Computer
Computer Diary
    RepRap


    "Our greatest fear is not that we are inadequate, but that we are powerful beyond measure."

    Marianne Williamson
    from her book "A Return To Love"

    Here my little rant and praise place, where the daily experiences of my programming work are expressed. I publish them with the idea that others might find it useful and benefit from it.

    2013/04/27
    KDE / Kubuntu 12.04: 10+ years terrible GUI, A Systemic Problem of OSS
    last edited 2013/12/14 01:56 (*)

    After 2-3 years not looking at KDE & Kubuntu (Ubuntu with KDE) I installed Kubuntu 12.04 - the installation I did via the Windows Installer was terrible:
    • Linux kernel failed to boot, I had to add 'nolapic' as boot argument (facepalm)
    • outdated packages, Firefox and Chrome failed to install, terrible experience. I upgraded Kubuntu to the latest, and was able to install Firefox and Chrome (not just Chromium).

    KDE started up, and first I wanted to change the keyboard layout, not use the swiss layout (CH) but US layout - I went to "System Settings" and I found this:

    System Settings


    Kubuntu 12.04 System Settings

    Where can I configure the Keyboard?

    You have a minute . . . you won't find it, there is keyboard icon but it's for shortcuts - there is no icon for keyboard configuration, there is no mention of keyboard at all.

    This is the state of Open Source Desktop KDE software in the year 2013!

    Instead to rumble silently, I took the some time to look for the G+ KDE Community, and posted this , and some informed me to use the internal search engine of KDE to find the keyboard configurator, Jos Poortvliet, from Open SUSE recognized the systemic problem, but his explanation made the entire issue even worse:

    (excerpt of the discussion)

    Jos: System Settings is a bloody mess, we all know it.

    Rene: If it's a mess you recognize (!!) why not fix it? (I would if I could) Can you honestly tell me why it's not fixed? Is it too complicate, or is there no consensus among the developers that it is a mess or nobody cares or dares to fix it?

    Jos: It is not fixed because it is HARD to fix. I know it seems easy but it's not, both for technical and practical reasons. ... the System Settings modules are NOT under anybody's control. They are maintained by the maintainers of the applications or subsystems they are associated with. Changing them would require the collaboration of (very rough guess) over 50 people from all around the world, many of which are not even active anymore, who live all over the world, do this in their free time, have different ideas about usability and have different priorities. It is HARD.

    Rene: I am fully aware this all seems like details to you, but from a user point of view, this is a major annoyance, and reason not to use KDE, as rediculous it looks like, and I can tell you, when people review KDE for schools or even business (I once made that call for a business solution using KDE), at the end the mess on the "Systems Settings" is reason not to use KDE. And I ask you to read this carefully: it's not disregarded because the icons are wrong, but because it's a mess, and if you can't get even "Systems Settings" coherent, how about the real challenges of the desktop coherently resolved? You communicate an immature state of software with it, even it might not even be true (most of you would argue the backend is great). 10 years is just too much to have this not fixed . . . (that perhaps is the terrible truth)

    If you cannot resolve the "System Settings" mess, something as simple as that, how can I expect that something more complicate as "System Settings" to be resolved?

    Jos: As I said, it is a hard issue. So saying "if you cannot resolve something as simple as the System Settings mess you can't resolve other things" is wrong: it is NOT simple. It just SEEMS simple to an outsider, that is what I tried to explain.

    Rene: Simple vs Hard: let me explain, of course I believe you when you say it's hard, you explained it, but this should not be hard, but simple. Something here has gone wrong, for 10 years, something as "simple" as a "System Settings" should be simple to fix, but it's obviously not - why not? Because of systemic software decision which allows 50 people able to contribute but at the same time stalling the fixing of a mess - this is the message behind our discussion.

    Jos: The way KDE is set up is the reason why it has been successful for 17 years now - it might have unfortunate side effects but I can make a good case that if we had set it up different, we wouldn't even HAVE a System Settings in the first place as KDE wouldn't have made it to 2013.

    So, while it is annoying, the way we work inherrently leads to the System Settings mess (and more like that) AND it is what enables us to be successful. Can't have your cake and eat it too. We simply have to put in the extra effort to fix this, it is as simple as that, and we will.

