Ubuntu & Linspire

[Scott Mazur]

On Fri, 09 Feb 2007 12:26:07 +1100, Daniel Pittman wrote
> Joe Hart <j.hart at orange.nl> writes:
> > Jonathan Jesse wrote:
> >
> >> Wow, lots of broad statements there.... I always thought the beauty
> >> of Ubuntu was that all the command line stuff, the "hardcore", is
> >> still there and one can use it if he/she wants to or if one is closer
> >> to a "newbie" then he/she can use the graphical options to
> >> change/update/configure their system.
> >
> > Well, yes and no.  Take a look at this:
> >
> > https://wiki.ubuntu.com/KernelCustomBuild
> >
> > They explain (sort of) how to do it right after they say it will void
> > any contract you have for support.  All "hardcore" users know that
> > compiling your own kernel is the best way to tune your system to your
> > own hardware.
> 
> All "hardcore" users are damned fools, apparently.  This is, not 
> least, because there is no statistically significant measurable difference
> between a generic kernel and one that you "tune ... to your own
> hardware."
> 
> > Why do I need floppy support, parallel port, serial port, bluetooth,
> > pcmcia, etc. if my computer doesn't have, or use them?
> 
> Do you /really/ miss the 77MB of disk space they consume?
> 
> Do you actually know that is the one and only resource they consume?
> 
> > Why is there a 'restricted modules' in the first place?
> 
> Licensing reasons.  "Hardcore" users, of course, don't care for that 
> law and disregard it.  Companies lack that luxury.
> 
> > Sudo/Root?
> 
> Good security practice?  Is it that hard for a "hardcore" user to run
> 'sudo passwd root' -- I mean, seriously.
> 
> These are the "hardcore!"  They know how to use the command line,
>  and it isn't like Ubuntu prevents you setting a root password -- or 
> even blinks if you do.

I agree (or have no opinion) about everything you've said up to this point. 
It's true Kubuntu doesn't prevent you from setting a root password (I've done
so myself, becuase that's just the kind of user I am).  But it's not true
there are no consequences to this.  Every (and I mean EVERY) configurable
option in KDE that needs admin rights prompts for a password.  Out of the box
that's fine (whether you agree it should be any old user password or root
only).  But once you set a root password none of the KDE password prompts
work.  Regardless of the password you type in (root or user) it's wrong and
does not authenticate.  So by setting a root password you are forced to login
as root to make admin changes for ever more.  And it's damned annoying being
prompted for a password in KDE when you know darned well it's not going to
work.  It shouldn't have to be that way.  Everyone should set a root password
just to understand how mucked this action makes your system before commenting
on how 'trivial sudo is'.  That by and far is my biggest grudge against
Kubuntu, and yes weighted against the things I like about Kubuntu, so far
things balance out.


> > Wacom devices in xorg.conf?
> 
> I guess "hardcore" users don't own Wacom tablets, but they do own USB
> mice, right?
> 
> I infer this because you whine about Wacom tablets being configured 
> to work "out of the box" but we don't hear complaints that xorg.conf 
> contains definitions for USB mice...

You don't hear complaints about definitions for USB mice because they don't
generate warnings about missing devices everytime you start an application in
X.  When they do, advice is given to fix the config, not 'ignore the error'. 
I want to be clear about something: Developers spend time making code work. 
They don't spend time making writing (let alone testing) 'exception' events to
ensure they've cleaned up properly.  What you see as a harmless X error that
means nothing and should just as well be sent to the NULL bucket, I see as a
hole in the code (in this case X, which is a big part of the system to have a
hole in).  Maybe the guy who wrote the code to check for missing devices (like
Wacom tablets) cleaned up when the code hits the error.  Or maybe he felt his
job was done by dumping the error message in the first place and thinking that
any idiot would correct the Xconfig before trying to continue.  Maybe he
generated the error message in the first place because he knew he didn't have
time to properly ignore it.  You tell a user to ignore it and some time later
program XYZ fails because X was left in an unknown state after a failed Wacom
device check.  Only now you've got no idea why program XYZ failed and so
begins the 'your code is broken' finger pointing game.  It's a fact. 
Developers don't spend time handling errors.  Ignoring the messages that was
given is just plain bad advice.  Encouraging it is irresponsible.

Scott

-- 
Registered Linux user #395249, http://counter.li.org
Nothing goes to waste when Little Fish are near!
(http://www.littlefish.ca)