will freebsd15 have default desktop?

Off the top of my head, I do recall KDE 3 having a kuser utility... Yeah, it was not the most reliable thing, but that did not make or break my DE experience.

To me, a DE needs to offer visually convenient ways to organize files - properly working drag-and-drop, sorting files by date or something else, looking at stuff side-by-side - and yeah, not crash if I do something weird with the mouse.
For a DE I do feel a user manager (and network, and service mgr) needs to be present. Otherwise it is basically just a glorified window manager and file manager. (Which is fine, many users use just that!).

Basically the KDE (and Gnome) in ports are so stripped down that they are basically just WMs at this point.
 
No there is no need for a default DE. FreeBSD should be minimalist and allow the user to decide to install it separately.
I agree. When I first starting using FreeBSD it was nice to install things rather than remove them after the system had installed. On Linux I spent more time removing installed by default programs longer than it takes to set up a FreeBSD.
 
For a DE I do feel a user manager (and network, and service mgr) needs to be present. Otherwise it is basically just a glorified window manager and file manager. (Which is fine, many users use just that!).

Basically the KDE (and Gnome) in ports are so stripped down that they are basically just WMs at this point.
Ideally what one wants is a full application suite. Remember DEs were originally created to give the user a CDE-like ability when CDE was still proprietary. And CDE never had such capability because the user (non-root) couldn't manage users, network, disks, filesystems, or anything else. Systems at the time were mult-user, like the Sun 2000 was in our basement at $JOB.

But today, DEs endeavor to be more like Windows, giving the user the ability to not only do job specific work but also manage the machine. Linux tries to do this. The problem here is that the Linux tools in the various DEs are Linux-specific, calling non-POSIX Linux system calls not in FreeBSD -- FreeBSD has its own system calls that duplicate what Linux does (or Linux duplicates FreeBSD's syscalls functions) but differently. In many cases the FreeBSD approach is superior.

Case in point. Another Open Source project requested FreeBSD change it's timekeeping implementation to that of Linux to avoid their implementing the same functions twice in their code (with #ifdefs). A senior FreeBSD kernel dev told them politely our approach was superior and that we wouldn't implement the Linux approach.

This is the problem with Linux admin utilities in the various DEs. In some cases we'd need to rip out our approach and replace it with the Linux approach. I suppose if one really needs a Windows-like experience on a UNIX-like O/S, they should use a Linux distro that implements this.
 
Why? I've never needed more than one user on Windows. I know it comes with a bunch of built-in accounts, but the only one of those I've ever used was Administrator.
Haha. Good point. Linux/FreeBSD already comes with "root". They can just use that ;)

But it is about managing expectations of those requesting a DE. They would expect to see it because "Windows and macOS has one". In other words, if FreeBSD did ever have a default DE, the masses *still* wouldn't be happy because it does not provide these expected things.

So what would happen is people would still whine that FreeBSD is "difficult" and "not user friendly" and yet we would have also messed up our nice clean base install. A lose-lose scenario.

Obviously the winning strategy is to not play the game and instead simply say "no" to a default DE in the first place.
 
Obviously the winning strategy is to not play the game and instead simply say "no" to a default DE in the first place.
Sure, we're in vigorous agreement here. I was wondering if I was missing something. Hardly ever mess with user accounts on Mac, too.
 
We have to realize that user management in Linux and FreeBSD is different. FreeBSD has master.passwd(5) which is duplicated stripped of secure data such as passwords into passwd(5).

Linux OTOH uses the SYSV approach of a shadow file. If a user manager app uses both files, passwd and shadow, it's missing /etc/shadow on FreeBSD. The exercise for people here is, find the source for such apps and modify them (with #ifdefs) to support /etc/master.passwd and propagate it to /etc/passwd, and maintain /etc/pwd.db.

