Skip to main content ↑ ↓
Topic: A fork in the road and the path not taken (Read 3004 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

A fork in the road and the path not taken

After giving consideration to the discussion in this thread on removing wayland from an artix installation:  https://forum.artixlinux.org/index.php/topic,8392.0.html

I've come to think that we might be finally approaching a fork in the road between different camps in the Linux ecosystem.  Obviously, systemd was a breaking point for many developers and users.  The major thrust of that project that provoked a counter reaction was that systemd was not just complex, but was shoehorning everyone else into the vision of a few coders and admins, mostly but not exclusively coming from redhat.  Systemd was a new paradigm that was monolithic and demanded control. 

As systemd broke into the econsystem and distro after distro adopted it, I had many conversations with other GNU/Linux enthusiasts and professionals.  It was interesting that RMS had a rather aloof view of events.  I think he is down on Linux in the general and felt he had bigger fish to fry.  But my friend, Rick Moen, from Menlo Park, Ca (home of facebook, Jerry Garcia and a small train station), was quite adamant that he had no need for it and saw it as a growth of the entire Free Desktop project, which he really no need for in his view and wanted no part of. 

Rick is a debian user, and once I decided I wanted to break from the systemd design and got disgusted with the rancor that followed Pottering et al, I searched for another system to use.  Rick insisted, and demonstrated,  that there was no need for another distro and that a systemd free system was simple enough to make with plain debian with the right aptget magic.  He was very unhappy with the Free Dektop group and saw them as the root cause of the systemd split.  He saw a number of initial concessions that Gnome gave to systemd that pushed systemd along.  And this has continued to be the case as we can read here:  https://blogs.gnome.org/benzea/2019/10/01/gnome-3-34-is-now-managed-using-systemd/

I personally dislike debian and didn't really agree.  I felt that even if it was true, there are directions in which systemd is growing that software which conformed to its design would eventually lead to a virtual lock in to the the systemd ecosystem.  I was ready, at that point to just through the baby out with the bath water (an American proverb for foreign readers)...as I did with Windows in 1998.  I hated it so much I didn't want to use it.  I searched out first slackware and found that it just didn't seem to work on my modern equipment any longer.  I was very disappointed.   I moved on to looking at Devuan.  Devuan is debian based and the developers where very difficult for me to work with.   I considered that a hostile environment.  I've never been a fan of the Debian culture.

Then I found Manjaro's systemd free system and that lead eventually to Artix.

Now the Artix developers can be a frisky bunch, like many successful Free Software projects.  Artoo has raked me several times both in private and in public, and he has strong political opinions.  But he has lead a group that has been responsive to users, avoided chewing off more than they could do,  or the political wrangling that has buried so many projects.  But there has been this dangling problem of increasingly incompatibility with systemd and the rest of the world, and Rick Moen proved to be right... much of this comes from the Free Desktop organization...and now Wayland.

After reading about the vaporware threats of gtk5 and the problems with Plasma when used  on a purely X based distribution, I am not sure how much longer we will be able to patch systems together without systemd without addressing a larger fork that would include major edits in large codebases like Plasma.  And if so,  where is the manpower going to come from to support it?

I see the Xlibre fork as an opportunity.  I think Artix should commit to an ISO which could support both X and Wayland as simply as it does with dinit and openrc.  I don't know how I can contribute to that, and it might well be that a pure Xlibre iso would not, in the long run, run a gnome or kde based DMs (which is asking and says a lot, I know).  But I think the need for a fork of the ecosystem is coming soon.  Systemd systems with their client side intrusions, are reaching a critical stage now where they are not a "Linux" OSs any longer.  And the software made for a systemd OS  specifically just may no longer be compatible with a traditional GNU/Linux environment and I consider that a detriment to my digital life. 

I'd like to make one more point about this evolution. I am not concern about the popularity of these decisions.  I am already a minority member of the computing public and I have been since Linus was an undergraduate.  It sort of makes me chuckle when systemd people think they have mindshare... as they say in Brooklyn... they have GOTS. 

The entire Linux installation base is a small percentage of the MS install base, or the Apple install base.  Android might dominate the overall install base, but I would hardly call that Free Software and it is not a versatile computing system, like  a Desktop  or server OS.  Linux made up 2.5% of the worldwide website hits (however that is guestimated) in February 2025.  So where does that put Systemd, Wayland and the Free Desktop :)

On Desktops, MS is about 74% and 15% is Macs.   So for systemd developers to point at communities like the Artix community has fringe nutters... well ... aint that calling the Kettle Black?  None of us are normal buy this measure.  I would be happy to be more abnormal than most if I was happy with the technology I was using and that is true about Wayland and Free Desktop software as well (which increasingly make me unhappy and feel hamstrung). 

https://www.youtube.com/watch?v=rubapQ221Jc

Re: A fork in the road and the path not taken

Reply #1 –
I don't think dropping Wayland would benefit anyone or hurt FDO for that matter. Most software and people will flow with the stream, whether we like it or not, and only conscious users will actively choose their underlying init and display server. The rest won't give a damn, unless inconvenienced by bad gaming performance or system failure. To them, it doesn't matter if their software is written by an <insert strong word here> or an <insert another strong word here>, so long as it works.

