[KLUG Advocacy] What users want from Linux?

Adam Williams advocacy@kalamazoolinux.org
Mon, 03 Nov 2003 21:16:38 -0500


> > >>I've been called several times by trade rags and asked for
> > quotes;  as
> Instead of giving quotes maybe try writing articles yourself.  Many
> in here show a great command of the english language.

Ok, your the second person to make that suggestion...

> 4. Simplicity
> This is one that is really misnamed by them.  What is really needed
> here is some GOOD documentation not GEEK documentation.  

You won't get a moments argument on this point, at least not from me. 
One real problem is that it is HARD to get documentation out there even
when one writes it.  Alot of Open Sourcers still have a dismissive
attitude toward documentation and documentation authors.

> 5. Hardware
> This one is mostly a non-issue, escpecially for common hardware. 
> Sometimes video cards can be a bit of a challenge.  Speciallized
> hardware might be a bit of a mess too.  I have a logitech webcam that
> is NOT supported by Linux either by Logitech or by any Linux user. 
> Granted it is a cheap camera but it is frustrating to not be able to
> use it.

True;  but I also think this is a pretty easily avoided problem.  It
really applies only to hardware purchased pre-switch.  This same agony
applies to upgrading to XP;  we've discovered lots of hardware that has
been legacied on XP.

> Printers can be a big problem however.  First there is a problem with
> driver availability if you are not using a common and highly used
> printer like an HP.  Second, I have had problems getting what I would
> call acceptable photo output from either of my HP deskjet printers. 
> It works fine for basic color and B&W printing but photos are poor at
> best.  CUPS is making things better but I have reverted back to
> either my Windows desktop or my MAC when doing photo printouts.  

I've seen this on some models;  but I don't typically work with
el-cheapo printers so it may be much worse than my personal perception. 
If it don't do postscript you ain't pluging the piece of crap into my
network.

> Printers also need a common interface.  I think that CUPS is doing a
> good job of workign toward this.  The interface to the printer
> subsystem needs to be cleaned up though.  And it needs to be common
> between KDE, and GNOME.  I should be able, in any program, to select
> print and get a nice drop down that would list all printers I have
> available, and it needs to work in ALL applications!!!!  Printer
> options need to be GUI selectable, not selected from a command-line
> type switch.  

Ximian XD2 provides this, and it is just pretty universally supported in
GNOME 2.4+ apps;  some really low level parameters including dithering
methods, etc... can be adjusted via a universal print dialog in recent
GNOMEs (this can really help to tweak photo quality output).

I wouldn't hold your breath on a cross-desktop printing subsystems,  it
would be really hard and no one is going to get paid to do it. 

> Having a directory for Linux (can anyone say NDS) would be a huge
> help for printing in a large network environment.  I am sure that IPP
> would be a very good alternative although I have done very little
> with it so far, so I will grant that there may be a really good
> option that people are not currently using (I refer back to the need
> for GOOD documentation also). 

I'm currently hacking on a CUPS<->SLP<->LDAP.  It is at the point where
it twitches and gasps in agony.  CUPS supports SLP,  there are a few SLP
provides that support an LDAP backend, etc... so this is really just a
gnarly configuration issue

> I think that Oracle collaboration suite is a good example of ease of
> use (from what I saw at the demo). 

Have you looked at OpenGroupware at all?

> > >>>>6. Cohesiveness
> I think the deal of this is the varying directory structures making
> software install and systems administration difficult.  I think that
> more than anything this is the biggest problem between the different
> distros.  (of course there is RedHat with that atrocity they call
> bluecurve bleeech)

Personally I really think this is a packaging issue.  I shouldn't need
to know, 99% of the time, where the package drops stuff.  If I do it is
a failure of the installation defaults and scripts as well as the
configuration tools.

> > >>>>7. Applications
> > >>>>"Support for Windows-based applications would be a definite
> > plus," God,
> Part of the problem here is vertical market apps that are only
> written for windows.  These applications need to either be ported to
> Linux or comparable applications written.  This is happening to a
> point.  There are some verical market apps but many are still
> immature and more need to be written.  

True,  but i REALLY favor a port over a hosting solution.

> The other problem is one that over time will change.  There are some
> areas that are still lacking meaningful applications.  An example
> (although not a corporate example) is movie editing.  Linux is way
> behind in lower end editing software (most of what is used in
> Hollywood, IMHO, is custom written in-house). 

I don't really know enough about that niche to say much.

> > >>>>10. Consolidation
> > >"Instead of having basic tools spread around the operating system,
> > users would like to see them all in one easy-to-find location." 
> I think the best answer to this point is YAST.  A single point to
> find all the different configuration utilities is a very good thing. 
> One advantage that Windows has (to a point) is the control panel
> concept.  Webmin is another good example of a single point of
> configuration.  Both of these significantly increase administration
> ease of use.  

I'd prefer an mcs.exe like utility.  The ability to admin a remote
machine is critical IMHO - something that red-carpet, yast, whatever...
don't offer yet.  And configuration tools for reliablities sake need to
be more than command line wrappers which can go awry in so many ways. 
The movement toward XML based configuration files will hopefully ease
this transition.

> Increasing acceptance of Linux will be dependant in part on people
> like KLUG users.  We need to look at articles like this for two
> benefits.  First, what can we do to listen to concerns and address
> them with changes in the system.  Second, what tools are available
> that the other person does not know about (Webmin, YAST???)
> Self-rightous pontification will not forward the use of Linux.  There
> are things to learn from these people, even if they do not express it
> well.  

My beef with the article (and other like it) is that it is full of as
much error and misconception as it is fact or correct representations. 
Technical journalism should be of a much higher standard than this.  I'd
love to see an article critcal of Linux that enumerates solid and
substantiated complaints and problems (and their ain't no shortage of
them).  Freshmeat's occasional commentary articles tend to be good
examples of constructive criticism.