Message: previous - next
Month: February 2012

Re: [trinity-devel] KControl and the confusing mess it has become.

From: Darrell Anderson <humanreadable@...>
Date: Sun, 5 Feb 2012 21:57:49 -0800 (PST)
> I use KDE 4.8 on one of my laptops and use it nearly the
> same way as I
> use Trinity.
> Could you just name such a "latest desktop fashion" which
> is:
> -newer than KDE 3.5
> -visible from the user (e.g. not the fact that KMail2 uses
> Akonadi), and
> -which use is *required* in the KDE4 desktop interface

Sure, I can do that. :)

Odd that you should exclude KMail and Akonadi, which pretty much pins the tail on the donkey.

I have tried KDE4. I have KDE4 installed in virtual machines. I am not naive about KDE4. I have a lengthy writeup at my web site, written a while ago with 4.5.5, based on my experience as a non KDE4 user learning to use KDE4. Before people rush to nit pick everything "wrong" and "short-sighted" with my journal entries, I can say with hindsight that I have learned a few things since that writeup and KDE4 has continued to mature --- but the writeup is legitimate for that moment in time based upon my experience at that time.

Yes, I dislike the way various configuration options function, such as customizing the panel, folder versus desktop views, etc. Yes, there is a method that, once learned, more or less works. More or less. That does not mean because I learned the KDE4 way of doing things that I like what I see.

Yet learning the KDE4 way of doing things does not excuse the way Nepomuk, Strigi, and Akonadi are now being embedded into the system, which are examples of the "latest desktop fashion" that is required to use KDE4. These three services reminds me much of the way Internet Explorer is embedded into Windows starting with Service Pack 2 or 3 of Windows 2000. Previous to that Internet Explorer could be removed with impunity and I did so regularly. Not so thereafter.

The basic Nepomuk and Strigi services can be disabled from running, but they can't be removed completely because developers use libraries from those packages. That is, developers are building their apps with those dependencies. I tried removing them. Things start breaking when those packages are not installed. To me that is poor design.

I see places where semantic desktop services like Nepomuk could provide benefits, such as enterprise environments or virtual offices. I am curious to read how others might be using that backend to coordinate work flows --- although I have yet to read detailed examples how people are sharing in that manner. To me the semantic desktop remains a concept that has yet to prove its viability. Possibly in five years developers and users might have this figured out. Yet I recognize that such services provide no benefit to many home or single workstation users --- so why should they be forced to install such packages?

Akonadi is a "latest desktop fashion" example that cannot be disabled. Starting any PIM app starts the Akonadi service. Although I see some users benefiting from this caching service, I also see that many never will and they are stuck. For example, I am not an email glutton. A few emails a day, even when I'm contracting off site. I don't run KOrganizer and KAlarm is sufficient for some basic reminders. I use Akregator but subscribe only to a few feeds. I'm not an information junkie. I also do not have or need a huge address book. I'm not that kind of user. I don't need (or want) that kind of information overload. I don't need to waste RAM running a cache like that because there is nothing to cache.

At release 4.8, I am still reading complaints from people using KMail. That is disheartening. Eight point releases and how many minor releases in between and a key PIM app is broken? Because I use KMail, I am not migrating to KDE4 when I know I am likely to experience breakage.

I could choose not to use KDE4 PIM apps, but I'm not interested in migrating to GTK PIM apps. Why bother with an integrated desktop environment when I am going to start mixing apps and widget libraries? I might as well use a basic window manager and have a full menagerie of apps.

If the developers had let the user decide whether or not to install those three backend services, I likely would be using KDE4 right now. I have been around computers a long time, longer than many people participating in this list have been alive. I recognized that the early push --- not by the KDE4 developers but by distro maintainers --- to include the early KDE4 releases as standard fare rather than in a testing branch, would be short-lived and eventually KDE4 would mature. I blame the distro maintainers with their unrealistic and insane bleeding edge mentality for those early perceptions about KDE4. Because of my anticipations toward that eventual maturity, some time after the 4.2.x series was introduced I considered moving to KDE4 --- eventually --- until I read that removing Akonadi was not an option.

