frans.englich at telia.com
Sun Aug 1 00:02:01 BST 2004
For long buzzed about is KAdmin, similar to KControl but
does only contain server administration functionality(which KControl have in
addition to user preferences), and has KParts as content instead of
KAdmin would have kernel configuration, boot manager, "pseudo modules"
linking to web admin interfaces, kuser, KIOSK(no, that's nothing you need as
a regular user) -- all integrated under one roof. KAdmin would typically be
installed on servers only.
Why is KAdmin needed and what problems does it solve?
* Administration software is more than simple configuration options -- it's a
working area. Some understandable needs toolbars, menus and so forth, and
have tried implementing it in KCModules(QWidget) and it doesn't work -- it's
like squeezing an elephant through a key hole. Administration functionality
needs the flexibility KParts offers; a widget in a dialog is insufficient.
* It is hard to design a KControl when it has such a wide role and demands on
flexibility and expandability. When administration functionality is taken
care of elsewhere, its scope is narrow, and designing the content(categories
etc.) and representation(the GUI) gets easier. KControl can become a simple
application for user preferences.
* As noted above, administration stuff lives between KCModule and standalone
apps -- the former to small, the latter usually overkill. KAdmin would reduce
the number of entries in the KMenu, group related together and bring
consistency -- like KControl. A unified way of handling root permissions, and
In short, having this diversity in KControl restrains everyone, and by
splitting the content in two and ensuring these two do One Thing Well,
everyone will be happier. Especially KControl needs it as a step towards
avoiding feature bloat.
This is no new idea, of course. There's SuSE's YaST, and kde-debian have stuff
too -- important is that "KAdmin"(whatever that will be) is part of KDE.
Currently my plan is to write this umbrella application KAdmin for KDE 4.
Input on its specification is of course welcomed -- but let's see if we get
I've probably missed tons of stuff already, so comments, Corrections,
questions and updates on what already is going on would be great.
More information about the kde-core-devel