KDE

KDE was a fine desktop environment at one time.  The 3.x branch was stable, snappy, and struck a balance between bling and usability.   4.x changed all that.

Bling became front and center, 4.0 was pushed and marketed with great fan fare with  only faint whispers to the dissenters that 4.0 was really “beta”.  Beta is too kind IMO.  I would consider 4.3 “beta” and everything before it “alpha”, but it’s not my project –so whatever.

What happened in KDE? Well tons of features were lost and it took almost a year to get them back in, some still haven’t returned.  Dual screen functionality broke, screen space usage was and is horribly abused (though the panel has gotten better).  The transition to QT4 was supposed to make everything faster, yet KWin is now probably the clunkiest, slowest window manager in the linux landscape.   KDE takes longer to startup over every DE out there too.  And for what?

So we can have a crappy Kickoff menu system? Have options on how desktop icons are used taken away (now fixed of course). ? Akondi server that’s an all around bad idea, and so we can turn desktop widgets round and round.   All those headaches over all this time for what? For bling.  Nothing else.  Just bling, because no one got any performance jumps out of this.

Even in 4.3, in Arch and OpenSuse, with nv, nvidia and mesa drivers the compositing effects in KDE barely act better than alpha quality, freezing up X.  And with the changes to X to disable Ctrl Alt Backspace, we now, just like our Windows brothers, must hard reboot in order to get our DE back up and running.  Ridiculous.

So my question is–based on recent experience, is such drastic re-write really the best approach at this point in time?  And on the assumption that it is (which i don’t think it is at this time) do we want to break this many things for the sake of bling?  Below are my issues with Gnome 3.0.  Gnome folks have to understand too, there’s a lot of people using Gnome now purely because of how breaky KDE has become.  If you are going to break the 2 major DE’s at roughly the same time, where are they to go?  And what negative effects does this have on the linux desktop all around? Because it seems to me the Gnome team and the KDE team are sabotaging the Linux Desktop.  Maybe unintentionally and with the best of intentions, but as they say, “The road to hell is paved with the best of intentions”.

Usability

This is where I think Gnome (in the 3.0 branch) and KDE 4.x have really lost it.  The menu systems.   This is something that over the years Linux got right, and everyone else got wrong.  Point blank.

The best part about the menu system of Gnome currently (2.2x) and of KDE of yesteryear (or recently added classic menu) is that apps are organized by type. So its very few clicks. Want firefox? Applications –> Internet –> Firefox. With the only actual clicks being on Applications and Firefox. 2 clicks, bam the thing is open in a hurry.

I can’t stop wondering, what problem KDE with Kickoff and Gnome with ..whatever this garbage is, are trying to solve? I haven’t RTFM and looked for the theories behind this, but I really wonder. I’ve never thought of needing any of the new features  There’s this story I heard that Palm used to employ “click counters”. People whose sole task it was to count the amount of clicks it took to do important things.  Gnome might want to use some of the money donated to them and hire some of them.

Or, just read this blog and understand that the OLD MENU SYSTEM MAKES SENSE.

Windows menu system has always been a complete mess because they allowed every vendor to have their own spot on the menu system. Hell sometimes the vendor gets multiple spots:

http://www.softpedia.com/screenshots/Vista-Start-Menu_1.png.  And that image is relatively clean compared some of the stuff I’ve seen.   OS X doesn’t even bother with application menu’s and has you browse a folder of apps and drop what you want on your dock bar. Again, terribly inefficient and time consuming.

Gnome and KDE got it just right. A menu system that’s simple, effective, etc.  But why let a good thing just sit? That’s no fun is it?

And now they’ve gone and messed it all up. Gnome3.0 is atrocious.   Now I have to ZOOM OUT my whole desktop, just to get to a list of Favorite Applications.  What about a non favorite? How many more clicks will that take?   This is just like KDE’s kickoff  which is equally as ridiculous, b/c now a simple “mouse over” won’t open a menu, you have to click to expand it. Of course they want you to use their nifty little search feature built in and favorite everything, and while that works for the most used apps, getting to the lesser used ones becomes a terrible pain. I’m glad they at least kept the classic menu around. I still don’t like KDE4 but at least they reimplemented that.

Annoyance

Aside from it being a usability problem, does no one else see that the whole desktop zoom in/out thing is a cheap gimmick? It looks super neat the first 5 times, hell, it might the first 50 times or 100 times.  But imagine using that for 6 months, or a year, or more.  Not so fun anymore, in fact, the dizziness of it is going to grow incredibly annoying.   No one in a work place is going to take a desktop serious if the users can’t just get their apps open to do their work.

