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

List:       kde-core-devel
Subject:    political issues (Re: The filedialog)
From:       Stefan Westerfeld <stefan () space ! twc ! de>
Date:       1999-11-04 20:30:52
[Download RAW message or body]

   Hi!

Disclaimer: I am sorry for bringing up political issues in a technical
discussion.  If you don't want to care about these things, just don't
read it. If you have read it and feel that discussing this is a bad idea,
don't reply.

I also have no opinion technically to the filedialog thingy, but something
to the discussion.

One position - lets call it the commercial-pragmatism - repeatedly agues:

Political issues are not at all important for the KDE project, and should
in no way affect the development. Only focus on the technical decisions,
and ignore all other issues.

However, I assume this position doesn't reflect the whole story. As an
example for that, just look how the Gnome project convinced people to
work on it. They did it by discussing the KDE licensing. So without
taking any position in that statement, one can say that our political
decisions also influence the future of the KDE project, and are - somehow -
important.

So to Prestons rules:

On Thu, Nov 04, 1999 at 10:09:25AM -0500, pbrown@redhat.com wrote:
> Get it done NOW! 
> Focus! 
> Use available tools rather than reinventing existing ones! 
> When making a suggestion, change "we should.." to "I will.."; grandiose
>    plans are useless unless you are willing to put in the work. 
> Improve iteratively. 
> Start with reasonable functionality and configurability and then improve
>    over time. 

These are only the technical rules mentioned here. So if the KDE project
wants to do decisions, they can't be the only guideline for deciding.
There could be something like:

- Prefer solutions that are accepted and/or widely used in the free linux
     world.
- Try to write widely reusable code where possible (no QPL).
- Do open development.

Qt is a toolkit. It's a good toolkit. But also, it includes things that
make development consistent with Qt, such as the Qt data types, the Qt
signals and slots model (with moc), the Qt marshalling, the Qt XML stuff
and so on.

While these make programming easy, they also do two things politically:
- they limit the reusability of the code to a Qt only world (much KDE
  code can be seen as written in Q++ which is mostly useless without Qt,
  as opposed to C++)
- if they are maintained inside the Qt CVS tree, they conflict a little
  with the goal to do open development (as most people for instance will
  not improve QFileDialog, except for those working at Troll)

I mean: there are a lot of neat things Troll Tech could add to Qt. Such
as a COM/CORBA/ActiveX replacement, to do consistent component development
with Qt (also worthy for the windows customers). Or a real network library
replacing kio. Or real HTML capabilities, on which Konqueror could base.
Or real text capabilities, from kword. Or integrating DOM. (And of course:
if a HTML-with-DOM implementation is already in Qt, using another thing
in KDE - like our HTML code - will add bloat - so the bloat argument is
in favour of using Qt solutions as soon as they are there).

But if the KDE project would follow all these, I think we would be deciding
wrong in the political section, moving from a free desktop to a set of Qt
example apps (which just wrap a bit of functionality).

As it stands, the KDE project IMHO can't fail to remain a great solution
for desktops ... the only way to achieve that would be compromising the
political goals, that at least for me are the motivation for working on
KDE.

Also, some arguments go like:
- free software developers just can't produce the quality one can reach
  in commercial entities like Troll Tech or Microsoft
- commercial software development is due to it's strength of maintainance
  inherently superior to free software development
- so in the long run, we should move away at least the fundaments of KDE
  from free software developers to companies

No I don't agree here. If you do hacking just to do hacking (as Matthias
once put it) - that means just for fun, like someone watches TV - then
you are probably satisfied with prototyping something, to see it re-
implemented a year later in a library written by a company, and your
(more open in the sense of development model) solution dropped.

But if you do hacking to achieve something that fits into your view how
software development & licensing should be done, you are probably not
happy with that.

   Cu... Stefan

PS: The "Trolls" happen to be really nice people - and without them KDE
would be not where it is today ;).
-- 
  -* Stefan Westerfeld, stefan@space.twc.de (PGP!), Hamburg/Germany
     KDE Developer, project infos at http://space.twc.de/~stefan/kde *-

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