|
Will BSD keep X11 in the future?
Hello,
X11 is to me really important. I like to have a system that allows to modify and to custom to our needs. However, if developers change all the time their libraries, it starts to be seriously annoying to work and to use Linux. Is BSD cleaner and the ultimate solution to run X11/Nix*/Unix? |
There is no single "BSD" development project. However many people conflate FreeBSD with "BSD" merely because it is the largest of the currently active operating system projects. You might be one of those who assume FreeBSD = BSD.
The four main active development projects -- DragonflyBSD, FreeBSD, NetBSD, and OpenBSD -- are all separate projects with separate staff, seperate goals, separate strategies, and separate direction. These systems have diverged over many decades, though all are based upon one of the releases of original Berkeley Software Distribution, or are forks of each other. https://en.wikipedia.org/wiki/Berkel...e_Distribution |
Quote:
BSD meant of course openbsd, freebsd,... |
Disclaimer: I gathered this information via the Internet, via Google, in about 10 minutes. An experimental Wayland or Weston development noted or available for testing does not mean that the individual project has stated any direction to remove or supplant their X11 software. For more information on strategy or direction, please refer to to the individual projects.
DragonflyBSD
|
If it doesn't include it on the basic install disc, you can always add it in a package.
In fact, the main Solaris installer disc doesn't come with X11/Gnome. You have to type Code:
pkg install solaris-desktop |
Quote:
|
I don't see X11 disappearing any time soon. Wayland is something designed by X11 developers, but is ultimately freedesktop.org stuff and thus "Linux driven" and dependent on Linux DRM/KMS drivers. It's in early development and while the Red Hats of this world may jump onboard sooner than most, it's not yet a given that it will eventually completely replace X11.
|
Quote:
It would be terrible to start about 20-30 years again development for make again WMs. All X11 applications would be completely waste... 20-30 years of waste of time. |
Well like it or not, simpler compositors seem to be lined up to replace the whole X11 kitchen sink and window managers. This is the kind of thing which other platforms had since the 1990s. I don't exactly look forward to it nor dread it, I find myself indifferent. If we get more responsive desktops, cleaner code and better security, it has to be better right?
|
Quote:
Soon or later, you will have to move to BSD... you won't have any choice. This is happening already with SYSTEMD. |
No, on the contrary, it doesn't come much heavier or complex than X11.
I would say that there's always a choice. Slackware Linux does not use systemd as is the case with some others and Gentoo Linux and Linux From Scratch gives the option to build without it. Red Hat and Debian would seem to be the wrong way to go if you want to avoid systemd. Linux went maintstream and some people don't like the result. gnome, KDE, et al and all the underlying automagic, messagebus and *kit stuff has existed for years, but there were few complaints or references to "UNIX philosophy" from end users. It's all inevitable really while you have much of these projects being corporate funded/controlled, rather than in the hands of hobbyists/academics. |
Quote:
People who come from MS Windows, willing for change, will install KDE or GNOME. People who do some research will install TWM, CTWM, GNUPLOT, and LATEX/XETEX/ whatever TEX and their bibtex, for serious stuffs. People, who are learning programming will install Java, Python, Mono,... believing that C/C++ is a loss of time. and people, who are responsible of famous Linux distributions, will bring Systemd, Pulseaudio, heavy libraries, and remove for fun, all the working part of Linux for modern unstable, most unwanted, programmes. Life of Linux, Everything's possible ;) :) |
Quote:
Quote:
Quote:
The language you learn and use depends on what you want to do. e.g. if you want to hack device drivers, then you need to learn C. But if you're writing some other application, then other languages and toolkits are what you want. Ideally you want to be linking to exiting code, rather than reinventing the wheel. This is how modular systems are put together and how the free software eco system "works". Quote:
|
Quote:
|
I think all projects welcome users and bug reports - that's a given. But software development projects are generally not democracies (or anarchy) where anyone can suggest anything (and then let someone else do all the work implementing that).
