GNOME 3.0 May Not Come Until September 2010

[mat69]

Please don't generalize, there is no such thing as the typical KDE or Gnome user.

Use what you want to use and be happy with that.

Evolution, Accessibility, GStreamer (with automatic codec installer), Seahorse, Orca, Tomboy, Cheese, GVFS, Brasero, fast user switching, tabbed Nautilus.

Well tbh Nautilus having tabs is not a groundbreaking feature -- rather a sign for the stubborness of the devs as that was a real old feature-request -- neither is fast user switching or Tomboy or Evolution (what groundbreaking changes were there, IMAP?).

Imo there are other, better examples that happened in the Gnome ecosystem, like PackageKit, ConsoleKit and DeviceKit. Great basis for future developement and used by other DEs as well. E.g. look at that new tool to manage your drives that uses DeviceKit, really nice that you can look at the S.M.A.R.T figures and make a test etc. from a nice gui. There happened a lot though imo Gnome devs -- or the ones deciding what comes in -- often are too conservative and don't use the potential Gnome could have --> options are not bad neither are they evil! Still their conservatism is also one of Gnome's strengths, its like Ying and Yang finding a balance and that is pretty hard.

On the window decoration part, nearly everything is possible with the changes Marco Martin (IIRC!) did, you could make the windows look like the plasma widgets etc. there just needs to be someone writing a theme for it. I have forgotten the name though.

Consistency, man, the point is consistency. If I hover on a button, I expect it to shine. I expect a tooltip to unfold and explain its function. If neither happens, it's broken and should be fixed (at least in the Gnome world, but I doubt KDE views this differently).

Don't generalize please. There are lot's of KDE devs that aren't really happy with situations you outlined above, there are also some who don't really like the new notifications etc. (heck I e.g. don't like the new "thingy" to add plasmoids, the dialog was not ideal either but could have been improved but that is just my opinion) but as it often is devs can be stubborn, especially on their brainchild. --> heck if I have compositing on I also want the option to turn transparency of my plasma widgets off, I don't like transparency.

On Gnome 3 what I'm missing is Gtk 3, there is nothing groundbreaking for Gtk 3 really, look at the changes of Qt, look at what we will have with Qt 4.6 in a few weeks --> animating has become so easy, I know there is Clutter, but clutter needs OpenGL Qt does not. On the Qt board there happened so much and in the following months there will happen even more (QMF for one).

[BlackStar]

Well tbh Nautilus having tabs is not a groundbreaking feature -- rather a sign for the stubborness of the devs as that was a real old feature-request -- neither is fast user switching or Tomboy or Evolution (what groundbreaking changes were there, IMAP?).

I listed both evolutionary and revolutionary features to counter the claim that Gnome is stagnant.

Tabbed Nautilus is an evolutionary feature in itself, but it has actually revolutionalized how I use my computer (actually, this was the single KDE 3.x/4.x that I missed in Gnome).

Evolution: several features have been introduced over the years. For me, the biggest is exchange support (I haven't been able to get this to work on any other open source mail client). Laugh all you want, but this is the feature that allowed me to stay with Gnome/Linux while working for a Windows shop.

Tomboy: this is an innovative little application that packs a punch. It just keeps getting better with every version, with remote sync being the killer feature. Dammit, I used to enter my notes into version control just to be able to sync them between my workstations!

Imo there are other, better examples that happened in the Gnome ecosystem, like PackageKit, ConsoleKit and DeviceKit.

No argument from me here - I just listed a number of applications that have changed my day to day experience when using Gnome.

Consistency, man, the point is consistency. If I hover on a button, I expect it to shine. I expect a tooltip to unfold and explain its function. If neither happens, it's broken and should be fixed (at least in the Gnome world, but I doubt KDE views this differently).

Don't generalize please. There are lot's of KDE devs that aren't really happy with situations you outlined above, there are also some who don't really like the new notifications etc. (heck I e.g. don't like the new "thingy" to add plasmoids, the dialog was not ideal either but could have been improved but that is just my opinion) but as it often is devs can be stubborn, especially on their brainchild. --> heck if I have compositing on I also want the option to turn transparency of my plasma widgets off, I don't like transparency.

Don't take my comment out of context. I replied to a very specific comment, which roughly said that (a) a specific example of inconsistency is fine (posted in the blog I linked) and (b) any user who is thrown off by this inconsistency is an idiot.