    Rene: I think you are missing something very important in your perception, and I like to elaborate on this more. I have been studying collaborative groups for quite some time, a group can be formed such that the diverse views aid to a final conclusion (this process might look trivial but is not, consensus that is). You have 50 contributing developers stalling a fix which affects ten or hundred thousands of user, or even more? This is a systemic dictatorship you have formed, 50 contributing developers stall a fix, and you tell me "AND it is what enables us to be successful", no, this limits to be really successful, to be adopted by a BILLION people. You allow a small group to dictate, based on sole technological system decision (how modules expose their configuration) the actual outreach of the entire KDE system.

    Jos: KDE can't be successful without contributors. We have a culture of bottom-up guided decision making and killing that culture kills KDE. That culture is what led to System Settings being what it is - changing that has to be done in a way that fits with our culture. You can disagree with that all you want but that won't change reality

    Rene: You confirm my worst view of KDE - that is systemic technological mess (obviously), but also culturally by those who contribute. I have no problem as long you KDE developers don't think you actually develop a viable alternative to Win7 or MacOS-X this way. When the mess is part of the cultural identity, so be it - I don't like the mess, I am wasting time and attention, and when it can't be resolved then it makes no sense for me to use it.

    (end of excerpt of discussion)

    You see, if KDE would be 1-2 years old, one would ignore it and hope someone gifted does create a coherent System Settings, but after 10 years something inherently not resolved means something at the entire decision process within KDE has gone wrong and that goes beyond my random rant but exposes something to really look at for Open Source Software for Linux.

    Systemic Lockdown in Open Source Software

    It is widely accepted that MacOS-X has done a tremendous job to simplify system settings, hide the complexity of it and expose it to average users. KDE, the Open Source Desktop software attempt, who merely copies Windows and a bit of MacOS-X, has tried now for over 17 years - and I have been using it apprx. 10 years ago for 2-3 years, and then left it, because no useability improvement has happened, even worse, nowadays you can't even find the Keyboard configuration.

    The brief excerpt of the discussion reveals a few significant issues:

    • there is a technical limitation
    • people who once contributed no longer be reachable, a common effort to resolve it impossible, they actually stall a resolvement
    • the mess is part of the culture to code KDE

    Now:

    1. is something to address - let's assume it's not just technical incompetence, but connect directly with the 2) - if it were just technical incompetence there would be no hope for KDE whatsoever.
    2. people who are no longer reachable should be replaced, as simple as that; package no longer maintained should be purged, period.
    3. this part is the worst, KDE developers seem not to care the useability, give no or little attention of it - the clichee of Open Source at its finest.

    As said, if KDE would be a startup effort, nobody would care, but over 10 years on-going mess, reveals the systemic problem of the KDE development in particular, illustrative to alike projects.

    Crowd-coding or Open Source coding does often give great results, Linux on the server is such a story, Linux runs on every Android phone, on millions of servers serving the web-pages we read, the SQL databases of Wikipedia are open-source - it's everywhere, but there are also the example like KDE, where the crowd is unable to bring something useable like a Desktop system to birth, it remains a hardly useable piece of software, even after 17 years. They are happy is has survived just that long, but life just seeing as survival isn't life really.

    To step back a little bit, one realizes the collective intelligence can in cases be quite dumb, as I see it in case of the KDE development, or it can be actually dangerous, like the mob who lynches a culprit; on the other hand, the collective can be mature and act responsible, like at the fall of the Berlin wall in 1989 between West- and East-Germany.

    The collective intelligence has to have a space and rules to express itself, I know this may sound very bizarre for some to read, but I learned this in Occupy movement in 2011 and also before with my experience running SpiritWeb.org (~1994-2003) with over 100,000 registered members. In case of Occupy it was the General Assembly , and a clear separation of power:

    • the facilitator/moderator of the General Assembly (GA) seem to have the power, but was just facilitator: giving the GA a structure, summarize and point out where things need still decision and elaboration
    • the actual power remained in the crowd, the GA was the form to provide the space to express the power of the crowd

    The GA has a clear structure, everyone can speak, when given the space to speak (one at a time); the facilitator summarizes and the collective acknowledges this, if the facilitator summarizes wrong, it can be voices until the spoken is heared and understood correctly.

