The most exhausting part about the whole X/Wayland debate for me is the condescending tone that's coming usually (though not exclusively) from the Wayland side.
The list of "wayland.fyi approved" software is extremely interesting, but is it really too much to ask to present it in a neutral fashion that doesn't make me want to switch back to X out of spite?
Indeed what an embarassing attempt of blog post, the part that got me to close the tab was:
Phoenix is a waste-of-time project (while quite impressive), rewriting a piece-of-shit protocol in a language with a new breakingchange every 5 minutes is not a good idea
Gee, the audacity to call someone else's project a waste of time, after just writing an article with about the same technical rigour and vagueness as a zodiac sign reading. I love dunking on Zig as much as the next guy (let me have unused variables for Christ's sake), but this is a new low, even for Wayland discussion.
Anyone who has checked out phoenix or any of the author's other repos sees that they know what the fuck they are talking about, and also have the battle scars to prove it. The phoenix FAQ contains a detailed writeup with the reasons behind the project, which are conveniently dismissed for what, a programming language knee-jerk?
The fact that this low-effort rant even got its own domain, might challenge probonopd's github gist of hell for the title of most unhinged post on the topic yet.
This post oozes the "no you're the worst!" vibe, and a big portion of it seems to exist only to highlight how bad X11 is. Almost at the very end is a list of libraries/tools that actually are what the SIG is supporting. Based on the bashing/supporting ratio, it seems the group is not so much about minimalist wayland, but more about how bad something else is. It's a pity, because wayland is not really perfect either, and I'd hoped they would have focused on making those rough edges go away.
Yeah I feel like this sort of messaging is often a bit of an own-goal. You spend more time talking about the competing technology than building a case for your own, overall the negativity is also very unappealing. A constructive framing is much more appealing. If you're building the future, that's enthusiasm I may want to be part of.
The author and contributors of this website, and of the many projects listed, are clearly also fans of 9front (hence the plan9 cursor and involvement with 9larp.net). My take is that the "attitude" in this blog is derivitive (pun intended) of the kind of thing you might find over on cat-v.org. So depending on your perspectibve, this could read as pure, unfiltered snobbery (and this could still be the case), or as a cat-v inspred, cynical and sarcastic attempt at humor, while still remaining righteously opinionated.
I thought I'd offer a bit of background, since I wrote swc and wld, which many of these projects are based on.
I started swc back in 2012, when wayland was a very new and exciting thing, and wanted to port my window manager to it. At the time, there were no options to do this. Other than occasional updates to keep stuff working, it has mostly stayed dormant since 2015 as my focus shifted to other projects. It lacks a bunch of new protocols, but mostly works ok with the clients I use.
Personally, I think wlroots is quite comprehensive and is pretty good at what it does. I also don't have very strong feelings about X11. I think what wayland really did well is provide the means to more easily build new display servers, and to easily extend the protocol. There's no server-side equivalent of xcb, so with X11 you'd have to implement all the protocol bits yourself. Now with the dri3 and present extensions, the display pipeline is quite similar, and in my opinion the other aspects of wayland are less compelling.
So while I don't agree with most of the criticism or how it's presented (I prefer to promote software I like), I do believe that this group is well-intentioned and passionate, though maybe a bit naive. I think they were looking for a small hackable wayland codebase, and I'm happy to see that they found it in swc. They are working to implement protocols like wlr-layer-shell, subsurfaces, and more, which I've wanted to do for a long time but never found the time.
This whole site should've just been a awesome-minimalist-wayland repo on GitHub, the commentary on the page actively detracts from what is supposed to be the content.
I think this page is largely missing the point of the main complaint I see against Wayland? Namely, that it punts too much to the compositor. Showing off its minimalism really isn't going to win people over.
I'm a bit surprised river is not included on this list, because breaking down monolithic Wayland compositor design into smaller, more focused components seems more in line with a minimalist philosophy? I guess maybe it has too many features implemented for it to appeal to the webmaster?
This feels a little cringe teenager. I feel we should be accepting of a little cringe because we were all teenagers once, and it’s good if new teenagers are getting passionate in this space. But I’ve no idea who actually wrote this and if they’re a grown adult or not.
