GNOME Shell 2.29 Brings A Lot Of Improvements

[val-gaav]

Then try it and find out by yourself.
As long as you don't explain the real purpose of "plasma" and what is it for, it's worth nothing.
What does it bring for me as a user that I won't be able to do with Gnome ?

From developer point of view if they used already stable plasma architecture with Gnomeshell then they would be a lot faster to provide a working stable version of it ...

Plasma is flexible and allows to construct different desktop shells with it. The three shells already constructed on it are the main kde desktop with it's panels,plasmoid widgets etc, plasma-netbook shell (which is a special use case shell for netbooks) and something called plasma mobile for N900 devices.

More here : http://en.wikipedia.org/wiki/Plasma_%28KDE%29

[BlackStar]

I guess this goes for personal preferences or how much you want ... If Qt4 integration is 90% for you then it seems for me it's 100%. On the other hand gtk+ is really a LOT worse. I even dare to say it doesn't even try to do the job at all .

Agreed on both points: a single Qt frontend is usually good enough and GTK doesn't really try.

As for different frontend for each platform I disagree. Firefox is one example how that doesn't really work. It has quirks and problems on Gnome were just as you say it should have 100% integration. Oh and how long does it take for FF to integrate with new Win7 ?

Heh, Firefox is the poster child of failed platform integration. There are much better examples: Chromium (yes, it breaks one specific paradigm (tabs on window decoration), but otherwise integrates perfectly on the respective OSes), Opera, Photoshop and MS Office (ooh, sacrilege!)

On the other hand, apps like Qt Creator, Amarok (Qt), MonoDevelop (GTK), Gimp (GTK) don't really feel native outside of their respective platforms.

Anyway I applaud the Qt4 way as it takes the burden of integration from application developers and does a decent job (I would say it is an almost perfect job, but I did try it only on Win, Xfce, KDE so maybe MacOSX has some problems just as you say).

Me too. Personally, I think Qt is the best cross-platform UI toolkit right now. My only objection is the (relative) lack of language bindings (nope, I'm definitely not going to use C++ for a GUI ) - but that's another story entirely.

[BlackStar]

From developer point of view if they used already stable plasma architecture with Gnomeshell then they would be a lot faster to provide a working stable version of it ...

Plasma is flexible and allows to construct different desktop shells with it. The three shells already constructed on it are the main kde desktop with it's panels,plasmoid widgets etc, plasma-netbook shell (which is a special use case shell for netbooks) and something called plasma mobile for N900 devices.

More here : http://en.wikipedia.org/wiki/Plasma_%28KDE%29

Great sentiment, but does Plasma work independently of the KDE infrastructure? If not, it's probably not the correct tool for this job.

[spykes]

From developer point of view if they used already stable plasma architecture with Gnomeshell then they would be a lot faster to provide a working stable version of it ...

This is a pointless assumption... In the same style, I could say the best way to make Linux ready for the desktop is to use Mac OSX instead.

[kraftman]

Then try it and find out by yourself.

It seems I will have to do this :>

As long as you don't explain the real purpose of "plasma" and what is it for, it's worth nothing.
What does it bring for me as a user that I won't be able to do with Gnome ?

It wasn't my intention to celebrating KDE's plasma over Gnome's shell here. After reading some Gnome's shell descriptions it reminds me plasma, so I wondered what are the differences.

[val-gaav]

Great sentiment, but does Plasma work independently of the KDE infrastructure? If not, it's probably not the correct tool for this job.

I remember the lead plasma developer mention that plasma can be easily striped down to have only a Qt4 dependency (so no kdelibs etc etc.)

I wouldn't call Qt4 a dependency on kde infrastructure.

Personally I think as much as possible " under the hood " technology should be shared among Gnome and KDE, and both sides should work on that.

[BlackStar]

I remember the lead plasma developer mention that plasma can be easily striped down to have only a Qt4 dependency (so no kdelibs etc etc.)

I wouldn't call Qt4 a dependency on kde infrastructure.

