On Wednesday 02 February 2005 16:38, icebox wrote:
> When I install/show linux to a (linux) newbye he´s always confused on
> why there are 10+ window managers (most of them even can't make the
> difference between window manager/x/operating system anyway).
This is purely a matter of packaging/bundling/distribution.
I think it is unfair to choose a distribution with a diverse software
repository and then complain that there are too many similar options.
If a user wants a pre-selection there are Linux distributors for that as well,
for example Xandros or Linspire.
Generic distributions are for people who want to tailor their system to their
own liking.
> I remember I spent some almost a week on the net after I returned from a
> 2 week hollyday in the mountains to find out why, suddenly x.org
> appeared why xfree is no longer good to use and which is better to
> choose.
xfree is still good to use, but the distributors felt more safe using the
x.org option when xfree.org changed their licence.
For the user it doesn't matter, they usually install the distributions X
server without looking for alternatives or even caring about them.
> IMHO the word join should be used much more often then fork.
Agreed, but that is already happing on a lot of fronts.
For example LSB works on establishing installation compatability criteria to
share between distributions, freedesktop.org works on unifying data and
technology between desktop implementations.
> I think that having to choose not only between gtk and qt but than
> between a commercial qt and a community one would only add to the
> confusion.
Not sure I understand this.
The choice between GTK+ and Qt is a technical one, so it depends on the
technical needs of the developer(s). That's not unique to Linux that happens
all the time when evaluating third party libraries for software development.
The choice between Qt commercial and Qt free is separated from the one above,
it is legal instead of technical.
It is more likely a decision in management than in the development department.
Very likely the availability of Qt free is a non-issue for companies, they are
usually used to third party libraries licenced commercial and might not even
be aware of free licenced versions.
> It's hard enough the companies are to choose between 2 or more widget
> sets (not forget wxwin and others) then package the application for 10+
> distributions, each distributions needing more than one package because
> they break compatibility almost at each version (remember most
> companies cannot offer source code).
True, but as I wrote above LSB2 is about to change this.
> Just as a bonus there are 4
> printing systems, 2 major kernel versions in use across distros, 2+
> sound systems (with more on the way probably) and so on and so forth.
I doubt that the kernel version matters to most applications, unless they do
something very low level but in this case they have the same "problem" on all
other platforms as well.
As for the print system, this is mainly a problem because applications decided
to implement system dependent things themselves.
Applications that use generic printing APIs like Java applications using the
javax.print print service or the KDE printing API or even just having an
option to configure a print command the postscript output will be handed to,
don't have to worry about which spooler manages the jobs, which rasterizer
creates the pixel information and which printer technology puts it on paper.
> IMO that´s one of the main reason why we see so few commercial apps
> around. A commercial developer is not eager to invest more time and
> money in packaging an application for 10 distros than he did creating
> that particular app.
Absolutely true, I always wondered why nobody came up with a packaging service
for software vendors. Sounds like a good opportunity for making some money :)
Observing that Windows applications often state that they run on Win98, 2000
and XP, I imagine that either there is some difference there as well, or
software vendors just like to put a lot of different names in their
compatability lists so the ignore any common base.
Might be a marketing problem, maybe marketing should pay for the extra
development effort they are generating with this ;-)
> apart and work together. It's much useful to improve an already
> existing app than reinventing it again... e.g. I see projects started
> for creating the yet another CD burning software every month or so.
Sometimes yet-another-project seems like wasted resouces, but often the new
project is a lot better because they know about the mistakes made by the
older project and can therefore advance further than the original project
could have.
Sometimes it is out of s shift in user expectations. My favorite example is
text editors.
There are two text editors doing everything you want a text editor to do, vi
and Emacs, but I really thank Christoph Cullmann for creating Kate and I
guess a lot of other users do as well.
And I think it is not a new phenomenon on Linux or for free software, it
happens all the time on other platforms as well, only there a new project
can't usually use internal knowledge of another project thus being even more
questionable.
> Why? We have k3b, we have xcdroast and more, why just not choose one
> and start improving it?
The task might be the same, the goals might not.
One of K3B's goals is good KDE integration, I doubt that would be true for
xcdroast.
My point is that availability of alternatives is not a problem on Linux
because it exists every where else as well.
The only difference is that it gets more attention, being more visible due to
the fact that a lot of users use generic distributions with huge package
archives instead of minimum distributions and painfully searchiing for
software themselves.
Cheers,
Kevin
--
Kevin Krammer <kevin.krammer@xxxxxx>
Qt/KDE Developer, Debian User
www.mrunix.de - German Unix/Linux programming forum
www.qtforum.org - Qt programming forum
___________________________________________________
.
Account management: https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.