Your Linux distribution, Slackware, for example is a benevolent dictatorship. You get to use Patrick Volkerding's personal Linux distro, designed according to PV's personal tastes - for free. Slackware is entirely financed by sales, via the Slackware store or any donations. You can buy merchandise or not and you can use it freely. What you don't really have is a say. You can suggest something, it might be considered, but when all is said and done - the man decides. |
There are many projects that like the 'unix way', & there are companies who have vested interests in certain distros.
These companies make money by offering technical services, & that is where systemd, (& to a lesser extent pulseaudio), come into the picture. They want to see monolithic software projects that are complex to administer for businesses, so that they can sell them their services. Linux, has been seen to be trying to get a share of the desktop market, & these companies would like that, but the average user just wants something that works as a (free) replacement for MS Windows. Linux may eventually go two different ways, but as long as the kernel developers prevent any contamination at that level, we should be able to have both. The BSDs have always been more unix like than Linux, & I don't see them changing much. |
X11 is the engineering equivalent of an 80 year old rickety bridge that's been shored up half as many times as it is years old. After they get everything straight with XWayland, you'll be able to run X11 apps there. Eventually, I expect the developers will stop updating XServer all together, except for XWayland.
Last I looked, the BSD guys were working on the Wayland port, probably modifying it to better work with their kernel and/or remove any GNU Tools dependencies. I don't know for sure but, if BSD is not planning to move to Wayland, then I have to wonder why they put all that KMS stuff in the FreeBSD kernel. I'm all for keeping things that work, except those that have become more trouble than they are worth to maintain. I'm sure CBM-DOS still works fine too, but there's little interest in continuing its development for good reason: it's ill suited for modern hardware. Just like X11 is ill suited for modern interface design. I'm sure there are people working on "lightweight" Wayland wm-compositors and not just KDE/Gnome sized projects. They'll be needed for IoT and embedded after all... |
Quote:
Quote:
|
Code:
X11 is the engineering equivalent of an 80 year old rickety bridge that's been shored up half as many times as it is years old. After they get everything straight with XWayland, you'll be able to run X11 apps there. Eventually, I expect the developers will stop updating XServer all together, except for XWayland.x11 will be supported for 1-2 years, and all BSD and LINUX world will end with its graphics. Borland graphics library was the same for long time ago. Graphical applications are the most unreliable things to make and to maintain. Why to programme if developer have fun changing the libraries all the 10-20 years? |
Quote:
Apparently, the developers whom are working on it think so. You know... sometimes they actually demolish whole buildings before building a new one too. |
Quote:
The example it no so much comparable. Building are livings which are touching only a localized region of population. "Informatic has become key of our living." If you would say, we replace current pipes (for water, under the ground) by another metallic/plastic pipes. But, this would happen readily tomorrow, everywhere on all over the planet. No one on this planet would have water. They would be construction jobs everywhere. Everyone would have to install again. Like in informatics, you would do it every 2-5 years regularly. |
Well you seem to be interested in minimal memory footprint and less bloat in your other threads, yet here you seem to cling on to code from the 80s, much of which is bloat, cruft, you name it... and isn't used on today's, mainly x86 based, workstations.
|
Quote:
|
Quote:
Only the applications and libraries that make X11 calls directly would need to be updated. That's "updated". Not "waste". Most of them interact with X via middleware such as, oh, GTK. For those, you'd update GTK so that it works with Wayland, and then leave the application as-is. |
Quote:
Developers might say, this is just a new library. Example. Let's take a cool example: Many companies run virtualbox/vmware to run old windows softwares for using very expensive equipment. Not 50-400$, I mean. Rather some millions ;) |
Change for the sake of change is pointless and, as you call it, a waste. However, not all change is pointless. Sometimes change is an improvement. I also do not considser updates to be changes. X11 has evolved with many additions and updates. I am not stating an opinion, but merely posing the questions: Have all the those additions made X11 combersome? If so, is it time to replace it with something new and simpler?
|
Quote:
Change must be made if there is a strong and good reason. The freedom of change is for everyone different. Since the distros will make the decision for you, the individuals have no choice. A group of persons can compile (developers) for individuals, if they have some time and energy. But this is not an obligation. There is likely no choice either. |
It seems to me that FreeBSD and other will keep X11 for some time. But will it be possible that we go through the new Y?
