[Panel-devel] The ALI: do we really need or want it?

Björn Balazs B at lazs.de
Tue Jan 10 14:49:19 CET 2006


Am Montag, 09. Januar 2006 23:19 schrieb Nicholas Kaye-Smith:
> The goal of a user is to:
> Handle data in a productive way.
> or Make it possible to handle data in a more productive way.
> generally speaking (where the word productive means the eventual
> satisfaction of the user).

That is true - and leads me to one conclusion: Users do not primarily want to 
use applications. They just have to use apps - because they are the framework 
in which a user can productivly handle content.

To give an example: 
Almost nobody will walk up to a computer and say - well I really want to start 
amarok, see what I can do with it.

It works the other way around: I want to listen to musik - how can I do that - 
ah: I have to start amarok and then I can select the music I want to listen 
to.

There has to be a translation done by the user: Listen to music means starting 
amarok. The user has to learn KDEish or Linuxish - namely knowing what the 
applications do.

So I would strongly speak for allowing users to fullfill their tasks - e.g. 
listening to music - speak the users language!

Therefore: For each peace of content there should be a changable default 
action (based on the users preferences and a well chosen system-default) and 
the possibility to easily specify different actions - not talking about the 
apps in the first place, but talking about the goals in the first place 
(Edit, play, e-Mail,...) - and then saying which app can be used for it.

I know: the KDE-Menu does support categorization - or is trying to speak the 
users tongue - but only at a very basic level at the moment. There is one - 
at some distros two - levels of categroization. So users only have to learn 
the categorization of the menu, which is only little better then having to 
learn Linuxish. 

But this has to be done in a more flexible way: Applications often fit into 
more then one category, fullfilling more then one task. Also the 
categorization is pretty rough at the moment. I have up to 20 Apps in one 
category, making visual scan more difficult then needed.

Therefore: For any approach on an ALI - which IMHO is definitly needed as a 
part of a greater concept - a categorization should be combined with a set of 
filters. This way it is quickly possible to narrow down the amount of 
resulting apps - having contextual information available for the resulting 
set of apps 

We are developing this kind of application-selection for the next release 
version (0.8) of the gentoo installer "kuroo", which is supposed to give you 
access to all apps in the portage tree (some 10000 at the moment) - 
http://www.kde-apps.org/content/show.php?content=16002 - please contact the 
main developer karim for getting access to the svn-version if you are 
interested in the new concept.

Some screens of the kuroo8 can be found here:
http://www.ergopool.de/kde/kuroo8_basicView.jpg
http://www.ergopool.de/kde/kuroo8_InstalledKDE.jpg
http://www.ergopool.de/kde/kuroo8_InstalledKDE_filterMail.jpg

Cheers,
Björn


More information about the Panel-devel mailing list