SoylentNews is people
SoylentNews
Chris Siebenmann, a UNIX herder at the University of Toronto CS Lab, asserts that the death watch for the X Window System (aka X11) has probably started:
I was recently reading Christian F.K. Schaller's On the Road to Fedora Workstation 31 (via both Fedora Planet and Planet Gnome). In it, Schaller says in one section (about Gnome and their move to fully work on Wayland):
Once we are done with this we expect X.org to go into hard maintenance mode fairly quickly. The reality is that X.org is basically maintained by us and thus once we stop paying attention to it there is unlikely to be any major new releases coming out and there might even be some bitrot setting in over time. We will keep an eye on it as we will want to ensure X.org stays supportable until the end of the RHEL8 lifecycle at a minimum, but let this be a friendly notice for everyone who rely the work we do maintaining the Linux graphics stack, get onto Wayland, that is where the future is.
X11, for all its advantages, also has several incurable design flaws relating to security. However, the major distros have not yet been in any hurry to replace it. Wayland is touted as the next step in graphical interfaces. What are Soylentils thoughts on Wayland or the demise of X11?
(Score: 2, Insightful) by Anonymous Coward on Thursday July 04 2019, @03:55PM (17 children)
When people were dissatisfies with how XFree86 handled it, X.org took over. If X.org no longer properly maintains it, someone else might take over.
Of course it might also be that someone creates a new protocol that has all the advantages of X, but none of its disadvantages. But that's much less likely, to put it mildly.
Anyway, if Wayland is the future, I don't like the future.
(Score: 4, Interesting) by canopic jug on Thursday July 04 2019, @04:20PM (14 children)
The X11 code is rather special. It requires in-depth knowledge and then a lot of work to make any kind of changes. Again, some of the flaws are with the design and those won't be fixed by patching even if someone skilled and motivated were to adopt part of the code base. What is needed is modifying or replacing the specification, and then new code. However, I agree with the other comments about the lack of remote access in Wayland being a deal breaker. VNC and the others just won't cut it for most use cases.
Another bit of special code is related to NTP. There are relatively few people capable of doing a reimplementation. Poul-Henning Kamp was one of the right people for the job and was looking at writing a replacement, after realizing that the legacy code was an intractible mess. However, he seems to have stopped blogging about it [freebsd.dk] a few years ago. A casual browse around the web doesn't turn up any recent code either and the ntimed project page has no links to and recent code [nwtime.org] and only mentions something about 2017, year which has already passed.
Money is not free speech. Elections should not be auctions.
(Score: 3, Informative) by Snotnose on Thursday July 04 2019, @04:45PM (4 children)
Eric Raymond [ibiblio.org] has been working on NTP.
When the dust settled America realized it was saved by a porn star.
(Score: 3, Informative) by canopic jug on Thursday July 04 2019, @05:07PM (3 children)
I'm not finding ESR's name on the list of NTP contributors [ntp.org]. However, PHK is on their active list.
Money is not free speech. Elections should not be auctions.
(Score: 1, Informative) by Anonymous Coward on Thursday July 04 2019, @07:05PM
Raymond is working on NTPsec, a partial rewrite/refactoring focused on hardening and limited to non-legacy systems.
(Score: 2) by captain normal on Thursday July 04 2019, @07:47PM
Did you read Snot's link?
Everyone is entitled to his own opinion, but not to his own facts"- --Daniel Patrick Moynihan--
(Score: 2, Informative) by ncc74656 on Thursday July 04 2019, @10:20PM
He's working on NTPsec [ntpsec.org].
(Score: 0) by Anonymous Coward on Thursday July 04 2019, @07:05PM
Chrony (https://chrony.tuxfamily.org/)
(Score: 2, Interesting) by Anonymous Coward on Thursday July 04 2019, @07:47PM (6 children)
There is also Arcan (stuff written for X11 or Wayland will run on it). It seems to be much more sanely designed than Wayland (for instance, network transparency works, and they didn't go full retard like wayland with every "window manager" needing to implement the equivalent of nearly the entirety of X11-- but I guess wayland proponents wanted a large barrier to entry to force people to not have choices, and didn't think a zillion lines of code duplication and the bugs associated with it were as important as limiting user choice (fuck redshat).
My guess is that with redshat backing wayland, they will pull the same games as with systemd to make stuff they have suffient influence over depend on wayland in an attempt to force other distros to use it. But, if Arcan can maintain wayland compatibility, then redshat might be foiled.
https://arcan-fe.com/2018/05/31/revisiting-the-arcan-project/ [arcan-fe.com]
https://github.com/letoram/arcan [github.com]
(Score: 0) by Anonymous Coward on Thursday July 04 2019, @09:42PM (3 children)
That is a single person's vanity project, iirc.