Will x11, such as tinywm source code be still compatible? DragonflyBSD Latest release: 4.8. X11: third party (dport or package) X.Org 7.7 Other direction:Wayland, experimental FreeBSD Latest release: 11.0. X11: third party (port or package) X.Org 7.7 Other direction:Wayland / Weston, experimental NetBSD Latest release: 7.1. X11: integrated X.Org or XFree86, or third party (pkgsrc) modular X.Org Other direction:Wayland, experimental OpenBSD Latest release: 6.1. X11: integrated X.Org 7.7 Other direction:Wayland, experimental |
Who can say...?
If you're committed to a switch to FreeBSD, then it's best to just use that and not bother fretting about what will or will not be supported in the future. In OpenBSD for example, I can't see the integrated X.org (xenocara) going anywhere any time soon. But if it does, it will be years after the Linux people abandon it and move to whatever 'replaces' it. Whatever "it" is, it will only replace X.org in the OpenBSD base system if it fits the necessary criteria/requirements. Even then X.org could still be available or would move to ports. In FreeBSD, where X.org is part of the ports tree, X.org will stay around as long as someone wants to maintain the port. With the *BSDs in general, it's more about technical exchanges and pragmatism. Fanclub type behaviour and ideological chit chat doesn't gain much traction. |
In the FreeBSD ports tree there is graphics/wayland, graphics/wayland-protocols, x11-servers/wayland, x11-toolkits/wlc (Wayland compositor library) and x11/kf5-kwayland (KF5 Client and Server library wrapper for the Wayland libraries).
https://www.freebsd.org/cgi/ports.cg...land&stype=all If Xorg is going away any time soon I haven't heard about it, which means next to nothing. It works for me so I don't really have a bias against it. I'll go with whatever they decide and learn to love it. I just rebuilt all 3 of my FreeBSD boxen tonight and timed how long it took each to compile Xorg from ports: AMD Phenom II x 3 @ 2.1GHz and 4GB RAM 3hrs 20 mins Intel Core 2 Duo @ 2,0GHz with 4GB RAM - 4hrs 30 mins Intel Duo Core @ 1.60GHz and 2GB RAM - 7hrs 30 mins Time is one thing I do have ATM so it's not really an issue for me. I was just curious and wanted to document it. I have an OpenBSD box with xenocara and can't tell the difference in the end product from Xorg. Only that I could boot directly into a desktop without having to built it separately. |
Quote:
|
Quote:
X.org Xenocara it is all the same based on MIT graphical processing. Actually "Y" might be similar? |
Quote:
You know how you can build a panel under org and get it on the screen without a running window manager? Can't happen on Wayland. LXQt is a good example, they're using kwindowsystem & kwayland libs from plasma to achieve that bit I think. Anyway, that's my understanding of the problem. I could be wrong, but then a lot of what I've read would have to be wrong, or I've completely misunderstood it. |
Quote:
|
Quote:
|
Okay, that first video was by a former X11 and current (as of it's publishing), Wayland developer. He explained that X isn't network transparent, contrary to popular belief. That it doesn't adhere to the Unix philosophy, it does one thing, but it does it badly (IPC). And finally, how fixing all that crap would break the X specification so badly, that you'd have to rewrite it anyway.
