Should graphics be a topic of rigor in FreeBSD?

Should graphics be a topic of rigor in FreeBSD?

  • Sure!

    Votes: 24 68.6%
  • No..

    Votes: 7 20.0%
  • Meh, I don't care.

    Votes: 4 11.4%

  • Total voters
    35
Just a questionnaire to gauge the audience here. Much like networking (TCP/IP) and storage (UFS|GEOM/ZFS) have been highly underlined historically. Some examples I have in mind would be display stuff (touch or kb/m), GPU compute, graphical applications (ie gaming), native BSD drivers (akin to iflib), etc. I think this would be a great opportunity for FreeBSD to be a more usable system. What do you think? thoughts?
 
My opinion is you can't make open source developers do anything. They have to want it.
I am quite happy having a FreeBSD desktop. That's as far as my concern goes.
GPU Compute and Gaming don't interest me.
I do have a stack of tablets with no evdev working. So that might qualify.
Truth is if I cared I would be running Linux on them. I have no real need for a tablet or touch.
I grew up on digitizer tablet and pucks.

What worries me the most is having working browsers on FreeBSD. They are dropping like flies.
 
Yeah, I think that graphics stack on FreeBSD is rather suffering from lack of attention. I am interested in getting GPU computing going under FreeBSD. Graphics acceleration for video players like VLC is one thing, and Blender renders would really benefit, too. Marazmista's radeon-profile and GPU benchmarking/overclocking are of interest to me.
 
I'm happy enough with what we have already. I'm mainly here for the server capabilities of FreeBSD. I like the flexibility of being able to install a compact efficient OS in less than 2 GB of disk space, then configuring it in whatever way suits my needs.

The MATE desktop is more than enough for my purposes. That said, I've always enjoyed playing around with KDE, which has been an available option ever since I first installed FreeBSD from the 4 Walnut Creek CDs which came with the Complete FreeBSD book by Greg Lehey, which I bought at Staple's office supply store way back in the '90s. I think FreeBSD is doing a great job of being what, it seems, it was always intended to be.
 
As for graphics, it's good enough. FreeBSD made great strides. It will be a while before Wayland parts replace parts of Xorg, then hopefully it gets better and less complicated. Cleaning up Gtk and Gnome related bloat would go a long way, but there's stubbornness around that.

It used to be, oh, harddisk space is cheap. But compiling time was still cut down by hours, just by removing bloat, and the programs ran faster, and be more reproducible. So, if harddisk space is cheap, I want that for storage, maybe sitting around, redundancy and for what's needed, not bloat, increased compile times, unnecessary install conflicts, inconsistent updates and crashing programs.

Graphics isn't a problem, unless you're playing the latest action PC game.
What worries me the most is having working browsers on FreeBSD. They are dropping like flies.
Palemoon and the associated email client without Gecko. Firefox and Thunderbird need only gtk3, but they rely on gtk2 because of some mixed up code. For most, including I, gtk2 and gtk3 are needed anyway, but it shows that the code is a mess, when it pulls in a dependency in order to function that it doesn't need.
 
I think you lost them to python2 build. But ok firefox will do. It has no python2 dependencies or qt5-webengine.
Note : even Chromium has some weird dependencies ...
 
I'm happy enough with what we have already. I'm mainly here for the server capabilities of FreeBSD. I like the flexibility of being able to install a compact efficient OS in less than 2 GB of disk space, then configuring it in whatever way suits my needs.
+1 here. But yes, concerned about the narrowing choices generally in technology with hardware, browsers, walled gardens/DRM etc. But at the same time guilty of lazy pleasures like iDevices and Netflix and Amazon etc. :sssh:
 
My newest machines are 2008 Thinkpad W520's that use Optimus and that's been working for me without any tinkering since 12.0 IIRC.

I voted yes because it should offer support for the graphics it's able to, all things considered.

I just converted this one to FreeBSD from Win10Pro because it was my gaming box and stayed offline. I haven't played computer games in two years or have any interest in it anymore.

I haven't even updated to the newest nvidia drivers, since everything is running like an atomic clock.
 
In my opinion, the complexity of graphics, software/gaming dependencies make it a tradeoff. The amount of time and code dependencies needed to support complex graphics ( and the software that requires complex graphics ) detracts from effort spent making a solid system.

I would prefer a secure and bulletproof system/networking stack and minimal graphics, rather than a mediocre/inconsistent system stack and high performance graphics.

