Learn about Linux
Download Linux
Get Linux helpSpecial Offers
Feature: News
Four Core continues series of KDE 4 meetings
Share
Print
CommentsAll the KDE 4 meetings, such as the recent Multimedia Meeting, are intended to build team coherence and direction as much as to write code. All attendees seemed to feel that the meeting was successful in removing the hurdles that so often face a development team working together from places scattered throughout every corner of the world. Attendee Aaron Seigo says, "I get the feeling more than ever that KDE 4 really takes off here."
Fair weather in Trysil, Norway, throughout the week aided both spirits and efforts. Seigo writes, "One of the goals for KDE 4's interface is to be more organic, which is to say to behave more like the real world which we are well adapted to. In keeping with that we went for a walk out to a meadow for one of our two group meetings today." Work was done in a rustic Norwegian lodge selected to minimize distraction. Attendees enjoyed beautiful scenery and catered food. Nonetheless, World Cup fever prevailed midweek.
On Wednesday, a reporter for the Hamar Arbeiderblad, a local newspaper, visited the hackers. The paper was apparently responding to an alert made by a local KDE user, who also visited the group. The resulting article, ""Building the Future of Computing in Trysil," can be viewed by paid users of the paper's Web site.
Hacking done at the meeting revolved around outfitting KDE's low-level libraries for use in a new and modern desktop. Tasks involved in such a transition include implementing the cross-desktop standard communication tool DBUS to replace KDE 3's DCOP, porting Qt calls to use the new version 4.2, and, of course, bug-fixing. Attendees also decided to remove a few dozen obsolete classes from within the code of kdelibs. With the libraries more or less functional, the team must decide how to evolve them to suit the needs of KDE 4. Seigo writes, "We had a couple of quite long meetings today about kdelibs: what should go into kdelibs; how should we handle the process of managing the public APIs; what are our goals and practices for binary compatibility ... not trivial topics even if the questions are easily stated."
Though the meeting focused on the fundamentals, the team could not resist working on and discussing some higher-level code. KDE's Human Interface Guidelines and accessibility tools were topics of great discussion. KIO continues to be worked on, particularly its seeking feature, which was first written during the KDE Four Multimedia meeting. A redesign of the KJob class should also benefit KIO, which uses it extensively, as well as emerging technologies such as Akonadi and Solid. Akonadi is a forthcoming personal information management storage solution first proposed during January's KDE PIM meeting.
To aid the switch to DBUS, attendees planned a graphical browser that will work much the same way KDCOP does today, allowing users to see all the information and settings that applications offer to each other. LiveUI looks to replace the GUI development tool XML GUI. A new composition manager is to be built into the KWin window manager to replace the independent kompmgr.
As attendee Kevin Ottens says, "We've still so much to do, but the improvements made in the last few days are really motivating."
Comments
on Four Core continues series of KDE 4 meetingsNote: Comments are owned by the poster. We are not responsible for their content.
KDE only delivers source packages, and they are all made very modular. You don't have to install more then you want/need in any of them(except the natural dependencies).
If your packages do not fit your need, file a bug with your distribution. Or change to one with better built packages.
After the install finishes and reboots I do an apt-get update then apt-get install x-window-system-core kde-core kdm . What I end up with is a fairly clean KDE desktop. It has some un-needed stuff but then lacks some sorely needed core functionality such as mentioned earlier. The core disk mounting and unmounting app, and the hard disk (gas guage) Kwickdisk/kdiskfree. Both of which are included as part of a much bigger bloatware meta package. So too, if you want a sound mixer and volume control, you have to install the whole Kmultimedia meta package which includes Noatun and some other (at best marginally useful) apps I wouldn't wish on anyone.
I don't know who's in charge of figuring out what goes into kde-core in this case but it's severely wanting in one application (common sense) as to what is core, and what isn't.
Pray tell, how is it that I might unbundle the Konqueror web browser in your scenario of so-called "VERY" modular choices. Please don't argue why I might want to remove it, just address if I can and how. Really, this Desktop bloat situation is a real sore point.
It's not really desirable to 'remove Konqueror' in total, because Konq is just a shell that holds a lot of things - one of which is a web browser. I guess you still want Konqueror the file manager for those quick drag and drop moments, and you will still want KHTML installed for KMail's message viewer and Kopete's chat window.
One of the topics at KDE Four Core was how to split kdebase into the really important things needed at runtime (eg the useful KIO slaves like ssh and smtp) and the apps like Konqueror, so if I haven't managed to persuade you, KDE4 will let you save even those few MB.
HTH
Bille
> for those quick drag and drop moments, and you
> will still want KHTML installed for KMail's
> message viewer and Kopete's chat window.
