LWN.net Logo

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 2, 2008 17:46 UTC (Wed) by martinfick (subscriber, #4455)
Parent article: The critics are wrong: KDE 4 doesn't need a fork (ars technica)

I am a little amazed at this guy's focus on the desktop (the actual place where wallpaper and
such go traditionally).  I rarely use my desktop because that would mean closing windows or
using an out of band desktop icon to get to it; but that's just me.  Although, I notice that
many windows users always used maximized windows, it seems like accessing the desktop would be
a pain for them too? Perhaps with plasmoids I might fine the desktop useful eventually?  But
why all this fuss over the desktop?  KDE is about application integration with frameworks, not
the desktop.

I haven't used KDE 4.x, so beware, off the top of my head the plasmoid thing sounds cool, but
not the major improvement that I might hope for, such as perhaps taking plasmoid like ideas to
the application level where it would be more powerful.  But also a project centered desktop
would be a major win.  A complete flexible hierarchical plasmoidal project/window/session
manager, is what we really need!

A more project focused desktop/window/session manager would allow me to organize all the
applications/files for a project together into a project session.  I would be able to work
with several projects at once and depending on which project had focus the entire desktop
session would morph to the layout of that project!

A project session would be saved into one file (a project file) and would be start/stop-able
without having to login and out.  A project session would make use of already existing
paradigms.  It would be configured so that all the associated windows were grouped inside
another window (MDI?), or if preferred, the project would add itself to the list of desktop
windows (perhaps now labeled projects?) when it is active (opened).  Add plasmoids to the
picture and have a project specific plasmoid enabled desktop inside the project session
window/desktop.

Allow me to not only have session enabled projects that save their state continuously and
safely (the way Firefox does) for easy recovery after crashes, but also allow me to save the
state at any point in time as a separate project which I can then return to later if I want
to.  In the mean time I could continue to move forward with the project without fear of losing
the previous state!

And when I say session based, I mean fully session based, with the ability to save everything!
Save file statuses (and contents), window statuses, application statuses (especially terminal
statuses), connection statuses...  Allow me to email a project file home and open the project
session at home!

That is where the desktop paradigm should move, the desktop should at least be able to deal
with a project!  But that's not all, let's move on to windows and applications, another area
where a technological framework such as KDE (or even QT) could be providing better
integration.

Applications widgets could be built with a more flexible layout mechanism that is easily user
configurable.  I should be able to move around every pane or widget from any application to
anywhere.  I should be able to unhook an email message list pane from my email client and hook
it onto the desktop like a plasmoid.  But, I should also be able to simply hook a terminal
pane inside of any other application, perhaps my email client?

Why keep applications so isolated from each other at the GUI layer?  Depending on my current
workflow I should be able to configure one window with exactly the panes that I need laid out
the way I prefer, and if I need one button from one other application along with those panes,
I should also be able to create a new pane with a place to drop that button in my window.
This is the kind of integration that we should expect from our desktop frameworks!

Please don't take this as a criticism of KDE, it is not!  I am simply venting my frustrations
at what I see as a lack of completion of a lot of very good ideas that Linux/KDE (and other
desktops) have had. None of this is revolutionary, I am just carrying normal ideas that we use
every day to their logical conclusion...

-A patient and hopeful longtime Linux KDE user. :)


(Log in to post comments)

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 2, 2008 18:58 UTC (Wed) by Kamujin (guest, #52482) [Link]

I understand that sometime you need to be willing to take a fresh look at things if you want
to improve an already strong solution.

Sadly, I think KDE's main approach to taking a "fresh look" seems more like resurrected ideas
that were properly discarded as inferior many years ago.

Visually, I see progress, but for usability and work-flow. It is just terrible.

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 3, 2008 6:22 UTC (Thu) by Janne (guest, #40891) [Link]

the problem here is that you are basing your judgment of KDE4 on KDE4.0. KDE4.0 was the very
first release of KDE4. Since it contained massive amounts of new code, it has some rough edges
and not all the new functionality is up & running yet.

Or in other words: could you base your opinion of KDE3 on KDE3.0 alone?

"Visually, I see progress, but for usability and work-flow. It is just terrible"

While that might be true for 4.0, I can already see the groundwork for new features that will
bring massive boost to efficiency and workflow. And again: what you can see at the moment is
KDE4.0 and few betas of 4.1, hardly enough material to conclude that KDE4 sucks as far as
usability and workflow is concerned.