Again, if you need a Linux utility that requires Linux-only interfaces the options are a) create patches to the upstream source to support *BSD, b) try to convince the FreeBSD community to abandon *BSD interfaces and replace them with Linux interfaces, or c) switch to using Linux. I don't see any other options.
 
We have to realize that user management in Linux and FreeBSD is different. FreeBSD has master.passwd(5) which is duplicated stripped of secure data such as passwords into passwd(5).
[...]
Again, if you need a Linux utility that requires Linux-only interfaces the options are a) create patches to the upstream source to support *BSD, b) try to convince the FreeBSD community to abandon *BSD interfaces and replace them with Linux interfaces, or c) switch to using Linux. I don't see any other options.
Agreed. This is why we don't have any fully complete desktop environments for FreeBSD as an observation.

In short, we are much further away from ever having a default DE for FreeBSD then I am sure many of the people asking for it realise.

So much has been stripped out of Gnome and KDE in our ports compared to upstream that perhaps we should reclassify them as window managers. Its the equivalent of i.e metacity and nautilus at this point.
 
Agreed. This is why we don't have any fully complete desktop environments for FreeBSD as an observation.

In short, we are much further away from ever having a default DE for FreeBSD then I am sure many of the people asking for it realise.

So much has been stripped out of Gnome and KDE in our ports compared to upstream that perhaps we should reclassify them as window managers. Its the equivalent of i.e metacity and nautilus at this point.
Why not install and run sysutils/desktop-installer? It will install the desktop of your choice, from a pick list.

I used this once to prep a desktop for a newb at OpenHack. I could have used this 30 years ago when I installed my first FreeBSD desktop machine. Then again, installing a Linux DE was nonexistent at the time as well. I think I used fvwm at the time.
 
I've used a bunch of different desktop environments on FreeBSD over the last ~ 30 years. They all work. Your choice of desktop depends on your preference. Just pick one and (pkg) install it. There's even a port/pkg that helps you set up a DE.

Unlike most Linux distros which come with a fixed DE, you can choose your DE. Instead of hopping from one distro to another, like a lot of people do, because the DE sucks, you can install one or more DEs on the same machine. I have a number of DEs installed so that when people ask about my FreeBSD laptop, like at OpenHack, I can log out of my current DE (CDE) and sign in using a different one.

BTW, Gnome sucks on FreeBSD just as badly as it does on any Linux distro. You have the same experience with the DE as you would on any Linux distro.
GNOME experience on Ubuntu Linux and Fedora Linux is first class imo.
 
Agreed. This is why we don't have any fully complete desktop environments for FreeBSD as an observation.

In short, we are much further away from ever having a default DE for FreeBSD then I am sure many of the people asking for it realise.

So much has been stripped out of Gnome and KDE in our ports compared to upstream that perhaps we should reclassify them as window managers. Its the equivalent of i.e metacity and nautilus at this point.
Agreed.
Maybe I'd wrote the same before, but intentionally repeat with stronger wordings.

If a DE is made defautl and any of (even a quite small and rarely used ones) the functionality the upstream documents is missing or working differently, and the user want the functionality, the user should think, "Why is this not working just as the upstream states? Ha! it's completely unusable! Why can they ship such a f*ckingly incomplete thing?".

But if it's not default, the user could (not always, though) think "Oh... This is why the DE isn't the default. I see. It's still under development. Luckily, I can use other functionalities as documented upstream. Keep trying and wait for the support."
 
Why not install and run sysutils/desktop-installer? It will install the desktop of your choice, from a pick list.
Because all of them are quite incomplete. I don't think any of these offerings are of much value compared to with what a beginner coming from Windows, macOS or even Linux would expect.

We can certainly recommend this script to new users but it should perhaps come with a disclaimer that our DE offerings are a "subset" of upstream.

GNOME experience on Ubuntu Linux and Fedora Linux is first class imo.
I do find it disappointing that the Linux community has only managed to produce a DE after decades of work that is still only well supported on a small selection of leading distros. If they took a more responsible, portable approach they might have something servicable.

