I genuinely think freebsd would be more popular with a purely bsd eye-candy desktop environment

I think that sometime some automation would be fine.

For instance the nvidia-driver in order to work requires a couple of lines in the RC.CONF...

You must read the handbook and do it by yourself, or the installer might be proactive asking if you want allow it to make the change and pleasing your laziness.

I do respect this choice making my laziness sad... 😇
 
i915 mode setting also requires editing kld_list in /etc/rc.conf.
That is what the setup stuff in GhostBSD and other typically do.
Figure out what hardware is installed, determine the best driver, setup the init files correctly.
Modern X works very well for a large percentage of cases, most folks don't need any special configuration files.
For the cases that need tweaking, it's often relatively easy once you know what you need.
Again, this is where a smart script could help.
 
Windows detects and downloads the correct driver at least to get a framebuffer working, if not modest acceleration. I don't know what linux does.
If you can see the installer after booting the install stick, you know that the framebuffer works. Heck, if you can see the BIOS, you'd know the framebuffer works. Linux does the same thing, BTW.
--
If Linux is Spam that you can eat straight out of the can, then FreeBSD is a nice cut of raw Kobe beef that still needs some cooking before it can be served. :p
 
The biggest "user issue" seems to be "what driver do I install". Do I use modeset, xorg driver, nvidia, which version, blah blah blah.
Ultimately these are all Nvidia issues (you don't have any issues like this with open-source drivers provided by FreeBSD). I get round this by not choosing Nvidia. Or perhaps a bug request could be submitted on Nvidia's forums? They could probably merge the modesetting and classic driver into one module if they deem your request worthy of their time.

Intel was a bit messy a while back (mainly because we started using Linux kind of driver infrastructure. This has cleaned up considerably since).
 
These aren't just Nvidia issues...

Gnome, KDE, XFCE and also LibreOffice requires to add a couple of configurations in a couple of configuration files.

It wouldn't bad if the package manager would be so kind to ask to you if you prefer that it makes these changes for you.
I think the meaning behind this is to preserve an educational aspect and I do not disagree with this purpose.
 
Arch Linux has been pretty successful in directing people to RTFM for any sort of desktop configuration, whilst gaining tons of popularity. In fact, their philosophy is grealy inspired by the BSDs. So what makes one think it's ok to have a desktop ship in base? Let alone the behemoth that is X.org. I highly doubt the committers would bother with that sort of burden. Seriously, someone should patch bsdinstall if they really care that much.
 
… GhostBSD …

There's so much to like/love about GhostBSD.

… what makes one think it's ok to have a desktop ship in base? …

At least one thing inspires thoughts of OK-ness:

GhostBSD!

– I'm teasing :) but only half-joking.​

As an example of a distribution that's user-friendly from the outset, GhostBSD excels. IMHO it's the gold standard.

FreeBSD base

I like the current separation. Kernel + base without a desktop environment.

For now, continue to make FreeBSD a better basis for desktop/notebook/laptop use cases.

… For instance the nvidia-driver in order to work requires a couple of lines in …

… Gnome, KDE, XFCE and also LibreOffice requires to add …

For fun, a list of ports where the package message includes the word edit: <https://www.freshports.org/search.p...own=asc&search=Search&format=html&branch=head>

… I think the meaning behind this is to preserve an educational aspect and I do not disagree with this purpose. …

👍
 
The target audience for FreeBSD is not pop culture, kids and their games. The target audience is those who need and want a solid, serious foundation for their work. FreeBSD does not create desktop software.
I understand your viewpoint. But I think that there is no neutral position - one OS can go forward or backward in popularity. To be vital FreeBSD needs as much users as possible, even non-specialists, kids, or random users who think that FreeBSD is a Linux distro. In short: More users => more attention from hardware manufacturers => more drivers/support for any hardware => more software written for FreeBSD, including desktop, business, etc. => more users and "marketshare". And this needs desktop support even if the target group are serious users looking for solid server OS without GUI. I am sure that nearly all users from this forum will be happy to see 200 million FreeBSD users instead of 200 thousand.
 
Just to clarify to those who are pro-"X11 in base". I do certainly see some benefits. OpenBSD does this and it really works well. In many ways X11 is quite bound to the OS if you plan to do anything graphical and effectively just another set of drivers. If anything, "out of tree" drivers like Nvidia are a bit of a mess so it seems counter-intuitive to have "all" of the video drivers partly out of the tree (really, only the proprietary blobs should be excluded).

This is where the Intel driver can be added into kld_list as part of the installer perhaps (though since we don't do this for i.e acpi_ibm on thinkpads, it would feel a little inconsistent)

If X11 was ever brought into base (long ago, it used to be a base "set"), it would need to be seriously cleaned up. No duplicate versions of LLVM, no dbus, no python, no wayland things as dependencies. It might even be easier to throw it away (or leave it in ports) and bring in Xenocara just because it is monolithic and easier to maintain for FreeBSD. This would still be a big undertaking and in this time the developers wouldn't even be thinking about how to get Xfce (the current winner?) in base, which is an even bigger mess (all DEs are).

But it also does run the risk of opening us up to people pining for X11 based-GUI installers and not even knowing why and just starting to make things heavy and messy.
 
  • Like
Reactions: mer
In short: More users => more attention from hardware manufacturers => more drivers/support for any hardware => more software written for FreeBSD
It's precisely the other way around.