I want access to my apps to be front and center, and organized properly. All this favorited stuff and recently used stuff and this dizzying desktop zooming just to access a menu is just unusable, un-necessarily resource intensive. it might be neat at first but would get incredibly old, and frankly.. it’s just one more thing that’s going to freeze a user’s system. In no way, shape or form is this a good idea.

Respecting Users Resources

And speaking of resource utilization, I’ve already mentioned my problems with KDE 4 because they decided to make a new window manager  run tremendously slow, and effects that freeze X.  And hell, I have a relatively modern and powerful video card!  Linux is used all over the world, and notorious for being used on an incredibly wide range of hardware.  One of the main advantages Linux touts is the fact that it’s snappy, it’s a solution to the bloat of Windows.   But with these changes, not anymore.

And in fact, I think Windows can handle their bling better than linux can handle its own.   Windows has a certification process for it’s OSes with OEMs.   Apple controls their hardware. Linux, nor any of the companies involved with linux has no such thing going on with OEMs. At all. And because of that, I think default compositing for the next 4-5 years is always going to be a horribly bad idea. Especially to this degree.  And especially in nations where the populace is probably not going to have  the latest and greatest hardware.  And these are the developing markets Linux needs to capture. This is an area where linux can succeed.  But if it’s hardware requirements rivals Windows Vista/7, you’re just shooting yourself in the foot.

Not to mention the massive amounts of problems you’re creating for the end users. One more thing to freeze up and strain a system.

And what about servers? Red Hat and Novel push their servers with a DE on top.  Red Hat is the #1 Linux company in the world and they prefer Gnome, and while Novel is more KDE centric they have done a ton of work to Gnome.  Both push their GUI configuration tools.  Are you telling me that my servers are now going to require some suped up video card capable of accelerated graphics just so I can run Red Hat’s or Novel specific tools? Are you f–ing kidding me?

Interoperability with Other Window Managers

I’d also really like to know why in the world is Gnome breaking compatibility with other window managers. This will be the end of Compiz, Xmonad, and Openbox being used within Gnome. It just seems very.. anti-linux to go out of your way to break interoperability so badly. There’s also got to be few open desktop standards they’re breaking or ignoring in order to accomplish this as well, i assume.  I honestly don’t know.

Gnome 3.0 and Ubuntu LTS

And i know this is more of an Ubuntu decision than a Gnome one, but like I already pointed out with Red Hat and Suse servers above, Gnome is not taking into consideration the massive amount of stuff that depends on it.  This version of Gnome is slated for Ubuntu’s 10.04, they’re next Long Term Support Release.

Experimental user interfaces do not belong in Long Term Support releases. PulseAudio did not belong in Ubuntu Hardy, and Gnome 3.0 has no place in 10.04, the next LTS release. Canonical’s appetite for experimentation, while commendable, verges on the reckless.  This does not need to be in a LTS.  At all.

It appears as though Canonical is leaning on features and novelty to sell their products, and stability and workflow are taking a backseat.  It seems this idealology has leaked into Gnome itself.  That will fill the cheap seats with people eager to try new things — who will be accordingly burned when their new toys don’t work like they’re supposed to. Then there comes the backlash, which gets aimed at “Linux” instead of “Ubuntu”, and “Ubuntu” over “Gnome”.  I don’t think Canonical or Gnome appreciates the gravity of being the mainstream ambassador of an operating system or Desktop Environment.

Either way, Gnome 3.0 will not be ready in time for an LTS release of Ubuntu Linux.

So what are your thoughts?  Is this the end of the Linux desktop (oh i’m so dramatic!) or will this spur a huge exodus to Xfce and various Window Managers?

If you’re like me, you want Gnome to be around for a long long time and you care greatly about this issue.  Maybe we can do something about it.   File these problems as bug reports with Gnome and the distro of your choice and let your opinion be heard.    Or lets organize a team to help Gnome 2.x grow in a healthy and gradual manner.

I dunno, I just felt someone needed to bring up the ridiculousness of all this.    We want more people in the linux desktop, we want more people pleased with the desktop.  Not less people.  We want better usability not worse.  Bling should NEVER sacrifice usability.  Ever.

I’m open to thought and opinions about what we (those that agree) should do.   I’m also open to being cursed at and shouted down.  (I”m used to it.)