[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: KDE Linux/BSD distribution
From:       Christoph Wiesen <chris () deadhand ! com>
Date:       2004-10-26 21:26:47
Message-ID: 200410262326.47526.chris () deadhand ! com
[Download RAW message or body]

Am Tuesday, 26. October 2004 22:53 schrieb Gary L. Greene Jr.:
> While this is a tit-or-tat, aspect, I really have to disagree. Debian is a
> good distro, but it's the FREE distro that aims more at the geek, instead
> of the general audience which KDE is aimed at. Additionally, (and I know
> that this is opinion) I prefer apt with rpm over apt with dpkg, notably
> since dpkg is harder for me to make packages for.
>
...
> > Sure I would like to say that Ark Linux or something similar would be the
> > way to go, but reinventing the wheel over and over again can be a very
> > overwhelming task.
>
> As Bernhard asked: where is the reinvention of the wheel?

Well it has already happened (the reinvention I mean, since ark is a full 
distro already and doesn't have to do any reinventing anymore), so in this 
case it's not a 'problem' at all anymore ;)

My point is more like this:
Users coming from windows mostly expect to have a selection of third party 
software available in a format they can use easily. While debian is not the 
best example (SuSE and Red Hat are) there is support from non-Debian people. 
That means that you can find more .deb files any user can install on their 
debian (compatible) system on kde-apps.org (just an example) than you can for 
ark.
That's not a problem for some people but a huge issue for others. Maybe ark 
can use other packages - i don't know, but usually distributions that are 
made from scratch (more or less) need packages that are build specifically 
for them. There are just so-many packaging systems a usual upstream developer 
or contributing packager will support - Debian is among them.

Seriously this can be discussed to no end - there are areas where ark excels 
and there are areas where debian takes lead, period.

I agree that debian (sid, testing, stable) are systems more or less targeted 
at the geek, but debian is a very nice foundation for user-oriented 
distributions as well - they just happen to get all the benefit of a huge 
community driven package pool.
Now I'm sounding like a bad debian advertisement, but I guess I mean it: Look 
at SimplyMEPIS (lacks polish unfortunately) and Ubunu both are great desktops 
and build upon the strength of debian. I don't think they (especially Ubuntu) 
had to make much compromises due to their debian nature.


>
> > I don't think ark is bad by any means (though it has
> > it's controversies like the thing they do (did?) with root because of the
> > Windows-like nature) just that it's much-more work than what would be
> > needed to build THE KDE Desktop.
>
> The changes can be backed out by editing the pam.d files and setting
> passwords. this is trivial for a developer IMO. As the security head of
> Ark, I can say that this is not a security hole. I've run a number of tests
> against the system, and found zero comprimises caused by setting a system
> up like this.
>
It's a very subjective point and can't really be discussed I guess - I just 
'feel' a user should know when running applications as root and be forced to 
think about it by entering a password.

> > I'd say with enough people (not that much at all I think) and enthusiasm
> > the Kalyxo team could produce a really great Debian KDE Desktop that
> > installs from CD very quick (Live CD or Debian-Installer).
>
> You've never worked on a distribution have you? it takes a LOT of work and
> help from a large number of hands. Unless you like the idea of 50+ hours a
> week spent on getting it working while having limited help.
>
No I have never, and i'm sure my comments show this, but I still think it's a 
different thing to build up a whole new distribution (like you did, which is 
a great thing - can't stress that enough ;)) and to build upon what's already 
there (a finished distribution called debian).

And that's why a KDE-centered 'Distribution' should not be a whole 
distribution at all in my opinion - too much can go wrong that prevents a 
timely and bugfree release when it's due. 

With debian if the first goals are simplified enough there would not be much 
more (ok, there would, but I'm simplifying here) than a nicely build KDE and 
patched Kernel needed to get things going for a start. None will have to 
think about  packaging samba and get the systems security straight because 
that's already taking care off.


What I suggest is a system that integrates completely into Debian itself, not 
a fork like Xandros or Lindows, so no work has to be done twice except the 
one you want to solve differently.

Considering the original intention of the thread starter (as far as I 
understood it) building a whole distribution around it would be overkill.

Maybe ark will get there, but I think right now you could not guarantee to 
have a complete and stable installation CD out the time KDE 3.3 hits and then 
again when KDE 3.4 hits.
This is speculation on my part of course but I don't have the impression that 
a KDE centric release cicle is what you aim at. Ubuntu does - for GNOME (new 
GNOME -> new Ubuntu release) - and that's what KDE could need as well... I 
can't see anything else but Debian doing this right now.

Cheers,
Chris
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]