Those are different things, no? KHTML is in kdelibs, which means when Konqueror (from kdebase) is uninstalled then KMail and Kopete should still work like before. Therefore, it should be no problem to remove Konqueror while the rest of KDE stays operable - there's a desktop that learned its lesson from other ones where the browser can't be uninstalled<nobr> <wbr></nobr>;)
Obviously the Konqueror file manager is a must keep item. However, Kmail and Kopete, as well as ALL other non (core) essential KDE add-on applications will need to pass the "best in it's funtion class" test. Cross platform apps have an advantage here as well.
Firefox's web browser wins by default here because as we prepare people for a move to a total GNU/Linux solution, we can minimize their upset and confusion by getting them used to using Firefox within their Windows environment. Same goes for the Thunderbird mail client or other of the cross platform email clients too.
Further, each of these and many other GPL'ed software applications are equally at home regardless of the choice of GNU/Linux desktops (KDE or Gnome). It's really just a matter of simplifying support by using apps that are universally available.
The winner of the GNU/Linux Desktop race is far from a sure thing. No one is a shoe-in. When you also consider that there remains many outstanding usabillity niggles and outright half finished core funtionallity on both of the Desktop front runners. The smart GNU/Linux early adopters are keeping as generic as possible where the extra apps are concerned.
Based on that explaination of purpose, Kmail? Not on a bet!! Anything but applications that are DE exclusive. Kpackage? No thank you, Synaptic will do just fine Thank You.
If KDE and Gnome would concentrate on their core responsibillities first, quit trying to scandalize would-be users comming from Windows with their mindless changes from the norm. "I got an Idea" (spacial nautalis type foolishness). Then screw around with the extras, as extras. They would be a great deal further along in the GNU/Linux on the Desktop acceptance game IMHO.
Linux isn't just about winning over Windows users. I've persuaded some people to dual-boot because KDE is shiny and has great extra applications (like KdeEdu), not a clone of Windows. If Firefox had just replicated IE, nobody would've used it. It's different and better so many switched.
I guess this is the most stupid comment I have read in a long time.
Since it is a meta package it is an almost empty thing, just depending on real packages.
Which obviously allows to install those packages one by one as needed.
One can even install the meta package first and then remove those applications that are unwanted
"One can even install the meta package first and then remove those applications that are unwanted."
Correction: One can "ONLY" install the Full bloatware meta package first and then remove those applications that are unwanted. There might in fact be some secret-sauce-incantation that might allow one to avoid this, but it certainly isn't, in any way obvious.
Since, achieving obviousness through irreducible complexity, is the holy grail of user friendly operating systems. Then, I guess some of you non-stupid people need to get back back to work because you are certainly missing your mark. At least that is, until you reach the point to where when you take a deep breath, your not feeling a tight warm and wettness around your heads, and are smelling your own crap.
If KDE is going to follow in Microsoft's footsteps. Then maybe they should continue to force feed everyone their extraneous add-ons, then magnanimously offer to allow the end user a configuration option that will remove the executables link icon.
$ apt-cache depends kmix
kmix
Depends: kdelibs4c2a
Depends: libasound2
Depends: libc6
Depends: libgcc1
Depends: libstdc++6
Suggests: khelpcenter
$ apt-cache rdepends kdemultimedia
kdemultimedia
Reverse Depends:
kde
In short, kmix does not require the kdemulimedia meta-package installed. Further, the only thing that requires the kdemulimedia meta-package is a complete kde installation (the kde package is a meta-package designed to give you all the official kde modules released with kde).
If you don't want bloat, make skip the meta-packages and just directly select your apps (e.g., kdm, kmix, kmail, or whatever). Also, remember that some dependencies can be met by multiple sources (such oss vs alsa enabled backends).
Like it or not, a web browser is now a part of a desktop environment.
If all else fails, try giving the users what they want, and how they want it.
And yet when another desktop bundles applications that's somehow a good thing? ROTFL.
Bille
KDE Is My Desktop Of Choice, BUT!!
Posted by: Anonymous Coward on July 12, 2006 01:28 AMIn it's current state, it's a bloated pig. Where many of KDE's apps (some of which suck BTW) are foisted upon the user. In some cases requiring the installation of whole huge meta-packages just to get one applet. For example: Kwickdisk/Kdiskfree should be available by themselves.
How about a clean Desktop with just the basic functionality, minus the Konqueror Web Browser. Then make available individual add-ons (applets).
I know, wishfull thinking. But, wasn't Microsoft hoisted by it's own petard for excessive bundling? Specifically their Internet Browser?
If all else fails, try giving the users what they want, and how they want it.
#