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 3, 2008 11:56 UTC (Thu) by aleXXX (subscriber, #2742) [Link]

Actually I think it might make more sense to compare KDE 4.0 with KDE 2.0 instead of 3.0. The
port 
from 2 to 3 wasn't that hard.

Alex


The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 2, 2008 20:37 UTC (Wed) by endecotp (guest, #36428) [Link]

I have entirely different ideas about how I want to radically change the way that my
interaction with my computer happens.

Our problem is that it's an enormous amount of work to even prototype novel interaction styles
(or whatever you want to call them).  It shouldn't be.  I'm not sure where we went wrong
though.

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 3, 2008 0:51 UTC (Thu) by flewellyn (subscriber, #5047) [Link]

That's a wonderful idea!  Code something!

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 3, 2008 8:29 UTC (Thu) by renox (guest, #23785) [Link]

>I rarely use my desktop because that would mean closing windows

And? In Windows, Win+D show the desktop (iconifying all the window) and I've configured KDE to
do the same (there's probably a native shortcut but I don't know it) if you don't like
keyboard shortcut you can have an icon in the taskbar to do the same thing, easy no?


As for the plasmoid, I don't understand how plasmoid are different from 'pinned windows':
windows that wouldn't be iconified in the 'close all' shortcut and kept in the background. Why
KDE needed to have something totally different instead of 'pinned windows'?


Also I agree that recoverable state should be something ubiquitous (system wide) but Firefox
is a poor example as it doesn't do it nearly as well as it should: if you close a window by
mistake you can't undo it (unless it was the only one opened).

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 3, 2008 10:25 UTC (Thu) by Duncan (guest, #6647) [Link]

> [On firefox: If] you close a window
> by mistake you can't undo it 
> (unless it was the only one opened).

Actually, yes you can.  There's an extension that does just that.  I don't 
run it as I don't need it (I prefer a real delete to trash, for the same 
reason), but I've come across it going thru the extensions list.

Duncan

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 3, 2008 11:30 UTC (Thu) by renox (guest, #23785) [Link]

Thanks for the information.
Unfortunately, this still doesn't make Firefox a good example for recoverability as this
feature should be available by default.

"Undo Close Tab" in Firefox

Posted Jul 3, 2008 11:52 UTC (Thu) by dash (subscriber, #6182) [Link]

As a matter of fact, Firefox is a good example of recoverability, as "Undo Close Tab" has been an available-by-default feature since v2.0. Right-click on the tab bar and it is one of the choices on the menu.

--
Dag Asheim
Linpro

"Undo Close Tab" in Firefox

Posted Jul 3, 2008 13:41 UTC (Thu) by renox (guest, #23785) [Link]

Yes, 'undo close tab' but not 'undo close window'(by default) so this isn't a good example to
follow as it is incomplete.

"Undo Close Tab" in Firefox

Posted Jul 3, 2008 19:39 UTC (Thu) by endecotp (guest, #36428) [Link]

> "Undo Close Tab" has been an available-by-default feature since v2.0.
> Right-click on the tab bar and it is one of the choices on the menu.

Not on mine.  Iceweasel V2.0.0.1.  (Ah!  Now I know the difference between Iceweasel and
Firefox!)

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 4, 2008 15:19 UTC (Fri) by pimlottc (guest, #44833) [Link]

Hear hear!  Project-oriented sessions are a feature I would desperately love to have, under
any desktop.  We're part of the way there - already there are a number of session-aware
programs that can save their state automatically on logout.

But instead of relying on every program under the sun subscribing to some predefined API,
perhaps we could take a page from the system hibernate folks.  Already there has been enormous
progress in the last few years, and kernel hackers have become quite astute at saving and
restoring system state with any manner of arbitrary programs running.

If we could take the next step and perfect a more fine grained method, then we could
selectively hibernate individual programs, allowing for session save and restore for any
program, regardless of its desktop without any effort from its developers.  Doubtless there
are a number of tricky issues to be dealt with, but I think it should be possible, and if
anyone is to get us there, our Linux hackers should be up to the task.

The critics are wrong: KDE 4 doesn't need a fork (ars technica)

Posted Jul 5, 2008 15:41 UTC (Sat) by aquasync (subscriber, #26654) [Link]

Yeah this would be a nice feature to have. DragonFly BSD's already got checkpointing, the
linux equivalent I'm aware of was CryoPID, but don't think its as stable.

Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds