Introduction
As some of you may know I have recently started a new blog called "My Ubuntu Blog" (www.myubuntublog.com).
Some of you may therefore be alarmed by the title of this post as leaving Ubuntu would be a strange decision to make having committed to a whole blog on the subject.
This article is actually a guest post from Paul Smith who left a well thought out and well written comment at the bottom of the article "Is Unity Bashing a hobby?".
Having read the comment I made the decision that it was too good to languish at the bottom of the post and so I asked for Paul's permission to publish his comment as a full article on this site, which is about Linux in general.
So without further ado here is Paul Smith's article "Why I Left Ubuntu".
Why I left Ubuntu
I was a great fan of Ubuntu and Canonical. I loved the pre-Unity versions of Ubuntu. I found the last Gnome 2 version to be especially functional and polished.
When Canonical switched to Unity on 11.04, I tried it and mostly liked it. Admittedly, there were some issues but I really liked the fact that Unity did a better job of maximizing the screen real-estate available to applications than any other desktop environment I have used previously. I was hopeful that the wrinkles in Unity would be worked out in the next version and was just about ready to pay for support from Canonical for all the systems in my home, mostly as a thank you, when Ubuntu 11.10 came out.
Ubuntu 11.10 seemed to be a lot buggier overall. Unity would do weird things to my applications and sometimes make the desktop unusable, forcing me to drop down to the shell to restart X. Pulse audio on this version was a dog and would simply not work with a sound card I’d been using successfully on Linux for about 5 years. I also discovered a number of newly introduced library compatibility issues that broke some of the commercial software I needed for my job.
The final straws for me was Canonical's decision not to include snd-pcm-oss as a kernel model (which I discovered with Ubuntu 12.04), breaking ALSA's OSS emulation, as well as the inclusion of Amazon search.
I now use Scientific Linux with the Trinity desktop since I really liked KDE 3. I find that I can easily get everything to work with that distribution and it is extremely stable. At this point the only thing I miss is the old Synaptic package manager and some features of the Debian package file format.
When Canonical switched to Unity on 11.04, I tried it and mostly liked it. Admittedly, there were some issues but I really liked the fact that Unity did a better job of maximizing the screen real-estate available to applications than any other desktop environment I have used previously. I was hopeful that the wrinkles in Unity would be worked out in the next version and was just about ready to pay for support from Canonical for all the systems in my home, mostly as a thank you, when Ubuntu 11.10 came out.
Ubuntu 11.10 seemed to be a lot buggier overall. Unity would do weird things to my applications and sometimes make the desktop unusable, forcing me to drop down to the shell to restart X. Pulse audio on this version was a dog and would simply not work with a sound card I’d been using successfully on Linux for about 5 years. I also discovered a number of newly introduced library compatibility issues that broke some of the commercial software I needed for my job.
The final straws for me was Canonical's decision not to include snd-pcm-oss as a kernel model (which I discovered with Ubuntu 12.04), breaking ALSA's OSS emulation, as well as the inclusion of Amazon search.
I now use Scientific Linux with the Trinity desktop since I really liked KDE 3. I find that I can easily get everything to work with that distribution and it is extremely stable. At this point the only thing I miss is the old Synaptic package manager and some features of the Debian package file format.
Why Mir concerns me
I do a significant amount of technical computing. I could care less if the same OS runs on both my desktop and my phone or tablet. I need a desktop that provides a good environment for code development, modeling, as well as a limited amount of CAD. I do this work on machines at work and, to a lesser extent, on my home systems. Until my phone or tablet can support a large amount of DRAM, many cores, and can plug into a keyboard and several large monitors, I don’t see myself migrating away from a desktop. Canonical’s direction appears to be to water down the desktop experience in order to make it more like the phone, the same bad mistake Microsoft made with Windows 8. Developing Mir is a result of this direction. I expect to use my desktop and phone for very different tasks and could care less if they use the same OS. As an user, Mir does not appear to offer me anything of real value.
Given the direction Canonical is taking, I am very concerned that NVIDIA and/or AMD will make X and Wayland second class citizens in favor of Mir. I would love to use Nouveau and similar open source drivers; however, they’re not functional enough yet, either for the software I use for my job or for recreational use with the games. Some of these games purchased from Loki Games dating back to the late 1990′s.
I am thrilled that Steam and other game developers are beginning to fully embrace Linux and would like to spend some money on these games. Assuming these games are coded to work exclusively with Mir, then buying these Linux games is not an option unless other distributions such as Scientific Linux, Fedora, or Debian also migrate to Mir. For legacy games such as the ones produced by Loki Games, I am concerned that Mir may not emulate X well enough. Full support for legacy applications that depend on X will be more of an issue if the Linux community's effort to develop X emulation is split between Wayland and Mir.
Given the direction of the rest of the Linux ecosystem to standardize on Wayland as well as the concern over OpenGL support from NVIDIA and AMD, I really wish Canonical would have worked with the Wayland team to reach their goals rather than going their own direction. In my opinion, trying to make the same OS and applications work on both big iron and small phones or tablets is just silly at this point. Given this, fragmenting the Linux ecosystem right now to save a little power on low end devices is just plain stupid. I understand Canonical's argument for the other issues, such as, the desire for a more extensible input system; however, Canonical should have been able to work through these issues with the Wayland team.
Given the direction Canonical is taking, I am very concerned that NVIDIA and/or AMD will make X and Wayland second class citizens in favor of Mir. I would love to use Nouveau and similar open source drivers; however, they’re not functional enough yet, either for the software I use for my job or for recreational use with the games. Some of these games purchased from Loki Games dating back to the late 1990′s.
I am thrilled that Steam and other game developers are beginning to fully embrace Linux and would like to spend some money on these games. Assuming these games are coded to work exclusively with Mir, then buying these Linux games is not an option unless other distributions such as Scientific Linux, Fedora, or Debian also migrate to Mir. For legacy games such as the ones produced by Loki Games, I am concerned that Mir may not emulate X well enough. Full support for legacy applications that depend on X will be more of an issue if the Linux community's effort to develop X emulation is split between Wayland and Mir.
Given the direction of the rest of the Linux ecosystem to standardize on Wayland as well as the concern over OpenGL support from NVIDIA and AMD, I really wish Canonical would have worked with the Wayland team to reach their goals rather than going their own direction. In my opinion, trying to make the same OS and applications work on both big iron and small phones or tablets is just silly at this point. Given this, fragmenting the Linux ecosystem right now to save a little power on low end devices is just plain stupid. I understand Canonical's argument for the other issues, such as, the desire for a more extensible input system; however, Canonical should have been able to work through these issues with the Wayland team.
About the author
Paul Smith is an electrical engineer with 23 years of post college experience. He wrote his first program on an MOS Technology KIM 1 in 1979. Paul has been using Linux since 1998.
Summary
I would like to thank Paul for allowing me to post this article and I hope you enjoyed it as much as I did.
If you think that you have an article worth posting on this site please feel free to get in touch at everydaylinuxuser@gmail.com.
0 comments:
Post a Comment