After a few days, I've gathered over 100 replies to The First Hyprland Census. I've asked a few questions to better direct Hyprland's development. Now, with all those replies, The report is done. It's available in PDF format if you prefer it like that, or, you can read it below. Thank you to everyone who has participated, and I hope the future will bring even more good things to Hyprland. But for now, here's the report.
Download the report in a PDF format here.
The First Hyprland Census
Since 09.08.2022, until 12.08.2022, I have received a total of 106 replies to the First Hyprland Census. Today, I’d like to share the results and at the same time address a lot of the reasons some people have pointed out as dealbreakers to their adoption of Hyprland into their daily usage stack.
Demographics
First of all, let me go over the demographics. From those 106 respondents,
82.1% were men. This is not surprising at all, considering men are more likely to be tech-savvy enough to even find Hyprland. The percentages here match closely what one would expect out of a project like Hyprland.
The age breakdown also showed no surprises. Most of our userbase, 54%, are young adults, likely students. Over 80% are people below 26, which makes complete sense, as people that age are more likely to experiment with software to begin with, but also have enough time on their hands to even be able to do so.
Question 1
With the demographics out of the way, we can proceed to the actual important parts.
Perhaps unsuprisingly, 71% of people have found Hyprland through r/unixporn. With the lack of any real recognition (yet) outside of that subreddit, it was expected to be that way. However, even with the lack of any major coverage on Youtube, 7% of people still somehow first found out about Hyprland there. The rest of mediums are mostly insignificant.
Question 2
I am pleasantly surprised to say that 46% of respondents actually daily drive Hyprland as their only display server of choice. Although bugs are very rare to me, the amount of criticism Hyprland has received over the time for bugs and instability made me feel like it was a major roadblock for most people. Apparently, not so much. We can only hope this statistic will improve as time goes on and more and more bugs get ironed out.
Now I’d like to take a moment to reply to a select few of comments from people who answered “No” to the question above.
“Battery drain is too high compared to other Wayland compositors or even DEs like KDE or Gnome.”
Battery life on my new Zenbook Pro 14” has been really good, and according to reviews online, matching Windows. It is important to remember to always optimize your system. Enabling new blur optimizations for Hyprland, or disabling blur altogether will help a lot. Also, remember to use something like tlp and auto-cpufreq to further optimize Linux itself for running on a laptop. This has made my laptop last around 7-9h with screen on 50% brightness, which I consider really good.
“XWayland windows have blurry scaling, and I have to use some X applications for work.”
That is unfortunately an issue with wlroots itself. We’re waiting for a Merge Request to be accepted that would address this issue.
“I'm on proprietary nvidia drivers and on any wlroot compositors except sway for some reason, there is screen flickering.”
“Flickering on nvidia”
“nvidia moment”
Nvidia has always been a major pain on Linux. However, recently, Nvidia is taking steps to do something about it, seems like. Some of those issues can be fixed by following the instructions in the Nvidia wiki page, (you can find it in the sidebar in the Hyprland wiki) but not all issues can be resolved, unfortunately.
“Lack some stuff, touch compatibility, layout (master/slave)”
Someone has not updated in a year, it seems.
“Not stable yet. Looking at the massive changelogs and commit history it's very much still in development, and I don't like to deal with the growing pains.”
Depends on your definition of stability. The config is mostly stable, and breaking changes are rare nowadays. Usually, if your config works right now in a way you like, it’s not going to change for a while, and even if it does, I try to keep it easy to make a small change in the config and make it work again.
“Too many rough edges to be used exclusively. Stability is important when using in a productive environment!”
What rough edges? It’s important to open issues if you find something that bothers you. I, and many others, use Hyprland exclusively in a productive environment.
“The blur is very poorly optimized, as compared to newm (which I daily drive) its not even 1% smooth overall”
Straight-up wrong. Newm, just like Hyprland, Wayfire, KDE, and many others use the same shader to render blur. Kawase. The method used by wayfire is pretty much the same as Hyprland uses, and newm’s is very similar. However, once you turn on the new blur optimizations on Hyprland, blur causes 0 rendering overhead on all tiled windows. Such optimization method does not exist in newm or wayfire, so this claim is completely false.
“I need some features that at the moment are not present”
Thank you for explaining those features in great depth.
“I've had difficulties with setting up per device sensitivity an a touchpad and trackpoint”
You can always ask on Discord or open an issue. We’re here to help.
Question 3
The best part of Hyprland, perhaps unsuprisingly, was the eyecandy. I was not surprised to see that, however, I am pleasantly reminded that people also love the configurability and the two sockets Hyprland exposes.
Question 4
With the worst part, it’s pretty clear that bugs are indeed a problem. What I’ve expected has been shown here clear as day. The rest of the issues are pretty uniform, but “difficult to say” and “there is none” are leading after bugs, so it seems that fixing most of the annoying bugs should make most of the people happy.
Question 5
As expected, since bugs were a major problem in Question 4, here as well “Fixing bugs” is the top answer. However, I am a bit surprised by the 17% of respondents that selected “adding more options”, as I’d say Hyprland’s configurability is already pretty high. Optimizing the code was nothing unusual, as most people think that you can optimize the code forever, and get something that uses 0% of everything and 0MB of RAM if you try enough. Unfortunately, optimizing a lot of things in Hyprland would result in either bad code for very small performance gains, or many, many hours of problems and headaches. For now, I try to focus on stability over the 12MB of RAM that I could possibly use less. Supporting protocols is another interesting one, as I can’t think of many protocols at the moment that Hyprland does not support, but are useful. It might be people who don’t really know what a protocol is, and thought “ah, X11!”.
Summary
Summarizing the results, it’s clear that what people want is more bugs fixed, and thus it will become the top priority for the nearest future. Of course, more functionality is also important, but it’s important to have a stable environment to be even able to appreciate the added functionality.
It’s also become clear that unfortunately a lot of people ignore or skim over the wiki. It’s a very important, even central place for knowledge regarding Hyprland, and the fact it’s being ignored so much is worrying. Perhaps a better structuring could be done to better appeal to people.
Another thing that can boost the knowledge of Hyprland is perhaps a youtube video demonstrating inportant parts, as a lot of people for some reason prefer those over reading.
At any rate, it’s important we continue to listen to our users, apply our own expertise on top of that and then try to create something all of us can enjoy.
Thank you to everyone who has participated.