Because X supports it and a ton of the web standards are copying X's features.
...I'm kinda kidding here but also kinda not. Tons of the web standards today trace back to Netscape Navigator's features, and those frequently did just forward what X did to the web javascript. For example, look at historical mouse handling in the web. Old Mozilla set MouseEvent.which to 1 = left, 2 = middle, 3 = right. Same as X.
(then the W3C looked at that and said, that's silly, let's make it outright stupid instead and dictated that left button should be... 0, but kept the others in order. So now you gotta make sure you don't misinterpret a value of 0 to mean left since sometimes it means none. That remains the standard today, though the property name changed to button, and now if you check the value of buttons, with the s, it has to use different values so they can be combined! Meanwhile, the chad Internet Explorer, copying what Windows does, had bitmaskable values the whole time.... until they yielded to "standards compliance". Sigh.)
The CSS color names? Directly based on X's color names. I even kinda think the css content-box model might come, in part, from X, given how if you create a WxH window, the window manager then adds a border and such to the outside, but... I'm probably stretching a legitimate coincidence there. Still, it is different on old IE, and Windows also does it more like how IE does it (nowadays called border-box), and this MIGHT be part of why.
There's some other things like how focus in/out works being similar to X, but I'm too lazy to fact check it all. Regardless, there's obviously some influence from X on the web standards, so I'm only half kidding when I say websites can set custom mouse cursors because X applications can. I don't think that's the actual real world reason, Microsoft shipped it first here... I think so it is probably because Windows lets you do it, but I betcha if X didn't let you, Firefox would never have shipped it and it wouldn't have survived to today, so I do half seriously think it is allowed because X allows it.
sny | 17 hours ago
The most exhausting part about the whole X/Wayland debate for me is the condescending tone that's coming usually (though not exclusively) from the Wayland side.
The list of "wayland.fyi approved" software is extremely interesting, but is it really too much to ask to present it in a neutral fashion that doesn't make me want to switch back to X out of spite?
vimpostor | 13 hours ago
Indeed what an embarassing attempt of blog post, the part that got me to close the tab was:
Gee, the audacity to call someone else's project a waste of time, after just writing an article with about the same technical rigour and vagueness as a zodiac sign reading. I love dunking on Zig as much as the next guy (let me have unused variables for Christ's sake), but this is a new low, even for Wayland discussion.
Anyone who has checked out phoenix or any of the author's other repos sees that they know what the fuck they are talking about, and also have the battle scars to prove it. The phoenix FAQ contains a detailed writeup with the reasons behind the project, which are conveniently dismissed for what, a programming language knee-jerk?
The fact that this low-effort rant even got its own domain, might challenge probonopd's github gist of hell for the title of most unhinged post on the topic yet.
kamme | 14 hours ago
This post oozes the "no you're the worst!" vibe, and a big portion of it seems to exist only to highlight how bad X11 is. Almost at the very end is a list of libraries/tools that actually are what the SIG is supporting. Based on the bashing/supporting ratio, it seems the group is not so much about minimalist wayland, but more about how bad something else is. It's a pity, because wayland is not really perfect either, and I'd hoped they would have focused on making those rough edges go away.
marginalia | 14 hours ago
Yeah I feel like this sort of messaging is often a bit of an own-goal. You spend more time talking about the competing technology than building a case for your own, overall the negativity is also very unappealing. A constructive framing is much more appealing. If you're building the future, that's enthusiasm I may want to be part of.
pfr | 10 hours ago
The author and contributors of this website, and of the many projects listed, are clearly also fans of 9front (hence the plan9 cursor and involvement with 9larp.net). My take is that the "attitude" in this blog is derivitive (pun intended) of the kind of thing you might find over on cat-v.org. So depending on your perspectibve, this could read as pure, unfiltered snobbery (and this could still be the case), or as a cat-v inspred, cynical and sarcastic attempt at humor, while still remaining righteously opinionated.
mcf | 3 hours ago
I thought I'd offer a bit of background, since I wrote swc and wld, which many of these projects are based on.
