Jump to content

KDE/KDE4/Goals

From KDE Community Wiki
< KDE | KDE4
Revision as of 19:39, 9 March 2016 by Alexmerry (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Proposed for Deletion

This page has been proposed for deletion for the following reason:

a list of ideas some people once had about what might go into kdelibs4 is not very useful

Please use the discussion section of this page to voice your opinion on this.


KDE 4 Goals

This page lists a mixture of ideas and suggestions for KDE 4 goals from the community as well as some details of actual changes based on progress thus far. Please do not assume that anything listed here is a promise for what will be in the KDE 4.0 release.


  • Qt 4.3 based
  • revised kdelibs API:
    • dump obsolete stuff
    • delayed BIC changes
    • Qt 4 API style changes
    • improve type safety by using the C++ keyword "explicit"
    • better separate core, network and UI classes (a la Qt4)
    • harmonise namespace and header utilisation in non-core libraries and KIO
    • determine how much of KNetwork can be folded back into QtNetwork (if at all)
    • kdelibs TODO list
    • kdelibs4 TODO Progess Chart
  • moving forward to self-explanatory application names
  • KDE3-support library (already started in trunk)
  • documentation helping to port KDE3 based applications
  • everyone introducing/rewriting a technology has to write documentation/tutorial about it
  • improved accessibility (all widgets have to implement QAccessibleInterface)
  • full support for meinproxy
  • new multimedia interface (Phonon, formely KDEMM) making KDE independent of one specific media framework
    • KDE/Qt style API for most multimedia tasks (everything a media player like Amarok needs should be supported)
    • better integration with existing solutions for multimedia/(pro-)audio on Linux
    • Phonon Backends
  • Suggestion: "Start menu" style desktop similar to Windows XP
    • Suggestion: try taking a look at e17 could use ideas from there. it's already very advanced compared to what we have these days. and i always use right click on my desktop to access the menu.
    • Suggestion: windows xp is outdated and I don't think many people like the start menu style -> make it innovative + integrate some kind of search into it too..
      • Right. SUSE 10.x has a nice search facility. I feel the KDE Menu should look like this.
        • Maybe have a look at this -> http://www.kbfx.org
          • But it would to Windows Vista-like Interface. Look for shotscreen of Vista RC1
        • It seems to be overcomplicated. There shouldn't be any obligatory elements that could be easily done elsewhere. Just the menu tree and (as traditionly done) log off/shutdown buttons.
    • Suggestion: the current Kmenu is almost there, but it would be nice to have in-menu applications search and some way to expand the menu by replacing the current menu folder with next one. To make the navigation easy we could put it the way it was in NeXTSTEP file manager: every opened folder (submenu) leaves an icon on top of the list, so clicking the icon gives a shortcut to the respective menu level.
    • Suggestion: make the KDE-Panel alternatively look like the suse-enterprise desktop panel under gnome. its neat, workflow oriented, integrated with beagle and who doesnt like it should be able to switch to classic view.
    • Suggestion: visit the start menu research talk at Akademy (info)
      • Preview of "Kickoff"
        • It seems to be overcomplicated. To pick up an app needed I need an extra click ("all apps"). The start menu should simply be the place where I can choose and run a random app. If I have some final set of favorites apps, I could configure an alternative menu or simple set of launch buttons elsewhere.
    • Suggestion: One thing the KDE users like is the easy configurability of the desktop. Try to make the start menu as adaptable to our taste as you can. We don't need another gnome clone desktop.
    • IMO development of the KDE4 "Start Menu" should start with Kickoff. After using it, the original KMenu, the Windows menus, KBFX and TastyMenu, I'm convinced it's the simplest yet most efficient option I've seen.
      • Still, I don't know if how the application menu was done is ideal. And probably want to avoid forcing people to install Beagle by default, unless it's going to be a major part of KDE4...
    • Suggustion: The start menu is an excuse for complexity, move in a simplicity direction similar to Mac OS X.
    • Suggestion: As the menu is currently a kind of tree, we could project the menu on a sphere - so as soon as you go to the next submenues, the sphere rotates to show the next level(s).
  • 2D menu and interface is outdated. We need something new (very questionable statement -- the full potential of well designed 2d desktop is not yet explored). Like 3D desktop. Like XGL-Compiz or Looking Glass. See Looking Glass project -> http://www.sun.com/software/looking_glass/demo.xml Also take a look at Beryl/Emerald for kwin (make kwin compatible?), fork off of Compiz -> http://www.beryl-project.org/
