Fwd: Re: Application duplication (was: Re: cdbakeoven)

[Matthias Welwarsky]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ralf Nolden wrote:

> I think we need to "outsource" certain apps to their standalone versions
> not included directly in the core KDE distribution modules like kdebase
> and have one solution (preferrably the best and simplest to use for the
> user) that ships with the base modules.
> 
> But who cares what I think when one is totally blind because he is the
> author of a program that potentially could be moved to an extra section
> *sigh*.....
> 
> Ralf

I'm totally with you here, Ralf. I think KDE needs to concentrate on being 
the best _Desktop_Platform_ ever. Note "Platform". But that does not imply 
that we have to strip all applications from KDE and only keep some basic 
ones. These need to be there for reference, and I'm thinking about 
developers here. That's how KDE has worked in the past: The code of 
applications has taught many developers on how to solve programming issues, 
which is kind of how open source programming works. This means we have to 
have a large repository of code easily available for developers.

On the other hand: Maybe a platform without applications is not worth 
anything, a desktop without users isn't worth anything either. That is: 
Every issue a _user_ has with KDE is valid and to be taken serious. I 
strongly agree with Dimi here: Users are task oriented, they don't care 
about applications. UML-speak: Users care about "use cases".

So, we need to have both: lots of applications as a reference for the 
developers, and task oriented thinking to fulfill users needs. We're in 
danger to create a geek-only platform here.

How do we comply to both of these requirements? I think the solution could 
be something like the following:

Lets keep the modules base, network, multimedia a.s.o as they are, but 
modify the way to qualify applications for these core modules. Right now we 
look at an application and say, OK, it compiles, the code complies to 
certain standards, and to a certain extend it "does what it says on the 
cover", so it can go into the distribution. We should raise the level so 
that an application only goes into the standard distribution if it's not 
only "technically working", but also actually useful and not offers 
functionality that a user would expect in an already existing application.

Next: How do we keep the large code repository for developers' reference? 
kdenonbeta is a very good place, but right now, an application that goes 
into kdenonbeta is basically lost - it's very difficult to distribute 
software separately once it's in kdenonbeta, or in any other of the CVS 
modules. Though the solution to this problem is extremely simple: just 
offer a "make distribution" target to each application toplevel makefile so 
that a programmer can easily release source packages (and maybe binary 
packages, too) whenever he thinks it's needed. This has already been 
discussed on kde-edu and will be appreciated very much.

Maybe we should even go one step further and rethink what should be in a 
"standard" KDE desktop:
- - a web browser
- - a mail program
- - a media player
- - a picture viewer
- - a.s.o

Note that I didn't mention any application by name here. Maybe we do not 
need different "modules" called base, network, graphics, and so on, just 
one module that has _one_ application for every requirement a user may 
have.

regards,
        Matze


- -- 
Matthias Welwarsky
Fachschaft Informatik FH Darmstadt
Email: matze at stud.fbi.fh-darmstadt.de

"all software sucks equally, but some software is more equal"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8wTqvANO+fpRuZ2IRAsGJAJ9rEZZh/22DQlharstlKLYWljICbACfZzC9
QzQaT9sYN4bCgqTPUPN+gkA=
=zsmq
-----END PGP SIGNATURE-----