But... perhaps this was the commercial vendors plan? I have my suspicions about LibreOffice/Callabora too. More confirmed when RH announced they were no longer to maintain their own RPM.
 
GNOME experience on Ubuntu Linux and Fedora Linux is first class imo.
That is if you want to search for apps in a search bar rather than pick them from a menu. I tossed Gnome when it was upgraded from 2 to 3, switching from it to KDE at the time. KDE was better but even more bloated. The Gnome UI on any platform is horrible.
 
Because all of them are quite incomplete. I don't think any of these offerings are of much value compared to with what a beginner coming from Windows, macOS or even Linux would expect.

We can certainly recommend this script to new users but it should perhaps come with a disclaimer that our DE offerings are a "subset" of upstream.


I do find it disappointing that the Linux community has only managed to produce a DE after decades of work that is still only well supported on a small selection of leading distros. If they took a more responsible, portable approach they might have something servicable.

But... perhaps this was the commercial vendors plan? I have my suspicions about LibreOffice/Callabora too. More confirmed when RH announced they were no longer to maintain their own RPM.
Hmm. LibreOffice is not Microsoft Office 365 -- I use LibreOffice on FreeBSD at home while 365 at $JOB. Additionally, none of our MUAs (mail clients) have the look and feel of Outlook, and none can work with an Exchange server.

If you're looking for a Microsoft Windows experience on a UNIX-like O/S such as any Linux distro or FreeBSD, you'll be sorely disappointed. I don't know what else to suggest except the obvious.
 
Personally though, I can't stand the Microsoft Windows interface. My main beef is the cut & paste interface. Mark, control-c, point, control-v, feel so time consuming when compared to select, move mouse and click middle button to paste. But I do get it. If a person grows up with an interface they're used to it.

For me, personally, I avoided Windows for as long as I could, opting for OS/2, switching to Linux and then FreeBSD when my career moved from IBM mainframe to UNIX. I never was accustomed to the Windows way of interfacing with the UI, and now that I have to at $JOB it annoys me.

So yeah, it really depends on what a person was "raised" on.
 
Yeah I just don't care enough to even bother setting that up.

As I have never done it I can't say for sure but I wouldn't be surprised in the slightest if some random application finds a way to do things in a non-standard manner and then suddenly the .Xresources approach doesn't work and you're again spending 12 hours trying to figure out how to work around that.
ln -s .Xresources .Xdefaults
I love this file
1728675568136.png
 
Not yet.

Though I can certainly see the commercial vendors trying to push it that way. A large focus on building it for the web via Emscripten. Some big money there with i.e Collabora Online.
It never will be 365. The UI will never be the same as 365, you'll never be able to use the same keystrokes and mouse clicks to accomplish the same tasks, i.e. the M$ Office course you took won't help you with some esoteric tasks in LibreOffice and most importantly, there's no Outlook or hook into Exchange. Yeah, it might be a facsimile but it won't be M$ Office. For users of Office the important things are UI related. I should know. My wife took an M$ 365 course. She, an average non-technical user, could never use LibreOffice and such users could never use a Linux/BSD DE. They can do the same things but the differences how to do them makes all the difference to this class of users.

It won't happen.
 
I feel like presenting a default DE enforces a way of using the OS that doesn't necessarily long-term benefit learning about the OS.

Like it's easy to drop-down Ubuntu and have a File manager that has Desktop/Music/Downloads/Pictures folders. Drag stuff into em, it works. All mainstream distros present that. Deeper, that's a XDG spec (not necessarily Linux-specific). FreeBSD having no desktop, making those folders manually "works" (and might be the only obvious solution at first), but more-proper to re-create the experience from Linux, there's xdg-user-dirs-update.

