable API for GUI applications that was still in widespread availability. > Of course it is too late now.
Never thought about it this way. I totally agree with you. It was GTK2 which should've been forked and maintained.