Personally I think as much as possible " under the hood " technology should be shared among Gnome and KDE, and both sides should work on that.

To the contrary, I feel it's better to develop common standards (freedesktop) and let each DE decide how to implement them on its own. Diversity is the driving force and the main strength of our community. Take that away and what are we left with?

Put in another way, if Gnome n+1 or KDE n+1 suddenly bombs, common standards mean that you can jump ship and keep on working. If everything was consolidated into a single, "super-DE", jumping ship would be much much more difficult. Not unlike the situation with Windows right now (yes, you could fork the super-DE, but the lack of standards would ensure your effort would be almost certainly behind the curve.)

Edit: interesting tidbit on Plasma, will have to hunt down some developer blogs.

[drag]

yeah, except that you just posted a big pile of bullshit.
Redhat always was a KDE hating gnome-house. From the start. It did not matter what KDE did - or how much gnome sucked.

Yes. Redhat hated KDE personally because KDE picked on them when they were children.

This is the type of crap that I was talking about.

Redhat spent _YEARS_ trying to figure out how to integrate KDE and Gnome into a usable desktop. Bluecurve and all that.

[quote[]
Novell has the Ximian gang onboard. Gnomefanboys who are well known for their open KDE hatred.
SUSE was VERY successfull being KDE based. [/quote]

Yes it was so successful it never became commercially viable and was purchased by Novell.

Novell came and the Ximian-gangsters tried everything to kill off KDE. Well done, assholes!

And ubuntu? If you really want 'easy' - compiz+gnome is NOT the way to go. But hey, there is a reason why ubuntu patches the crap out of gnome.

Uh-huh.

See: No rational thought.

You think that Redhat has a personal grudge against KDE? That Novell gives a flying crap either way? They are both created with the primary purpose to make money (and Redhat has a hard-on for open source); beyond that they could give a shit less. Your making yourself sound f-ing nuts, man.

[drag]

I'm truly fed up to read always the same reactions of Gnome haters/KDE fanboys, each time Gnome related news are posted.
You really have no other argument that a list of KDE techno when trying to convince people of KDE superiority ?.

Yeah.

If people want to talk about KDE then talk about how much KDE kicks ass and what aspects of KDE are the best in your opinion.

It is not necessary to denegrate Gnome to make KDE successful. It's not a zero-sum game (a zero-sum game is were one side must lose to have the other win) or a competition. KDE can exist and be successful quite outside of any involvement from Gnome or anything else.

It's perfectly possible to run QT application in Gnome and visa versa and when people follow FreeDesktop.org standards in their applications then the integration is not really that bad.

Beleive me... It was not always that way. For a long time there were real incompatibilities to were you had to choose one or the other and there were always a few window managers that would break KDE applications and running KDE applications in Gnome would do nasty things like break your sound and cause KDE applications to be used when the user prefered some Gnome ones.

If you look at the people that actually put time and effort behind Gnome or KDE they do it because they want to or are paid too. There is very little hatred between the developer groups... they have differences in opinions, have different approaches, and are competitive, but they don't hate one another.

[drag]

Agreed on both points: a single Qt frontend is usually good enough and GTK doesn't really try.



Heh, Firefox is the poster child of failed platform integration. There are much better examples: Chromium (yes, it breaks one specific paradigm (tabs on window decoration), but otherwise integrates perfectly on the respective OSes), Opera, Photoshop and MS Office (ooh, sacrilege!)

On the other hand, apps like Qt Creator, Amarok (Qt), MonoDevelop (GTK), Gimp (GTK) don't really feel native outside of their respective platforms.

Having identical 'Look-n-Feel' is really really really overrated in my honest opinion.

If taken just by itself Gnome has _far_ more unified look-n-feel compared to everything else out there, except for KDE.

If you just take Gnome apps and Gnome-related applications... Epiphany instead of Firefox, Abiword and Gnucash instead of OpenOffice.org, etc etc etc. Then it is _Very_ uniformed.