That process was abstracted by presenting it all-nice from the start, but now I have a better idea how it's done, after learning about it when it wasn't presented to me all-nice :p

Then there's Xorg vs Wayland, and stuff on Xorg like evdev vs libinput. GNOME is Wayland/libinput. Mostly everything Xorg is libinput by-default. For me, Xorg/evdev works best. Fedora is trying its hardest to make Xorg not exist as an option, and has presented GNOME/Wayland/libinput default for years, against the best user-experience (they're doing it to push Wayland development; Xorg dev is deprioritized as maintenance burden; regardless of Xorg working better, Fedora pushes Wayland to the end-user).

But the point is is that a default DE presented enforces a set of standards that may or may not be optimal for the end-user, and abstracts the process allowing a user to "fix" it or even know there's anything to fix.

I trawl through dmesg occasionally looking for something to change and can usually find something to "improve" in-general if I look at something hard-enough (like GSK_RENDERER=vulkan/ngl/gl), but I don't suspect most people do or would like doing this, especially to need to fix someone else's default presentation (I got heavy into Xorg/Wayland/libinput/evdev after literally having my basic 1000Hz mouse feel floaty with every mainstream Linux distros's GNOME/Wayland/libinput default 2016-2024 and "everyone" as end-users still claiming it superior somehow :p; I finally find this fixed/tolerable Fedora 41 Beta)

I like the DIY from the ground-up approach to needing to choose Xorg or Wayland, choosing the DE/WM you want, setting your own standards for tech (evdev vs libinput; maybe even avoiding XDG/ xdg-user-dirs-update), and the openness of experimentation (even though I like it, maybe you might find evdev sucks after trying it :p)

Particularly with FreeBSD, I like that Xorg and Wayland were laid-out as relatively un-bias options in the Handbook, and like that Xorg wasn't vilified. I like that the DE choices were laid-out, and liked being able to pick Xfce on my own. I had to battle on Linux to get an ideal set-up down (openSUSE TW Xfce has close to my ideal DE/desktop experience but I don't like oS's core), but did my ideal set-up from the ground on FreeBSD cleanly! I know what I want and didn't have to spend time undoing someone else's mess :p
 
For me, if I'm forced to choose a Linux distro to try, I'll first drop distros which has non-Mate DE as default. If none remains, drop distros which don't have packages (or something equivalent with ports) of Mate DE.

This way, having default DE would make potential users who don't love the DE to leave from it.
 
It never will be 365. The UI will never be the same as 365
No, not the trivial UI stuff. I meant that the core developers and funding is really yearning for an online "cloud" office product. I predict this is where much of the efforts will be in future from all LibreOffice's sponsors.
 
This thread is hilarious. All this talk about upstream Linux projects as if we have any control or influence on the matter. Who cares. The Linux community controls all GUI experiences used on FreeBSD.

How can we improve FreeBSD-based grassroots efforts like helloSystem? I'd wager we'd probably get further; all DE biases aside. That project has a better chance of attracting more application (that's the point of a desktop OS, right?) support than pointless bike shedding here.

Or, get a mac. And enjoy your life. :)

Btw, with that giant slush fund the Foundation has for improving driver support. I sure hope they write thorough documentation on actually writing device drivers for future hardware. People underestimate the importance of documentation; especially for driver development.
 
This thread is hilarious. All this talk about upstream Linux projects as if we have any control or influence on the matter. Who cares. The Linux community controls all GUI experiences used on FreeBSD.

How can we improve FreeBSD-based grassroots efforts like helloSystem? I'd wager we'd probably get further; all DE biases aside. That project has a better chance of attracting more application (that's the point of a desktop OS, right?) support than pointless bike shedding here.

Or, get a mac. And enjoy your life. :)

Btw, with that giant slush fund the Foundation has for improving driver support. I sure hope they write thorough documentation on actually writing device drivers for future hardware. People underestimate the importance of documentation; especially for driver development.
I agree with the idea that a lot of forum users may be happier using another system rather than FreeBSD. Namely those who want the system to be more like a different system.
 
Back
Top