1. 50

  2. 24

    This guy is such a champ.

    Found something that he wishes was better and took it upon himself (for over a year!) to research solutions and organize work to help loads of people, for free.

    1. 4

      I just signed up here today to thank you for this comment. It has been added it to my list I look at to cheer me up when I’m having a hard day.

      It’s been an interesting process to motivate myself to work on a willpower-intensive goal as long as this one. Especially since the end result is uncertain. Even the issue continuing to get attention is uncertain. So it’s really hard to get motivated sometimes. But the enthusiasm of people like you helps carry me, and I’m optimistic that this long-term goal will not continue forever. Ideally it will continue < 2 years, and continue to help Linux adoption thereafter. :D

      1. 1

        Oh no, thank you!

        If I was a better person I’d be helping all the open source projects out there that I use all the time. (Hint: I’m not) But you are! I was really impressed with your no-nonsense practicality

        I hope you’re getting the help and support you need. Maybe Ubuntu or Red Hat would sponsor you if you asked - they definitely should.

    2. 17

      IMO this is a potentially great time for desktop Linux. I’m not suggesting the entire Mac fanbase is abandoning ship, but there are a sizable number of us who are growing more and more frustrated with the directions Apple is taking both in terms of hardware and software for the Mac line.

      The author nailed it - What desktop Linux needs right now is polish.

      1. 7

        I don’t think you were being ironic but I kinda love that you’re saying 2019 Is The Year Of Linux On The Desktop ;-)

        1. 5

          The reason the year of the Linux desktop has become a joke is exactly because of the point I made - it very much lacks polish. Polish is what end users demand in their environments.

          Windows notoriously used to lack polish in a big way, but that’s changing.

          Linux is getting better too but slowly because there are 2 major and about a grillion minor desktop players, and they can barely agree on how to make things like cut and paste work, much less offer things like desktop app scripting and integration or consistent theming and accessibility.

          1. 2

            and they can barely agree on how to make things like cut and paste work

            I don’t think that’s a fair complaint. Selection copies, middle click or ‘Insert’ key pastes. It’s different than Windows and OSX, but it’s a different OS, so some things will just work differently.

            Copy and paste of graphics or formatted text is more complicated, but support for that is hit or miss on OSX and Windows too.

            1. 1

              I do. That’s the X11 accepted standard. Different desktop environments and applications had their own. The freedesktop.org folks have improved the status quo here markedly, but it used to be a mess.

              Also, this was one example of the kind of polish I mean. You can invalidate my example, but my point still stands.

          2. 1

            I’ve been using Linux on a desktop for many years, at this point I think that Linux on the Desktop has already arrived.

            What I am still waiting for is the Year of Linux on the Laptop. It seems like the trackpad driver is the final problem to be solved, and I’m really glad to hear about this new project.

        2. 10

          Reading this and realising crummy Linux trackpads is a big part of what drove me to tiling WMs, many years ago.

          (Off topic, but) I wonder if a similar effort for Linux desktop responsiveness would be popular. Heavy background system load (I/O or CPU) is still able to drive my i7-6500 driven X11 interface to multi-second-latency for mouse clicks and keypresses. Have tried -ck “desktop latency” kernels and currently using -zen kernels, they’re a little better but it still happens to me most weeks.

          1. 5

            Yep, it’s a problem. When I/O load starts maxing out, my X11 just completely locks up also (AMD Phenom II here, so slightly older but still beefy enough by modern standards.)

            Interestingly as un-mature as Haiku is in certain respects, I’ve never had this kind of lockup on it. But the kernel schedulers were specifically designed for GUI usecases, so, that is probably a large part of it. I don’t think Linux will start really prioritizing that anytime soon.

            1. 4

              I definitely have the same issues, and have since I started using Linux. It feels like it’s only gotten worse, to the point on lower end systems, it can take 30 minutes to get to a VTY to kill the program causing the lag.

              1. 3

                Ouch! I’ve been there where it takes maybe 20 seconds to get to the VTY, but not 30 minutes.

              2. 3

                Heavy background system load (I/O or CPU) is still able to drive my i7-6500 driven X11 interface to multi-second-latency for mouse clicks and keypresses.

                Wow. I don’t think I’ve ever experienced this. I use a stock kernel. Some thoughts/questions:

                1. Are you sure you aren’t running out of RAM? (Or swapping, if that’s even a thing any more.)
                2. What window manager and/or desktop environment are you using?
                3. Does the latency occur in all applications it’s just specific ones?
                1. 2

                  I was half hoping someone else might have some clues for this. :)

                  1. Fairly certain. I have 16GB of RAM and I’ve experimented with and without a 4GB swap partition. Have checked free -m during/after such events and usually more than half is free.
                  2. i3 window manager, no desktop environment (I run xsettingsd). Have wondered if maybe this is somehow too minimal and I’m missing something that would make it better…
                  3. Everything, particularly in complex apps like browsers but even in terminals/emacs or pressing the i3 keys to switch workspaces becomes laggy. I have wondered if it may be related to Intel embedded graphics (HD520+modesetting driver).
                  1. 5

                    I’m stumped. I run a similarly stripped down environment (Wingo as my WM, no DE, but it’s a “classic” non-compositing WM just like i3) across many machines. They run the gamut from i3 to i7 Intel CPUs, with between 8GB and 64GB of RAM. Some of them use a AMD graphics card to drive three monitors and others just use the Intel embedded graphics to drive one monitor. None of them experience lag like what you’re describing. (When I used to use Chrome, it could be laggy at times, but that was specific to Chrome.)

                    The only other thing that might be different between our setups is that I run a compositor (compton) on top of my WM, mostly to smooth things out and support transparent windows. But I don’t quite understand how this could eliminate the lag you’re seeing, but maybe worth a try?

                    The only other lag I can think of is that sometimes my WM lags a tiny but perceptible amount when switching workspaces where one of the workspaces has a lot of windows on it.

                    What terminal emulator do you use? I use Alacritty now (with tmux), but I used to use Konsole from KDE, and I don’t really notice any lag difference (that is, neither lags for me).

                    Lag sucks though. Bummer. Wish I had better ideas for you.

                    1. 1

                      Thanks. I will try spinning up compton, just in case.

                      EDIT: Just started using urxvt recently, was using sakura. Haven’t noticed any difference yet.

                2. 2

                  I’ve often wondered, but for a while assumed we may have just covered up the problem with the inevitable march forward in hardware performance, if there would be any value in constructing a desktop operating system with hard guarantees on input/feedback latency.

                3. 3

                  It’s nice to see this issue getting more attention. I would love to switch fullly over to Linux on my old 12” Air, but the trackpad issues just made it unpleasant. I hope this effort garners more attention, and generates more opportunities to learn about touchpad drivers.

                  1. 1

                    I don’t know much about robotics but would it be possible to create a robot hand that can perform some fixed set of movements and provide reproducible results when testing this - against itself and against Mac touchpads? The resultant effect on a Mac can be measured (pointer distance moved, when down/up happened, how far scrolling went) - at least video recorded for comparison.