Fwd: Installation Problems and Useability

Matthias Kalle Dalheimer kalle at klaralvdalens-datakonsult.se
Fri Aug 23 22:54:11 BST 2002

----------  Forwarded Message  ----------

Subject: Installation Problems and Useability
Date: Fri, 23 Aug 2002 23:30:05 +0300
From: Eli Wapniarski <eliwap at newmail.net>
To: kalle at kde.org

I originally tried sending this email to webmaster at usability.kde.org but the
email came back.

I have become so frustrated with installation of KDE. To sum up, I do not
believe that the correct approach to making a desktop easy to use is to have
a completely new Linux distribution to get easy to install installation of
any desktop, or to have to recompile KDE entirely from scratch to get bug
fixes is not an easy thing to do either. Besides, compiling KDE from scratch
is not quick, or easy.

I think the main focus of useability needs to be on installation.

I am using Redhat 7.2 and KDE

I have been a Linux user for about 4 or 5 years now....
I have been using linux primarily because I saw it then and I see it now as
something that can very potentially be the future. However, over the past
year I started feeling comfortable around Linux to recognize the difference
between what is a lack of knowledge on my part, weaknesses in the operating
system, and some problem out there (Internet particularly).

This particular year has been a source of nothing but frustration. Under the
KDE 1.x series I was able to download the binaries, run a script, the script
would install KDE and presto, it would work. The RPMs for KDE 2.x worked
quite well. I had to compile from sources KDE 3.0 and it worked except for
KDE Bindings. I was told after submitting a bug report that I would have to
recompile the whole thing again from CVS or wait until the next version. I
elected to wait for the next version. I was unsuccessful at compiling 3.0.1
or 3.0.2. Nothing critical here. I am running kde 3.0 just fine.
Except for a few minor crashes from konqueror. Things have working relatively

Now.. there is a security problem with konqueror which has been fixed in
3.0.3. I tried to download the RPMs from your sight for Redhat 7.3 when I ran
rpm -Uvh --test there were all sorts of dependancy issues. That's OK I
 running Redhat 7.2 and Redhat compiled the kde libraries with funny names.
 but the one that really got me was that kdebase-3.0.3-7 required
 kdelibs-3.0.3-10 or higher which of course does not exist. I realize that
 this is a Redhat problem, but...

I am not a programmer, but I do know the difference between statically linked
libraries and dynamically linked libraries.

Having an installation as big as KDE and having the libraries dynamically
linked is a real pain in the butt and is creating a major mess for you guys
and for me. I think it would be a good idea to get your libraries statically
linked. Why?... Because it would make the release of bug fixes much easier.
You wouldn't have to ask individuals to download an entire version or compile
an entire CVS library just to get a small bug fix in one or two library
files. It would go a long way into getting an installer like up2date and
red-carpet available for KDE and that would make the installation of
applications a whole heck of a lot easier as well. Because, ease of use
depends on the ability to install KDE  and if a user can't install it, well,
every other useability issue is irrelevant.

I realize that this would make KDE downloads significantly bigger, however, I
think that the trade off on ease of installation is worth it.

I realize that you guys are not getting paid for the trully fantastic work
 you are putting into KDE. However, I do not believe that I am going to
 compile another version of anything this big again. My experience with KDE
 3.0 is enough. I simply do not have the time to do it. In my opinion
 currently, Gnome 1.4 was better than 2.x And this isn't saying much. I
 really do not like Gnome. And if... I can't get something were I get the
 security fixes installed quickly and easily for my desktop applications
 well. It looks like the dark side has already won. And the point of Linux
 as a security issue also becomes irrelevant.

Thank you for your consideration.


Eli Wapniarski

Matthias Kalle Dalheimer
President & CEO/VD
Klarälvdalens Datakonsult AB

More information about the kde-core-devel mailing list