FyneDesk – Linux desktop environment in Go
github.com
github.com
For full desktops: Plasma is great, but I personally dislike it for a reason I can’t put my finger on. GNOME is great once I install about 12 plugins, but then those break on updates, and I cannot stand the default GNOME experience. XFCE is “fine” but lacks some refinement. Budgie is on life support it seems. I haven’t tried enlightenment in a while. Never tried Deepin.
It's a continuation of LXDE which was absorbed into another project when LXDE was discontinued, iirc
Best obtained via LMDE
There were no things from the Mac I was missing with Unity. There are still things on the Mac that I miss from Unity (most notably, the multiple dots indicating the number of open windows, and the setting where the screen that does NOT have the default window can be subtly dimmed)
That said, this project looks like it needs a visual designer. Strong "programmer art" vibes from the lack of coherent spacing. Some consideration for padding would go a long way.
Please keep the useless touch-friendly interface as an optional addition on desktop environments. Better yet, contain it to EXCLUSIVELY for mobile apps/usage because a desktop computing environment should NOT have half the screen full of annoying whitespace! I'm trying to do work and see as much content as possible at once.
macOS Big Sur and above, GNOME3+, and even Windows 11 all have such infuriating interfaces. The frustratingly excessive padding added to the window title bars and context menu take up the whole damn screen!
Maybe on tablets and touchscreen laptops running GNOME or Win11 where you use a stylus for writing/drawing, I can sort of get it.
If you use a keyboard and mouse to do work, however, the context menus should be compact and dense so you don't have to keep scrolling and rearranging your desktop workspace to find and organize things.
I have, however, studied this exact thing. A target need to be at least 48dp in diameter to be accurately clicked the first time. The same is true for a touchscreen. Bigger is better on a touchscreen, whereas there are diminishing returns past 48dp for accuracy in clickable UIs.
Despite my interest, I am curious about the motivation for another DE. Is the primary goal to use Go and/or the Fyne library to create a functional DE? Or, is there a larger vision for this DE that Go and/or Fyne enables?
a persistent window, aka panel, is the conventional way to provide that.
it's simply a conceptual extension of the 'prompt' into a gui environment. nothing has surpassed the panel concept in the past forty years that it has been dominant.
i expect the 'panel' is the primary feature vehicle of any desktop environment - they are nearly synonymous. nearly every window manager also contains support for panels, even if they are not part of a desktop environment that specifically provides them. what do you use?
As for the full screen bug you experienced, I don’t recall ever seeing that but Windows 9x had a plethora of weird bugs so it wouldn’t surprise me if you had issues.
and microsoft does all kinds of weird shit nobody has an explanation for.
Would love to see something where this was the principle? Very little philosophy or opinion on how individuals do their own desktop, emphasis on "you should be able to add and take things off easily as you want." The metaphor being, you know, an actual physical desktop. It's an EMPTY table. Put what you want on it and nothing you don't.
(Like openbox perhaps, just with a little more friendliness towards well integrated panels etc. Which might be construed as "opinion" -- but seems like most WMs FORCE a lot of the things on you)
I'm more of a 'install i3wm, change a few things in the config, done' kind of person these days.
That entire field is somewhat unserious because it conflates engineering and actual usability with fashion, and often can't distinguish between the two.
That's exactly why I'm saying focus more on the modularity.
1. Panels are fixed in place and don't have title bars etc. They're the only window type like that so you'd have to go to extra effort to make a UI to "lock" windows and unlock them or whatever.
2. If you delete the launcher then how do users start apps? Some old DEs let you right click the desktop but that's not very discoverable.
Much easier just to have settings somewhere do let you say which side (if any) you want panels on.
Really though, Windows, Gnome, Plasma, MacOS, they're all basically the same and all feel so cumbersome and limiting. Desktop metaphors have been stalled for so long, and when Apple took a shot at something new with Stage Manager they managed to downgrade the experience even further. I'm so disappointed by the state of things.
It's especially annoying in the world of tiling window managers. There are few options for a desktop environment with a solid tiling WM. Regolith is one (based on i3 / Sway), and I thank those who develop it, but it's patching together a bunch of independently developed tools and you can tell when you try to change some things and have to deal with disparate config files.
I think its the vim effect where the userbase just has a high demand for customization and don't mind a lot more set up and configuration to get there.
One particular problem is theming: I'd like to set a theme in a single place for all the elements of my desktop environment. Not sure if it's fair to expect a bunch of independent opinionated developers who like to use weird graphics toolkits (when not drawing everything by themselves) to come up with a unified solution.
Another problem is the shortcuts. It's the same problem in (neo)vim when you build your awesome IDE-like setup based on 53 unrelated packages: you end up with an illogical, overlapping soup of keybindings. I think you need someone to take charge of unifying things to provide a coherent experience.
Maybe with net/rpc, D-Bus or gRPC, replacing the use cases of dynamic code loading from those environments.
Literally.. the machine has 40 fast cores and 384GB of very fast memory. Wow, just.. kind of nauseating. It could be a case study, I'm sure, because this exceeds even the compiler bloat of Electron.
This means one rather slow compile to start and then it’s back to super fast!
You may not like material design (I personally could take it or leave it), but visual design itself matters.
But I can't really tell how TFA is Material Design, it does not really look like any of the versions.
See:
Just like a hole in the road matters. /s
I just can't wait for people "designing" material design to grow old. Then they can tell us again how great Material Design is.