I started swc back in 2012, when wayland was a very new and exciting thing, and wanted to port my window manager to it. At the time, there were no options to do this. Other than occasional updates to keep stuff working, it has mostly stayed dormant since 2015 as my focus shifted to other projects. It lacks a bunch of new protocols, but mostly works ok with the clients I use.
Personally, I think wlroots is quite comprehensive and is pretty good at what it does. I also don't have very strong feelings about X11. I think what wayland really did well is provide the means to more easily build new display servers, and to easily extend the protocol. There's no server-side equivalent of xcb, so with X11 you'd have to implement all the protocol bits yourself. Now with the dri3 and present extensions, the display pipeline is quite similar, and in my opinion the other aspects of wayland are less compelling.
So while I don't agree with most of the criticism or how it's presented (I prefer to promote software I like), I do believe that this group is well-intentioned and passionate, though maybe a bit naive. I think they were looking for a small hackable wayland codebase, and I'm happy to see that they found it in swc. They are working to implement protocols like wlr-layer-shell, subsurfaces, and more, which I've wanted to do for a long time but never found the time.
gonz | 5 hours ago
This whole site should've just been a
awesome-minimalist-waylandrepo on GitHub, the commentary on the page actively detracts from what is supposed to be the content.rbuchberger | 3 hours ago
As annoying as this particular website is, handing a small and independent one over to a gigantic company is pretty much always a bad thing.
Relax | 12 hours ago
I think this page is largely missing the point of the main complaint I see against Wayland? Namely, that it punts too much to the compositor. Showing off its minimalism really isn't going to win people over.
sakurina | 12 hours ago
I'm a bit surprised river is not included on this list, because breaking down monolithic Wayland compositor design into smaller, more focused components seems more in line with a minimalist philosophy? I guess maybe it has too many features implemented for it to appeal to the webmaster?
FreeFull | 10 hours ago
Probably because river depends on wlroots, and the wayland.fyi folks don't seem to be fans of how big wlroots is
jmtd | 4 hours ago
This feels a little cringe teenager. I feel we should be accepting of a little cringe because we were all teenagers once, and it’s good if new teenagers are getting passionate in this space. But I’ve no idea who actually wrote this and if they’re a grown adult or not.
z3bra | 51 minutes ago
I'd lean more toward the teenager side. The author also made a video about this topic, so see for yourself : Make Wayland great again.
jlarocco | 18 hours ago
Why are websites even allowed to set a custom mouse cursor?
adam_d_ruppe | 11 hours ago
Because X supports it and a ton of the web standards are copying X's features.
...I'm kinda kidding here but also kinda not. Tons of the web standards today trace back to Netscape Navigator's features, and those frequently did just forward what X did to the web javascript. For example, look at historical mouse handling in the web. Old Mozilla set MouseEvent.which to 1 = left, 2 = middle, 3 = right. Same as X.
(then the W3C looked at that and said, that's silly, let's make it outright stupid instead and dictated that left button should be... 0, but kept the others in order. So now you gotta make sure you don't misinterpret a value of 0 to mean left since sometimes it means none. That remains the standard today, though the property name changed to
button, and now if you check the value ofbuttons, with the s, it has to use different values so they can be combined! Meanwhile, the chad Internet Explorer, copying what Windows does, had bitmaskable values the whole time.... until they yielded to "standards compliance". Sigh.)The CSS color names? Directly based on X's color names. I even kinda think the css content-box model might come, in part, from X, given how if you create a WxH window, the window manager then adds a border and such to the outside, but... I'm probably stretching a legitimate coincidence there. Still, it is different on old IE, and Windows also does it more like how IE does it (nowadays called border-box), and this MIGHT be part of why.
There's some other things like how focus in/out works being similar to X, but I'm too lazy to fact check it all. Regardless, there's obviously some influence from X on the web standards, so I'm only half kidding when I say websites can set custom mouse cursors because X applications can. I don't think that's the actual real world reason, Microsoft shipped it first here... I think so it is probably because Windows lets you do it, but I betcha if X didn't let you, Firefox would never have shipped it and it wouldn't have survived to today, so I do half seriously think it is allowed because X allows it.
refi64 | 16 hours ago
Is this rio's cursor? Haven't sen that in quite a while.