Perhaps one day we'll be forced to make a choice of display server (as a distribution I mean, for I have already made mine), but I'm not getting that sense of impending doom, yet.

Re: A fork in the road and the path not taken

Reply #2 –
Quote from: nous – on
I don't think dropping Wayland would benefit anyone

I am not suggesting that.  But I am suggesting that we (and when I say we, or course I mean you...which is sad) commit to a full iso based on X in addition to those we already support that are increasingly dependent on wayland.  We should try to make sure all the packages work equally well, as much as possible, with both environments.

I know it is a big ask.  I also know that despite the majority of users who "just want things to work (then why are you using Windows... are you stupid?)"  that there is a core audience for the flexibility of X.  Losing that flexibility would be an extinction event for a certain form of computing, and further pushes people into a systemd ecosystem.

Is that making any sense you you?

Reuvain

Re: A fork in the road and the path not taken

Reply #3 –
Quote from: mrbrklyn – on
Quote from: nous – on
I don't think dropping Wayland would benefit anyone

I am not suggesting that.  But I am suggesting that we (and when I say we, or course I mean you...which is sad) commit to a full iso based on X in addition to those we already support that are increasingly dependent on wayland.  We should try to make sure all the packages work equally well, as much as possible, with both environments.

I know it is a big ask.  I also know that despite the majority of users who "just want things to work (then why are you using Windows... are you stupid?)"  that there is a core audience for the flexibility of X.  Losing that flexibility would be an extinction event for a certain form of computing, and further pushes people into a systemd ecosystem.

Is that making any sense you you?

Reuvain

I can't envision what an "X ISO" would be, other than shipping a DE/WM that uses X, which I'm pretty sure it's all the ISOs already, except for base (because it doesn't come with a graphical environment) and maybe Plasma.

Re: A fork in the road and the path not taken

Reply #4 –
I think Artix project currently is doing the right thing, provide choices (including init systems, display servers/compositors, etc) and let the users decide what they want to use. I applaud the Artix dev for picking up xlibre as soon as it's available.

Re: A fork in the road and the path not taken

Reply #5 –
Quote from: nous – on
Perhaps one day we'll be forced to make a choice of display server (as a distribution I mean, for I have already made mine), but I'm not getting that sense of impending doom, yet.

Would love to know what your choice is (assuming its not Artix of course).
Supercalifragilisticexpialidocious

Re: A fork in the road and the path not taken

Reply #6 –
Unless upstream [and I mean the big corporation(s) behind FDO] try to lock out "dissenting" people from Wayland or Xorg, I don't see any reason to change anything in our current packaging pipeline. We are fully aware some packages unnecessarily depend on Wayland, but as long as they don't force users to switch to it, we shouldn't intervene.

@cds I wasn't clear perhaps, I meant I've already switched to xlibre.  8)

Re: A fork in the road and the path not taken

Reply #7 –
ever thought revamp the community qt and gtk ISO options of giving users a choice during installation to let users want to have since we already abandon entire systemd for openrc but its not easy for users to pick sides when other iso's do take lot your time to make each init than not do what devuan did of having 4 init's options during installation and we can put 2 window manager options for users to pick. XLibre or Wayland. We need ask Vaxry since i know even he get attack by wayland from Red Hat Freedesktop org that they ban him to participate any freedesktop that he became on his owned. We could ask him to join us as XLibre did. His C++ skills will benefit us greatly since i notice huge uptick in our ARTIX community interest lately and shown by distrowatch we are ranking higher than 220 clicks than 77 clicks 6 months ago.  Thats over 300% that something we never seen anything like this before since We all separate manjaro. Think About it, the skies have no limits what we can do for entire community that YES WE CAN MAKE BEST DISTRO ...WE JUST GONNA HAVE TO MAKE IT VERY SIMPLE FOR ANY AVG USER WANT TO GIVE A TRY OUR TAKE. I JUST DON'T CARE WHAT FOSS or any Freedesktop says now since they all are in bed with Red hat corpo team.

The reason i say this is Dinit is becoming much mature and much powerful than we all thought in first place. It even surpass S6-init that we suppose to believe its our golden perfect child but when Dinit came around its very similar in simplistic as systemd. Similar language what systemd speaks without the bloat in it. Its not hard to learned when it took me 15 minutes to learned vs S6 takes me a week to understand it but i never like 2 part init when dinit does it all in one part loader without extra commands. Its suggestion i ask if you think openrc should be replace for Dinit as main community releases and continue make separate smaller isos that seem requires lot effort to make complete desktop. Or we can combine all of them into community mega iso so users don't have chase too many packages over pacman manager. And i really want to do something about replacing Octopi into our owned gui packager that not crash often and Not cause more headaches for any arch users since cachyos packaging works quite well for artix setup. Better than expected in performance with less memory usage than systemd systems




Re: A fork in the road and the path not taken

Reply #8 –
Quote
Its not hard to learned when it took me 15 minutes to learned vs S6 takes me a week to understand it

In fact, Artix wiki makes s6 much easier to understand and follow. It took me some time to read documents on skarnet to understand s6, and took me even longer to read obarun documents to learn s6 & 66. :)