I don't believe that KDE developers are happy with such inconsistencies either, nor do I believe they are the result of stubborness. In all likelihood, they will be fixed sooner rather than later.

I do believe, however, that Gnome developers take a more proactive approach to such issues - which is also why they are delaying Gnome 3.0. It's simply a different approach to development - release earlier with some issues (KDE 4.0) or hold back and polish (Gnome 3.0). (Note: both approaches have their merits, I'm not advocating one over the other).

On Gnome 3 what I'm missing is Gtk 3, there is nothing groundbreaking for Gtk 3 really, look at the changes of Qt, look at what we will have with Qt 4.6 in a few weeks --> animating has become so easy, I know there is Clutter, but clutter needs OpenGL Qt does not. On the Qt board there happened so much and in the following months there will happen even more (QMF for one).

An OpenGL dependency is not bad now that most (relevant) open source drivers have reached ~1.5 level. Given current trends, I'd expect more and more applications to use hardware acceleration as time passes (Qt included).

If you are willing to look out of the box, there's a lot of innovation going on on the Gnome side, too: Moonlight on the desktop is simply awesome - and it doesn't require OpenGL either. (I expect a Clutter-based renderer will be implemented in the future, right now there's a Cairo and a OpenVG(!) based renderer.)

[Joe Sixpack]

Never said otherwise. It's intricately tied in the sense that if you take it out, half of Gnome will break.

I'm really not trying to argue with you, but that comment is completely insane and so is the reasoning behind it.

If you remove libjpeg or libxcb half of Gnome would break. The same for glibc. None of the aforementioned packages are part of Gnome or KDE. Let's not take it back to 2003 when every package written in Gtk+ or began with the letter "g" was called a Gnome app. The "g" stands for GNU - not Gnome.

[BlackStar]

I'm really not trying to argue with you, but that comment is completely insane and so is the reasoning behind it.

If you remove libjpeg or libxcb half of Gnome would break. The same for glibc. None of the aforementioned packages are part of Gnome or KDE. Let's not take it back to 2003 when every package written in Gtk+ or began with the letter "g" was called a Gnome app. The "g" stands for GNU - not Gnome.

Note, I said "intricately tied" which means something different than "part of". Gstreamer is intricately tied with Gnome, because Gnome is its primary consumer. For example, I really doubt that Gstreamer would break its API/ABI without first communicating with the Gnome developers.

Gstreamer is obviously not part of Gnome in the sense that it relies on Gnome libraries or infrastructure - it is a depenency and a significant one at that.

[bartek]

It would be nice if GNOME 2.30 would have the option to login in to GNOME 3.0 (alfa/beta/...)

So users will have the option to switch whenever they want from 2.30 to 3.0 and back. More people will test and give their opinion and comments on the upcoming 3.0 release..

[mat69]

Don't take my comment out of context. I replied to a very specific comment, which roughly said that (a) a specific example of inconsistency is fine (posted in the blog I linked) and (b) any user who is thrown off by this inconsistency is an idiot.

I don't believe that KDE developers are happy with such inconsistencies either, nor do I believe they are the result of stubborness. In all likelihood, they will be fixed sooner rather than later.

Sorry, guess I completely misunderstood your initial post or maybe mixed something up.

An OpenGL dependency is not bad now that most (relevant) open source drivers have reached ~1.5 level. Given current trends, I'd expect more and more applications to use hardware acceleration as time passes (Qt included).

Qt can use OpenGL, the difference is that it is not a hard dependence. And as KDE strives to be multiplattform having a hard dependence on OpenGL is not a good idea. Also I'm not sure if Clutter supports OpenGL ES.

If you are willing to look out of the box, there's a lot of innovation going on on the Gnome side, too: Moonlight on the desktop is simply awesome - and it doesn't require OpenGL either. (I expect a Clutter-based renderer will be implemented in the future, right now there's a Cairo and a OpenVG(!) based renderer.)

Just to be sure that I don't misunderstand you here again, but what do you mean with Moonlight?

[kwanbis]

Hello, new to the forum.

Long time COMPUTERS user (spectrum, ti994/a, commodore 128, ibm pc xt, amiga, etc, etc), and long time OSES user (DOS, CPM, DR.DOS, Windows, OS/2, BeOS, linux, OSX, etc).

As some have already say, beauty is in the eye of the beholder. For example, in general, y don't understand what is the obsession of some to "dark" themes.