Beyond that your comment seems to be on point. And things are likely to get worse now that Big Blue is finalizing its buyout of Red Hat.
(Score: 5, Touché) by c0lo on Friday July 05 2019, @02:11AM (2 children)
What do you think Linux was in the beginning?
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 1, Interesting) by Anonymous Coward on Saturday July 06 2019, @06:18PM (1 child)
Torvalds was more than happy to accept patches from day one. The Arcan dev seems not so inclined and more into staking his own fief.
(Score: 0) by Anonymous Coward on Sunday July 07 2019, @01:00PM
Except there is nothing to support that claim? there is literary a page on where to contribute? Seems much more like ’I do this for these reasons, other than that, IDGAF’ which is just what Linus mail read like. If you want actual jonestown cult building, look at wlroots.
(Score: 0) by Anonymous Coward on Friday July 05 2019, @11:13PM
Arcan is really really interesting. I spent a big chunk on the wiki alone and a fair bit more on the demos. The technical detail behind how it survives and rebuilds itself on crashes and how that was a foundation for the network transparency is jaw dropping. Found this https://gfycat.com/totalsecondhanddoctorfish [gfycat.com] that the author had posted on IRC. Bad video but that is damn transparent networking, literally drag and drop between machines.
(Score: 2) by DeVilla on Sunday July 07 2019, @04:21AM
I'm still dreading the forced client side decorations. I already hate when a gnome3 window pops up for some reason and doesn't implement something like "window shading" or "no raise on click" correctly because they chose to write extra code to not cooperate with my desktop environment. The wayland world is going to break some many things I take for granted.
Does Wayland have the windows behavior of not let you move a window if the program is hung the way windows did? I know most of the current apps that are using client side decorations currently seem to turn "invisible" when they program hangs. At least with "normal" X1 apps you get the window boards if not a gray window.
(Score: 2, Interesting) by Anonymous Coward on Friday July 05 2019, @08:46AM
Speaking of NTP...
Notice how NTP's timeservers are all DNS based? To be clear, what's really happening is people are donating timeserver services, and they're using the NTP.org domain to do it.
Now notice how ntp.org is now DNSSEC based?
What happens if the time is too far off, with DNSSEC based records? That's right, they DO NOT WORK!
So if your clock is out of sync too far, and we're not talking even weeks here, more like days or EVEN hours depending on the state of the DNSSEC record, you can't use NTP to update your time, because NO DNS RECORD!
What fool thought DNSSEC for NTP servers was a good idea, I'll never know. But it's indicative of people not thinking things through today. DNSSEC! SECURE!!!
And of little use, if you can't get the record.
And yes, you can disablr in your DNS server, eg bind, but of what use is that? Then why have it?
So now I have to use IP addresses on all my servers, and hope they don't change... otherwise no NTP for me!
(Score: 0) by Anonymous Coward on Thursday July 04 2019, @09:40PM
X11 will probably stand and fall with Keith Packard. Much like Linus Torvalds, he has been sticking with one project for much of his life. And that project being some variant or other of the X protocol. He event was brought in as a consultant by Valve to work on bringing VR headsets to X11.
The rest that has worked on it in more recent years have been far too hung up on frames pr second and the narrow use cases of the major DEs.
(Score: 4, Interesting) by Anonymous Coward on Thursday July 04 2019, @10:55PM
Keith Packard was one of the major fork proponents, and if you go and look at his pedegree, he was with X since X11R3 or so, before they had even formalized it into a hardware agnostic protocol (Early X versions had all sort of endianness and other issues, as well as major performance regressions when run on a big or little endian platform for a feature expecting the other. Most features were big endian at the time, which caused issues on little endian platforms. This all got fixed between R3 and R6, but there were many incompatible breakages along the way.
Personally I think the best solution at this point is an X11R7 that fixes some of the issues in the protocol (particularly around synchronization performance) and fork off new revisions of the libraries that allow some new function calls (stubbed out but added to the older R6 Xlib fork, so apps can be API compatible at the expense of performance, or just do that in the toolkits if anyone would actually support backwards compatibility...)
End result: X11R7 gets performance gains comparable to wayland, backwards compatibility can be maintained, albeit along a slower codepath, and new optimized programs can run at speeds that couldn't be obtained with the previous synchronization features.
A tangentally related idea: It's time to reimplement Display Postscript. It would actually be useful, performant, and scalable enough on modern GPGPUs, would allow wysiwig screen to paper printing, and it would negate a whole slew of kludges in modern display toolkits caused by relying on raster graphics for everything.
Just my 2 cents as a long time user and a sage of esoteric display subsystems.