    Going back to KDE, aside of the technical issue they have to resolve themselves (or even not) it cannot be, that absent people who left the project stall the resolvement of a mess - this is, as I said above, a dictatorship, a passive one due the absense. This is a major flaw in the community awareness and understanding of the KDE development community.

    To sum up: a crowd itself cannot express its collective intelligence by the sole gathering, it requires a setup and a structure to allow the parts or individuals of the collective to voice their perception, views, and opinions, and have an instance who reflects and summarizes what has been voiced, bringing it together - without, all opinions are still there, but the parts are not aware of it, and no reflection of such can happen - no self-awareness of the collective is possible. If people who no longer participate an effort are given still influence, this particular undertake stalls, which KDE illustrates well.

    One has to pay very close attention to the details, as I experienced in the GA at Occupy Zurich myself, the facilitator defined the tone, the atmosphere which allowed even shy people to voice their point of view; this was an immense win for me to see the responsibility of the facilitator and the service he or she was doing, it wasn't an issue of power anymore or pushing a personal agenda, but help the collective to find its way, to become aware of itself and realize its power and intelligence.

    I don't know the KDE community well enough to say that they lack a figure like Linus Torvalds, who operates as that facilitator for Linux - although his people skills are nearly non-existing but his technical skills are excellent - so his technical expertise and focus what Linux should do goes not havoc, whereas in case KDE, there seems no such figure, neither direction . . . headless.

    PS: Jos has joined the efforts to create slim version of KDE: KLyDE addressing some of the issues, but it has been one of many attempts according his statement in the past 5-7 years to resolve the known problems, none of the past succeeded.

    Update 2013/08/21: A few months ago I ranted about the quality of Linux Desktop approach of KDE, which after 10 years could not deliver a decent desktop. I recently came across ElementaryOS - an Ubuntu-based desktop with slick and intuitive GUI - something I have been waiting for more than 10 years for Linux, finally something to hope for.



    All posts or individual posts:

  • MetaFS - Dealing With Metadata the Proper Way (2013/12/14 01:00)
  • My Cellphones & Smartphone (2010-2013) (2013/09/02 23:15)
  • KDE / Kubuntu 12.04: 10+ years terrible GUI, A Systemic Problem of OSS (2013/04/27 12:18)
  • UNIX Man on Windows 7: VirtualBox + Ubuntu + LXC (2012/11/27 19:24)
  • Metadata - The Unresolved Mess (2012/07/10 20:59)
  • Cellphone Networks: Thieves, Insanity & Crap (2010/01/26 13:57)
  • MacOS-X for a UNIX Man with a PC (2009/09/26 20:43)
  • Windows XP for a UNIX Man (2009/09/22 18:28)
  • Server Counting (2009/05/18 11:50)
  • Automatically Geotag Photos without GPS (2009/04/22 08:28)
  • Rebirth of FastCGI (2009/04/15 17:17)
  • Online Advertisement & Income for Web-Site Owners (2009/03/18 22:10)
  • iPhone JavaScript Frameworks (aka Avoiding Objective-C) (2009/03/14 22:08)
  • Google - The Almighty Tracker & Advertising Blocking (2009/03/12 22:09)
  • How To Save 300MB RAM (2009/03/07 22:07)
  • Verbosity of Programming Languages (2009/03/06 22:06)
  • Problems with MacOSX (2009/03/03 22:03)
  • MacOSX: My First Steps (2009/02/24 09:57)
  • Catch 22 with HDD/DVD Recorder Medion Life (2009/02/24 09:27)
  • Kubuntu 8.1 as guest on VirtualBox MacOSX host (2009/02/24 01:33)
  • VirtualBox vs VMWare Fusion on MacOSX (2009/02/24 01:19)
  • SQL vs GREP with 230K lines (12MB) GeoLite (2009/02/23 20:10)
  • Kubuntu 8.1: Eye-Candy & Memory Waste (2009/01/24 09:57)
  • Firefox 2.0.x / 3.0.x - Memory Waste (2009/01/22 19:34)


    [ post new entry ] (only for administators)

    Title:

    Text:

    Tags: (separated by commas)

    Date (optional):

    Password:
  •  


    .:.




    Copyright 2007-2016 © by René K. Müller <spiritdude@gmail.com> / Google+
    Illustrations and graphics made with Inkscape, GIMP and Tgif