I mainly am interested in FreeBSD as a server and hypervisor. So, as long as I can display a nice resolution XFCE or KDE desktop, I am pretty happy.
 
Graphics should not be mandatory, but definitely should be fully functional.

Now Freebsd has reached a stage where graphics is easily usable, and with a few quirks, as good as any other system.

I am glad I can use Freebsd as either or both;
  • purely as server where graphics is redundant or frivolous,
  • as a functional desktop and everyday environment.
Why should there be a question?
 
Intel, amd or nvidia ?

Any. This could include ARM based GPUs, or any emerging GPUs. I envision a framework where one can just re-use code to create new drivers and omit features you don't need dynamically. See iflib and I/Okit.

Why should there be a question?

Because such functionality relies on chasing upstream that FreeBSD has no control over. DRM itself is bad design IMO too.

Case in point from this article. Imagine all the duplicate feature gunk from each vendor because of it.

Even some Linux users understand it's brokenness.

Screen Shot 2021-07-03 at 9.06.49 PM.png


What worries me the most is having working browsers on FreeBSD. They are dropping like flies.

Agreed. It's incredibly frustrating how rigid software is nowadays. Using Redhat a controlled toolkit to render it's UI is just a dumb decision.
 
In my opinion, the complexity of graphics, software/gaming dependencies make it a tradeoff. The amount of time and code dependencies needed to support complex graphics ( and the software that requires complex graphics ) detracts from effort spent making a solid system.

I would prefer a secure and bulletproof system/networking stack and minimal graphics, rather than a mediocre/inconsistent system stack and high performance graphics.

I mainly am interested in FreeBSD as a server and hypervisor. So, as long as I can display a nice resolution XFCE or KDE desktop, I am pretty happy.
Fair enough. I think the same reasoning can be applied to file systems too. GEOM was created to solve a specific problem. With a native graphics ecosystem we can take our path on things. Maybe I’m thinking too intrinsically about it, but I like the idea of not to having to depend on Linux for our basic needs. Just research and build.
 
It goes through the LinuxKPI, another support layer, like the Linuxulator. It's actually intelligent stuff, but DRM itself needs to die in a giant raging ball of fire.
 
Yes as long as the web says FreeBSD is an operating system used to power modern servers, desktops, and embedded platforms. I think drm-kmod came to fix that.

Off-topic: There is a lack of a community in the main languages, which may be a symptom of their moderate popularity.
 
Would it be nice if there was a comprehensive native graphics stack for the BSDs? Yes.

But, writing device drivers is easiest with documentation. Yes, you can reverse engineer, but it's not easy.
Documentation for current hardware is difficult to obtain legally. Most companies want you to sign NDAs with onerous penalties and will cost you big bucks.

The manufacturer could write drivers, some do as sample, some are actually trying to be nicer to OpenSource (Linux, BSDs, etc), but I'd bet a coffee that the latest Linux driver for Intel/Nvidia/AMD lags behind the latest Windows/Mac, simply because of money.

Then there is the issue of License. Will Nvidia provide documentation so someone could write a full featured up to date driver under BSD-3Clause? I'd say "not likely".
The Linux drivers that make up drm-kmod and the linuxkpi are about the best documentation we can get but they will still lag.
 
I don't mind loading a blob for my video or audio card.
It is bizar companies like Nvidia are difficult to convince of an open standard/library.
 
I said "meh I don't care" because yes, on one hand that would be awesome. FreeBSD has a bit limited use for me in its present form. On the other hand it could become a bloated behemoth.

I need a good support of my GPU for my graphical desktop use case:
- When I connect a flash disk, I expect it to automount
- I expect a 3D game to work
- When I connect even a rare piece of hardware (e.g. musical hardware) I expect it to work out-of-the-box
- I expect my bluetooth headphones to work out of the box with some easy graphical application to connect to it
- When I connect old-style wired headphones to a front panel of my PC, I expect the sound output to be automagically switched there
etc.

This, unfortunately, drags a lot of bloat with it. FreeBSD would became what Ubuntu or Windows/MacOS are nowadays. Hm, not sure I want that.

I praise FreeBSD to be a very coherent and minimalistic system which works for me as a server and as a workstation for programming / data analysis. For the latter use case I really need just Openbox, Firefox, Python, C, Lisp ... and a good IDE (vscode does the job, PyCharm worked for me in the past on FreeBSD, others are probably available too ... and there is Emacs if there are not).

Great graphical support is not a priority for FreeBSD for me.
 
Back
Top