KDE 4.7 – a First Look At Beta 1

A few days ago, the KDE project shipped the first beta of the upcoming 4.7 release. Reader dmbkiwi submits a link to a rundown of what 4.7 looks like, snipping from which: "Previously it was Gnome that was the steady plodder making minor incremental changes through the 2.x series, building stability and only adding minor features. However, with the recent releases of both Gnome Shell and the Unity desktop on Ubuntu, the Gnome/Ubuntu side of the desktop linux equation has made radical and controversial steps away from the well loved Gnome 2.x series, leaving KDE 4.x as the 'steady as she goes' option."

Comments

The interface doesn't need to be changed much

[elucido]

People who use KDE are typically coming from Windows so the default should look similar. However the good thing about linux is customizability. As long as we can customize it to look however we want most of us will be happy.

Gnome and Ubuntu Unity have removed the linux edge of customizability. It's only a matter of time before I switch from Gnome 2x to KDE 4x. The next big step for Linux would be to take advantage of 3d rendering to improve functionality further. The zoom is something I use on a regular basis. Perhaps being able to flip windows(frames) and being able to write on the back of them would be a useful feature as well. There are plenty of ideas for functional eye candy but I think linux is at the point now where it shouldn't look towards Windows or OSX for new feature ideas, and it shouldn't try to fix an interface which isn't broken, it should just be adding new features and options, new eye candy which increases usability, and new more powerful abilities, such as intelligent agents that a user can program to automate certain tasks such as burning a DVD, searching several search engines to find certain information on certain topics, all of this could benefit from agent based AI.

I suggested this to the linux community years ago and their excuse was there wasn't enough bandwidth. It's 2011. The majority of the country is broadband now. There is enough bandwidth to build an intelligent agent into KDE and if they wont do it then I might just go ahead and do it for them.

(For anyone who doesn't know what an intelligent agent is, http://en.wikipedia.org/wiki/Multi-agent_system an agent is a robot, in this case multi-agent is multiple robots which search for and process specific information you tell it to. http://en.wikipedia.org/wiki/Software_agent )

The agents in a multi-agent system have several important characteristics:[4]
Autonomy: the agents are at least partially autonomous

Local views: no agent has a full global view of the system, or the system is too complex for an agent to make practical use of such knowledge

Decentralization: there is no designated controlling agent (or the system is effectively reduced to a monolithic system)[5]
Typically multi-agent systems research refers to software agents. However, the agents in a multi-agent system could equally well be robots,[6] humans or human teams. A multi-agent system may contain combined human-agent teams.

Re:The interface doesn't need to be changed much

[elucido]

Thats actually not a bad idea. You could fit roughly 3 times the information (I hope my math is correct?) in the same space if you move from squares to cubes.

Re:The interface doesn't need to be changed much

[billcopc]

I still don't understand why KDE and Gnome are such big deals. Maybe I'm too Windows-centric, but what I expect from the GUI is simple: a launcher/taskbar widget, configurable window management and theming, and a handful of integrated utilities or configuration panels that govern common functionality among all apps (e.g. network shares, security defaults, notification prefs, video accel).

Beyond that, the rest of KDE seems like truckloads of cruft to me. I find the bundled apps largely deficient in functionality and stability, they're like "store brand" knockoffs of specialized 3rd party apps. Rather than wasting so much effort on these bastard subprojects, why not deliver a solid API and widget library that allows 3rd parties to properly integrate with the look and feel ? Let the GUI people focus on building the GUI, and let the app people focus on apps.

KDE 3.5 was fast, lean, maybe a little hard on the eyes but it did everything I needed without getting in the way. Everything since then has been a bad acid trip through OSX envy and good-old-fashioned programmer-designed atrocity. Just look at Windows 7, they pared it down from Vista to be as simple and efficient as Microsoft can be. Less baked-in functionality, but plenty of hooks to extend it IF AND WHEN NEEDED. Isn't that supposed to be the Unix way ?

Re:The interface doesn't need to be changed much

[seyyah]

People who use KDE are typically coming from Windows so the default should look similar.

Where does this notion come from? I've see in it before and I doubt it has any merit.

In fact, I would expect that the majority of people coming from Windows use Gnome since it is the default DE for Ubuntu and other popular distros.

Re:The interface doesn't need to be changed much

[VortexCortex]

Uh... I already do this.

Re:The interface doesn't need to be changed much

[hairyfeet]

Because, and I'm sure to get hatred for pointing this out, a lot of the developers out there seem to suffer from Cargo Cult Usability where they implement basic ideas without understanding the underlying structure which is why you saw a previous poster talking about "dept store knockoffs" because when you implement some of the front end without the underlying structure it feels like a badly done copy.