Yet it's pretty obvious that in many cases Windows and OS X have better usability in practice. Wether it's because people are more familar with it or the UI designers are more talented or whatever... it's just how it is. Technically KDE or Gnome may have aspects that are 'more correct' or 'less correct', but what matters is what happens in practice, not theory.

Apple regularly breaks it's own usability rules.. even in applications it embeds in OSX by default.. it's far from uniform. And when you look at Windows... MS Office vs Window Media Player vs Internet Explorer. It's all over the map in terms of UI shapes, placements, and terminology. Many of the favorite programs that people have don't integrate very well at all and don't have proper look-n-feel.

Despite Firefox not integrating well with Gnome it's still the most popular browser out there behind Internet Explorer. VLC is a favorite on OS X and Windows AND Linux, yet it's UI and how it goes about handling media is very different from any other application. There are whole hosts of reasons why this is so.. however "right" or "wrong" it is it's simply the reality we have to operate in.

What really matters, in terms of improving the UI and having slick interfaces, is usability testing and having developers work with real UI designers and taking a scientific approach to user feedback. This means paying people to come in and use your software.. give them tasks to accomplish, see how well they do, find out what sucks and change it.. rinse and repeat. It's basically impossible for a programmer to design a proper UI on his own becuase his knowledge of the system is much too high to be able to sit back and figure out how a real user will approach and use their program.

This is the biggest problem with Free Software desktops.. There has been very little to none involvement from UI designers and very little to none usability testing.

As far as I know only Gnome has formal usability testing from Sun Microsystems and Novell.. and lessons learned from those people lead directly to the much-hated "Gnome HIG" that demands simple user interfaces and limited dialogs and configuration options. Of course there remains massive more work to be done on that forefront.

When faced with real users and real testing it often turns assumptions on their heads about the 'right' or 'wrong' way to approach UIs.

-------------------------------------------



The biggest problem you run into with making things 'effortlessly' cross-platform is that you must design your UI to only be the 'lowest common denominator'. You can only use functionality that is present on every platform you want to support. If you want to use a UI feature that is not present in Linux then you cannot use it in OS X or in Windows.

A couple of examples:

WXwidgets. http://www.wxwidgets.org/

That is a _VERY_ cross-platform toolkit. If you write your applications using it then it'll integrate well and look almost exactly like a Windows or OS X or Gnome or QT application.

But it's functionality is also very limited. Say, for example, I want to write a applicaiton that listens for USB plugins and notifies the users about it and can perform some action (like writing a Live Linux image to it or have it look for usb-mass-storage-using media players). If I was using GTK then that would be very simple to use the Gnome-style application loop and listen to Dbus messages and react to events happenniing in the system.

But Wxwidgets cannot provide any sort of functionality like that. Even if the look-n-feel is there I'd still have to program specific code for each platform and it would, in fact, be much more difficult then if I just stuck with QT or Gnome stuff.

Java ---

Java has very strong cross-platform support. So strong that it is very recommended to application developers that they cannot use any more then one mouse button!! Why? Because Java is suppose to be write-once run-everywere and some platforms that you'll run on will not support using more then one button.

------------------------------------------


The only way to really have easily cross-platform KDE aplications is not to have proper integration into the UI of the OS... the only really way it can remain very easy and have the application developer not have to worry about platform specifics is to essentionally port the entire KDE infrastructure from Linux to everywere else. (which they have pretty much done. You can run KDE on Windows and run KDE on OS X; at least experimentally)

Even then though there are going to be limitations on what you can do. For example... OS X, Windows, and Linux all have dramatically different sound systems with lots of different capabilties and quirks. If you have very advanced sound application (say your targetting studios) then your going to be forced to make a bunch of platform-specific code no matter what you do.

Other then that if your interested in cross-platform applications that integrate VERY natively the best way to accomplish that is often to design your application to seperate out as much of the application logic from the UI and just write a bunch of platform-specific code for each platform you want to run on. Which is obviously not commonly done.