    • Criticism - 3D desktops like Looking Glass, while pretty are often inefficient. Because monitors display in 2D, 2D desktops would seem to be the easier and more productive than their 3D counterparts. There are many promising concepts from both Beryl/compiz and Looking Glass such as the desktop cube. However a fully 3D desktop environment would not be the way to go.
    • Strong criticism - there are many instances of desktops completely unsuitable for such eye-candy, in terms of work productivity and simple hardware limitations (especially laptops, with weak 3D acceleration and small screens, or even high-end laptops with CPU throttled down while on battery power), so the main goal should be an ergonomic 2D interface. Remember, one doesn't use a desktop enviroment for the sake of using it - it's a tool and interface for other things. Good DE/menu/whatever interface-related is the one you don't event notice during your work. There is place for optional gizmos like a 3D desktop, but they come after the serious, useful stuff.
    • BumpTop is a really great prototype of a 3d desktop, even if KDE4 most probably will not go into 3D space. Probably KDE5?
      • I like the idea of marking icons very much. It's a pain in the a** to mark files with the rubberband tool if you got a pile of files which in no sorting order line up easily for using the rubberband efficiently. I often wished I just could draw a freehand rubberband around the icons to mark - this would be cool to have in KDE. Clicking each icon with Ctrl held down is no alternative, and acrobatic combinations of Shift+Ctrl and multiple rubberbands are errorprone and even less intuitive. I've learned from my supporting issues at work that most office users even don't know about the Ctrl+Click feature. The idea of just drawing a freehand rubberband around icons is simply intuitive.
      • Critisicm: Have you got no other thing to eat up your performance? I think, this concept is very nice, but look at the movie. Does that seem more ordered?!
    • I find the eye-candy of Xmetisse very nice - it gives a nice argument against the Vista "Wow"-effect.
      • The possibility to copy pieces of a window to another location, use the windows even if minimized/rotated etc., is very good.
      • As soon as the desktop is 3D, there could be such eye-candy (demos!) as having some small physics-engine - the windows run around some center (kicker?), and when they collide, they shear each others menues off ...
      • I think it could be a major productivity boost if workgroups share a common desktop - but each has an own look unto this "universe". So I work on my spreadsheet, but if my co-worker asks me something I just scroll around to his window - better than VLC or FreeNX.
  • Suggestion: Do something revolutionary... make the GUI more like what Symphony OS (http://www.symphonyos.com/cms/) and Kuartet
    (http://kuartetdesktop.sourceforge.net/) use the entire screen realestate, the idea of nested start menus, and icons to launch/manage programs is old and not useful, most desktops are either mostly empty, or crowded with a mix of program icons and folders (i really like something like this:
    default1.jpg )
    and please, don't use anymore plastic but something new :)
    • Try something really new... and useful. This mockup is beautiful:
      http://www.kde-look.org/content/show.php?content=28476 - totally rework the general way that we think about our computers- in terms of what we aim to do, create, communicate, etc... putting email, im, etc, together makes lots of sense
    • Strongly oppose. KDE must work on computers whit 256 MB RAM+, so 3D effects must be optional (and there is problem whit non-free drivers if you want 3D). My idea is that you can switch between normal and kickoff/kbfx/gnome like/Non-cascading version.
  • revised or replaced configuration system
  • replaced IPC/RPC mechanism (D-Bus instead of DCOP)
  • Evaluate QCA as the cryptography backend for KSSL
  • Tenor (context search framework)
  • switch to kdom/khtml2/ksvg2 (kdenonbeta/ksvg2)
  • icon rendering
    • evaluate agg2 based icon rendering engine (no more pngs)
    • evaluate native Qt-SVG rendering (with Arthur) (on the fly)
    • evaluate cairo based rendering (on the fly)
    • compare on the fly rendering performance with different backends
  • RTL support and WYSIWYG in Kate/KOffice thanks to Scribe
  • more graphical effects/possibilities thanks to Arthur
  • new KDE control center design -- standalone control center needed?
  • freedesktop.org shared MIME database
  • rewritten Human Interface Guidelines
  • KDE Accessibility Guide
  • KDE Corporate Identity Guidelines
  • reconsideration of modules structure/which applications to include
  • better integration of python scripting - including build system
  • KJSEmbed integration in Kicker/KDesktop/KWin/KParts
  • Integrate Kicker, KDesktop and SuperKaramba into one streamlined application (Plasma)
  • Improve desktop/hardware interaction (Solid)
  • Native RPC bindings for KJSEmbed
  • Support for SSO (Single Sign On)
  • Switch to standard gettext
  • support for html background
  • Centeralized spellchecker, many non-english users expect to change spellchecker from only one part. Thanks to Scribe.
  • Command line options enhancements
  • Database Abstraction Layer
    • Evaluate QtSQL
    • Evaluate KexiDB
    • Evaluate knoda a full featured and working KDE database frontend with a rich set of command line options and python scripting.
    • Advanced data-aware widgets
    • Make support for relational databases data sources as easily available for every data-driven KDE app as KIO currently is; provide KControl module for database connectivity (a framework for storing and sharing db connection data is available, as well as generic widgets)
  • framework and property editor (already reused by Kugar And Kexi:
    koffice/lib/koproperty/ in SVN)
  • Shared Actions Framework
  • multi-engine service discovery/publishing system
    • LAN: support for zeroconf, SLPv2, Lisa and samba(?)
      • add samba and NFS in context menu
      • add NFS support
    • Internet: support for wide-area zeroconf
    • new procotol built on top of existing IM protocols (needs extending KIMProxy)
  • system-wide awareness of network connection state
  • OSX-style Dock support
  • Gnome Deskbar / OSX Spotlight style search applet
    • Possable Beagle integration
  • make use of Qt's Interview classes in Konqueror's file views. This could clean up and reduce the code and make the views more memory efficient.
    • separate the contents from the display (model/view concept)
    • consistently integrate metainfo and extended attributes in all views
    • make the views more consistent with the current task. Leave only relevant menu options, have the ability to change the menu (not just the toolbar) for each view.
  • Windows and MacOS X Ports of KDE parts
  • Evaluate Replacing, merging and/or removing some KDE applications
  • A backup resource shared between all KDE applications would be nice. Users keep asking about it. For example, they want to backup all their Kontact data. Or some app settings, or backup all settings and so on.
    • For this, have a look at dirvish: And also: Timemachine
    • and see rdiff-backup (gui less, but powerful as timemachine. Keep is a kde frontend to rdiff-backup)
    • Kamion is an application and library for backing up selected application data.
    • as part of the kde4 experience seems to be about going all sci-fi, i couldn't refrain from linking this article (arstechnica.com) here which draws a detailed picture how a file system like zfs (sun.com) could help any time machine like tool. and boy is it impressive.
  • Doc improvement: kio_help: share help from high-levels dialogs or KParts. I explain: for example kdelibs/KSpell2 config dialog is used by many apps but each app has to rewrite the doc for it. I would like a doc in kspell2 for the config widget and then that doc would be used by all apps using the dialog. Same for shared widgets in module/libs (like KOffice Autocorrection dialog is shared between KWord and KPresenter, i'd like to have 1 doc only for it) - Contact annma for more precisions!
  • Substantial Help/Manual/WhatsThis improvements. Make help more accessible to users, make it more easily discoverable, bring detailed usage hints closer to where they need to be applied (to the dialogs + widgets). Make "WhatsThis"-completeness a requirement for all KDE-core apps (definition in HIG). See here for some ideas:
  • add support for html backgrounds with clickable links
  • add support for selinux
  • better acl support
  • show filenames as links in the backtrace view of the KDE Crash Handler (jstaniek)
  • Easier Printing
    • Option to preview print output and remove/rearrange pages in this preview (just like Fineprint - Windows only, but test it, it's so useful)
    • Possibility to print multiple (also landscape) pages on one page (in the correct order!) (also with preview of course)
    • Are we using native printing, or are we all going to have to migrate to CUPS?
  • Improve Power management and laptop power support
  • One thing about the oxygen icons, i saw the last version of folders, but have you in plan to always keep them? i think is better that icons when the folder is empty and something in the folder (like a piece of paper maybe?) when there are files
  • Nepomuk-KDE as alternative, system wide, data-backend for KDE Applications. This can act as global store for applications including PIM, desktop indexing (including MP3/OGG file information), and ATOM/RSS reader storage.

This list is not assumed to be correct or complete and is subject to change any time.  :-)