If you ask me, i rather have the Windows 2000 look than the 7 look.

But then again, i like OSX look better.

I'm really not a heavy Linux user, but i use it from time to time.

I'n fact, i'm going to replace Windows7 with Ubuntu/SUSE/Fedora/other soon.

When i started, i tested both Gnome and KDE, and what i remember is seein KDE and think "ugh, 1000 options to do the same thing", and that it looked too much like windows. This was with kde3.

Since then, i've been using Gnome when i do use linux.

I would be checking kde4, but the point is i rather have simplicity and elegance oveer flashiness any day.

I prefer ThinkPads look over Alienware's or HPs ones: minimalism rules.

My 2cents.

[BlackStar]

Qt can use OpenGL, the difference is that it is not a hard dependence. And as KDE strives to be multiplattform having a hard dependence on OpenGL is not a good idea. Also I'm not sure if Clutter supports OpenGL ES.

It does support OpenGL|ES. Given that OpenGL is supported pretty much everywhere, I don't see why a hard dependency is a bad thing. Even if you lack hardware support, you can always fall back a software-based OpenGL implementation (no, soft OpenGL doesn't have to be slow!)

I think that's much better than writing separate backends for GDI, XRender/EXA, Quartz, OpenGL, OpenGL|ES *and* a software rasterizer to boot. This hasn't been easy for Qt (check their blog) and - even after all those years - many of those backends are still lagging behind the software rasterizer.

Maybe it's a better idea to focus on a single, well-optimized OpenGL & OpenGL|ES implementation, rather than implement a dozen half-baked backends.

Just to be sure that I don't misunderstand you here again, but what do you mean with Moonlight?

(Donning armor)

http://www.mono-project.com/Moonlight

(Trolls incoming in three, two, one!...)

In all seriousness, this is a great project. Easy to animate. Runs on the desktop (on Linux at least). Runs on the browser (everywhere!) It can be used directly by many programming languages instead of just C or C++ (sidestepping the "sucky, outdated Qt/GTK bindings issue")

[mat69]

I think that's much better than writing separate backends for GDI, XRender/EXA, Quartz, OpenGL, OpenGL|ES *and* a software rasterizer to boot. This hasn't been easy for Qt (check their blog) and - even after all those years - many of those backends are still lagging behind the software rasterizer.

Granted that makes sense

Maybe it's a better idea to focus on a single, well-optimized OpenGL & OpenGL|ES implementation, rather than implement a dozen half-baked backends.

In all seriousness, this is a great project. Easy to animate. Runs on the desktop (on Linux at least). Runs on the browser (everywhere!) It can be used directly by many programming languages instead of just C or C++ (sidestepping the "sucky, outdated Qt/GTK bindings issue")

I have no clue why this should be an innovative Gnome project.

[ciplogic]

In 7 years, can you honestly name me 5 major improvements?

User features:
- taskbar overhaul: blinking notification to aplication, drag and drop, drag and drop between workspaces
- menubar reorganized to have much more clarity from two entries to three entries
- open-save dialogs are much better, can be added new sources (like search integration with beagle/tracker)
- standardized theme and set of icons that make any GNOME distro to give a non-distro specific Gnome look (Clearlooks)

Technology:
- support for composite (also cause of cairo) and which enable some applications as gnome-do to look nice on a composite desktop
- desktop search API
- help technology replaced
- improved accesibility with Orca
- telepathy framework
- PolicyKit integration
- Freedesktop integration

Gtk:
- toolkit: switching from Gdk to Cairo
- mono interoperability
- switching from gvfs to gio
- GtkBuilder replaces Glade (which went to version 3 that do not generate source-code as it's predecesors, but was a separate library)

Hardware API abstractisation:
- D-BUS to send notifications as low batery, but not only. Is a IPC framework that will be adopted later even by KDE
- swith hardware API to Hal (and maybe later DeviceKit, but may be a part of Gnome 3)
- switch from ESD to GStreamer

This seems to be all I remember, in short, sorry for what was skipped, but comparing GNOME 2.0 with Gnome 2.28 is plain silly. Is just like saying that GCC 4.0 is the same with GCC 4.4 and is use the same Gimple language and it stores it's definitions in SSA form, and we need to release GCC 5 as GCC 4.0 do not get enough fast binaries (compared with compiler X) and GCC 4 is based on the same bad technology.