This next video is by a Tzien & Enlightenment developer. He explains many more of the problems with X. (It's cooperative, it's insecure and there's no way for the system to tell or act on an application that's behaving badly.) https://youtu.be/CeAnSSyv20Q |
Quote:
Looks to me another SystemD story for all distributions. |
Wayland has nothing to do with Systemd. Weston does, but it's a reference tool for testing the wayland stack. When someone makes a Wayland compositor they can use Systemd, elogind, consolekit2 or whatever they like.
I should point out that what the two big toolkits (GTK & Qt) are currently doing is bad. They're making hybridized compositors that will run both Wayland and X11. It's like they haven't even looked at XWayland, which already did all that work for them and in a sane way. The way they are going about it it will be a repeat of the way we still need both classic X and XCB, when XCB was intended to replace alot of legacy X components. We'll never get rid of X11 with the setup they're developing. |
Quote:
Linux should keep SysV, X11, Motif, ... libx11-dev, intrinsic, ... old good things, just because they are good and not heavy coded. GTK, QT, KDE,... can do whatever they want, wayland too, but distributions shall offer a possibility to have old efficient softwares but also shining softwares. Freedom of choice is Freedom. |
So, just like the GTK/Qt guys. You didn't bother to even look at XWayland. X is not good, it's terrible, that's the whole f'n point. People assume it's good because it works. Opening your front door by taking off the hinges with a screwdriver works too, but its far from ideal. Crank starting your damn car would work too, but I doubt most of us would want to bother with it. All that's without even mentioning that it was designed around the idea that software could be trusted, it cannot be, or maybe you live in a different world than the rest of us.
|
My problem is that I imagine things like this...
A native Wayland compositor like Openbox, only it includes a customizable panel. No need of a full widget kit, it already contains everything needed for standard desktop duty. There's no need for compton, or compiz or some other hack, because it's already a compositor. X11 is there, but running as the XWayland client, so you never have to deal with it's video driver or input B.S. again. It's small, it's clean, it's written in C and has no need of introspection BS. It's memory footprint is smaller than xfce, unless you're running some old X application. It has python & golang FFI's for the panel, so putting together an applet is simple and most of support structure lies outside of the gui. The only thing you must write for a applet is the code for what's displayed in the window. There's nothing built in... no volume control, no cpu monitor, no weatherbug, no battery monitor... that that stuff comes as little applet front-ends to other external programs that already do all that crap anyway. Finally, there's a small & simple xdg based theme changer. It doesn't even come with a terminal! Guess I could call it the DIY Wayland Desktop. :P |
Quote:
|
Quote:
|
Quote:
will-bsd-keep-x11-in-the-future - Probably, but actually, maybe not. Nobody really knows. Thread solved. |
Quote:
Now, what I find funny is that anyone who would threaten to fork and keep it alive is doomed to fail. Why? Because it's difficult and few people understand X's inner workings as-is. What is funny is that the minute they begin to understand how to maintain it, they'll see exactly why it needs to be replaced. As for the BSD developers... They're better than most and I suspect it wouldn't take them long to estimate that it is better to leave sleeping dogs lie in the case of X11. Certainly, none of their business clients are interested in at as the Tzien developer so eloquently pointed out in the video I linked. |
Quote:
|
Depends on how long it takes for the Wayland ecosystem to develop. It's slow because things work differently. Qt & Gtk are well underway to tooling support for it. But, there is another angle to this, the back-end is changing too, and also for good reasons.
Code:
Current: Kernel Interface -> OpenGL -> Mesa -> X11 -> Widgets -> DesktopIf we don't get all this done within a decade we'll find ourselves in a quagmire of jury-rigged APIs. |
Here Xeratul, I dug this up just for you. Proof that Motif apps can still run in a Wayland world without running X11 as the primary display server.
http://www.h-online.com/open/news/it...d-1619051.html (Note: I've done this myself under Weston. X apps run fine, better if you count the fact that they no longer tear and accordion.) |
Quote:
Wayland says that they wont support X11 long, just a early beginning. And then you have a serious issue houston. |
Quote:
|
| All times are GMT -5. The time now is 06:43 AM. |