I am sure that nearly all users from this forum will be happy to see 200 million FreeBSD users instead of 200 thousand.
I'm impressed you think FreeBSD approaches 200 000 desktop users.
 
That's all of them at the moment.
Isn't it only the Linux KPI stuff and the Xorg "hooks". The core parts of many video drivers are still in the kernel (to allow for i.e a virtual framebuffer). Though reading back, I had subsequently expanded that bit.
 
I don't think it's fair to consider unaccelerated video a contender in the desktop case at all.
Yep, I pretty much agree. This is my argument for why Xorg being part of FreeBSD base isn't the worst thing in the world. Currently we have this slightly strange situation of video drivers partly being included in base, partly being in ports (and some of them are even tied quite closely to the kernel versions).

Plus, if we can get Xorg in base before Wayland compositors might make any ground, it might make people think twice before jumping to the broken junk ;).
 
In a perfect world, X11 shouldn't include any "drivers". Historically, accessing some graphics card wasn't much more than accessing some memory-mapped framebuffer memory, and X could do this running as root. But that's a weird design. Any hardware drivers should be independent from some windowing system, the latter should just use abstracted interfaces (and this is what's happening today, e.g. with the "modesetting driver").

X11 shouldn't be in base either, for several reasons, for example these:
  • It's not the only option, at least in theory. It's a mess carrying a lot of deprecated stuff. Something else will replace it eventually, at least that's my bet
  • Base should be small. It should include anything to make the system functional enough for an admin to carry out important tasks. Nothing more.
A general-purpose OS must be suitable for a desktop/workstation. This doesn't mean all software necessary for that must be part of it, it just means software necessary for it must work.

Isn't it only the Linux KPI stuff and the Xorg "hooks". The core parts of video drivers are still in the kernel (to allow for i.e a framebuffer).
Not really. Of course, drivers for EFI, VESA framebuffers are present, but nothing else. Drivers based on Linux DRM/DRI/KMS (for Radeon, AMD, intel, ...) were completely removed and are developed as a separate project, available as a port.
 
X11 shouldn't be in base either, for several reasons, for example these:
Yes, you're not wrong. I think this is why it was removed in the first place. I think the cons of Xorg in base, do outweigh the pros ultimately.
  • It's not the only option, at least in theory. It's a mess carrying a lot of deprecated stuff. Something else will replace it eventually, at least that's my bet
I suppose the same could be said for most things in base. For example GCC -> Clang. Granted Xorg is a lot more messy.
 
I personally don't want X in base or any kind of GUI more complicated than bsdinstall. I just think a script to do boring, automate-able, and "whoops I made I typo" stuff when installing X would be a benefit to FreeBSD.
 
But some will say that's too ugly and it needs to be replaced with a lovely GUI installer. That works out the box on any graphics and has Wifi support built-in. Etc.
Work on a different installer started in April of last year, possibly earlier:

Also, FreeBSD is actively supporting a transition to Wayland. I posted the link on my profile back in December:

A bit too late to start making noise about 'Xorg in base', I'd think.
 
There's so much to like/love about GhostBSD.



At least one thing inspires thoughts of OK-ness:

GhostBSD!

– I'm teasing :) but only half-joking.​

Only the current latest version is 100% based on FreeBSD, one of things that made me prefer FreeBSD over GhostBSD was OpenRC as default init. That made impossible taking advantage of FreeBSD forum to troubleshoot GhostBSD issues.
 
Gnome, KDE, XFCE and also LibreOffice requires to add a couple of configurations in a couple of configuration files.
Checked. First two with a blank, brandnew, basic, default & clean 13.0/amd64 installation:

Gnome: pkg install xorg gnome3 - you've just got to fire up DBus first (f.e. with service dbus onestart), and the desktop can be used. I've been pointed to it by just reading the error message when trying without DBus (which I found already being installed as dependency).

Xfce: pkg install xorg xfce4 - runs out of the box.

LibreOffice: Never had to configure anything for it, works out of the box.

Haven't checked KDE (doesn't like my VNC), but: As far I can see it's not that bad on FreeBSD.
 
  • Like
Reactions: mer
I am sure that nearly all users from this forum will be happy to see 200 million FreeBSD users instead of 200 thousand.
It's hard to measure how many people use a desktop operating system. The best way I know of is to look at web browser statistics. To begin with, about 2/3 or 3/4 of all worldwide web traffic originates on mobile (meaning Android or iOS), so desktop use is already a small fraction. Further breaking desktop use down, I found the following statistics: About 76% is Windows, 16% is Mac, 4% is unknown, 2.5% is ChromeOS, 2% is Linux, 0.01% is FreeBSD, all other OSes are smaller. That's not a joke: FreeBSD is 200 times smaller than Linux, which is already only a few percent of total. The total number of clients using web services is several billion (a large fraction of all humans on the planet are still not connected to the web, amazingly enough), so multiplying this out puts the number of FreeBSD desktop users at about 50,000, give or take a factor of 2 or 3.

And personally, having more desktop users on FreeBSD will not make me happy. At best, it is irrelevant to me, as I chose to not use FreeBSD on the desktop, instead I use it for servers. Actually, I think a lot of desktop users is a net negative, since in many cases desktop users tend to be less experienced and less thorough. A lot of traffic I see on this forum or on Reddit about FreeBSD desktop users is just "distro hopping": users who think that FreeBSD is sort of another Linux distro, they want to try it once, it doesn't work the way they expect, the get frustrated or angry, post a few questions, and then give up.
 
Back
Top