My own testing showed that to be true. I could not start simple apps like KAlarm or Akregator without starting the Akonadi backend service.

When I discovered that removing the Strigi package started to break things, I frowned some more.

Recently I read an online forum discussion about a user trying to build a non mainstream KDE4 app that, with the latest release, required Nepomuk to be installed. A deeper frown.

KDE4 will not run on older hardware. An idle, stripped-to-the-bones KDE4 desktop (along with the underlying operating system) requires more than 512 MB of RAM. That more or less excludes people on fixed incomes with older computers and people in developing regions of the world from using KDE4. From my perspective KDE4 has become a geek's playground and geeks almost always have bleeding edge hardware. Understandably so because time is money. The faster development proceeds the faster the results. Better hardware improves that kind of environment. But that kind of environment is not what non developers use, which is where software should always be tested. Fellow geeks using bleeding edge hardware and sharing similar opinions seldom provide good feedback for usability testing.

And so the story goes.

Redesign Akonadi, Nepomuk, and Strigi as true optional packages for users and I will reconsider my perceptions about KDE4. Until then I will devote my energy toward building Trinity.

Please notice I have shared my experiences and opinions. I have not written that KDE4 is crapware. I have written that I disagree with some fundamental KDE4 design decisions. Decisions that, for now at least, seem irrevocable.

Nor have I called anybody names. I disagree with certain design decisions but have not called anybody stupid or insulted anybody's mother.

This list is for Trinity developers. That means opinions and enthusiasm here will lean heavily toward that desktop environment and not others. Why Trinity exists is no secret and we here should not back off from those reasons. Trinity exists because those involved do not like the design decisions made by the KDE4 developers, such as those I just shared.

Yes, during discussions we should pause to check our facts before making statements, but often in human interaction the line between fact and opinion is forgotten --- especially when those involved in a tight community share similar opinions for being involved. Human nature 101. Nothing new about this behavior trend any more than a person during a game rooting for a visiting team can expect some razing from the home team crowd. Humans are social creatures and seek those who share similar perspectives and views. In those discussions we should remind ourselves that this is the Trinity developers list and we should expect that kind of human interaction. We should not apologize for why we are here or even feel compelled to do so.

Every time somebody here in this list uses the symbols K, D, E, and 4 concurrently and in that order in a comparative manner to Trinity does not mean the person is a bad person or is misinformed. Such statements should be received and viewed from the simple premise that this is the Trinity developers list, not the KDE4 developers list. People are allowed their opinions. Such statements are not always derogatory although in my opinion, KDE4 developers deserve some criticism. Criticism yes, but not hatred. And no, I'm not going to join the KDE4 discussion lists to help change KDE4. End users have been trying to persuade the KDE4 developers to reconsider their design decisions and they have refused. I'm interested in improving Trinity, not KDE4.

As long as Nepomuk, Strigi, and Akonadi are required I am not interested in trying KDE4, 4.8, 4.9 or 4.whatever. I'm not interested in trying GNOME (version 2 or 3), Unity, LXDE, or Xfce, on desktop computers (other devices such as tablets is a different conversation). For desktops I am interested in Trinity.

I don't join KDE4 discussions with the intent of pimping Trinity any more than I would join an Ubuntu discussion list to pimp Slackware. Joining the Trinity discussion list means being expectant that participants are going to be favorable toward Trinity more than other desktops. To join and act otherwise is acting naively.

The world is big enough for different desktop environments and different operating systems. Users and proponents of each should focus on their preferences and leave others to do likewise.

I am not going to hold my breath or pretend to walk on glass every time somebody here shares an opinion favoring Trinity over KDE4. That is exactly why we're all here.

Calvin, I responded to your request for suggestions toward improving KControl. Let me know when you need help with testing.