Take Gnome for example. they have ripped off (homage, whatever) a lot of the Mac GUI without understanding how the structure ties in which makes it 'off". For example they have the traditional Mac menu in the correct placement but their DE is windows based and Macs are application based which causes it to make no sense. Since Macs are app based all apps use the same menu at the top whereas with gnome each app typically has their own menu layout or at least did last time I tried it with Ubuntu 10.04, which makes having the top menu kinda pointless.

With KDE they seem to follow the Windows model but yet again they don't implement the features, just the look or at least that was the case last time I tried it (again with Ubuntu 10.04) because while they had a lot of the familiar layout they didn't have breadcrumbs or Readyboost or superfetch or many of the other features that makes Windows more usable.

So I think Canonical is in the right here, even if it falls flat or takes a while to get solid, in that the way to go is to forge a "Linux centric" model where you have a completely different GUI that won't make people feel "cheap knockoff" when features they are used to in Macs or windows aren't found or are different. By switching to Unity+Wayland they will have a different "look and feel" to other OSes, thus removing some of the preconceived notions people have when you use a Windows or Mac centric layout.

Re:The interface doesn't need to be changed much

[Risen888]

I still don't understand why KDE and Gnome are such big deals. Maybe I'm too Windows-centric, but what I expect from the GUI is simple: a launcher/taskbar widget, configurable window management and theming, and a handful of integrated utilities or configuration panels that govern common functionality among all apps (e.g. network shares, security defaults, notification prefs, video accel).

You were deprived of a proper desktop as a child. You know nothing of multiple workspaces, the ability of your applications to share their data with each other, even the simplest things like changing the color of your window decorations is beyond your ken. It's like you were raised in a cage.

I find the bundled apps largely deficient in functionality and stability, they're like "store brand" knockoffs of specialized 3rd party apps.

I think you're out of your mind. Okular is the best document viewer I have ever seen. Show me a pdf reader that does a third of the things that Okular does, and does them half as well, and I will eat my hat. Kontact is absolutely gold. Even the file manager has been doing things for three years that Windows Exploder still can't even imagine doing. Marble...well, I was gonna say Marble's the best at what it does, but actually, it's the only application that I'm aware of that does what it does.

why not deliver a solid API and widget library that allows 3rd parties to properly integrate with the look and feel

Yeah, we got that. We've had it for years. Have you looked? No you haven't, have you?

KDE 3.5 was fast, lean, maybe a little hard on the eyes but it did everything I needed without getting in the way. Everything since then has been a bad acid trip through OSX envy and good-old-fashioned programmer-designed atrocity. Just look at Windows 7, they pared it down from Vista to be as simple and efficient as Microsoft can be.

"foo n-1 was the best thing ever, new is crap, Windows 7 is shiny." Okay then. Use Windows 7.

Re:The interface doesn't need to be changed much

[fnj]

I don't think KDE4 is overblown except in one respect. If you could just can that imbecilic desktop and replace it with a single simple folder view like in KDE3 and Gnome2, where you can put launchers and objects, KDE4 is basically perfect. Now, I haven't been able to figure out how to rip out that crippling piece of garbage from KDE, but I am sure the KDE team could easily add a single radio button to allow the user to just enable or disable it. It's like how they let you switch the start menu to the normal, useful "classic style," instead of the godawful new style which copies one of the most HATED and DESPISED "innovations" of Vista.

In all other respects, I see no fundamental flaw with KDE4. I find it in no way mysteriously slower or more ponderous than KDE3. I am just perplexed when people claim this. Obviously, the first thing you do when you bring it up is completely turn off all the "desktop effects" horse shit, and then it works fine.

And as far as I can see KDE has some wonderful apps. Kate, for example, is a superb multi-document text editor. Gnome has nothing remotely comparable. I know of no standalone one that is better. KDE Office would be a wonder if we weren't spoiled by Open Office, so I admit I don't use much of it. Obviously, I use Firefox instead of Konqueror (usually). But I see no way in which the KDE guys have built a less than first class API for 3rd partiesa to properly integrate with. If they won't do it, and instead use the GTK horror, it's hardly KDE's fault. KDE's is vastly superior in every way.

Konsole is so many orders of magnitude better than Gnome terminal or anything else, that it is like the adults vs the kindergarten to compare them.

If you really and truly want a bare desktop with no cruft at all, you just use Xfce or LXDE. But I must warn you that they have substandard "little things." The clock cannot be adequately customized. The other applets are similarly deficient. I suppose we could port forward all the superb Gnome2 applets if we had the energy, but gosh darn it, I just want to USE a desktop that is neither INSANE nor DEFICIENT as it is.

Re:The interface doesn't need to be changed much

[fnj]

I agree almost across the board. My only issue is that I think Gwenview is a piece of crap from a UI standpoint compared to the old Kuickshow. Kuickshow was an inspired PERFECT app. You can still compile kuickshow under KDE4, and it still works perfectly. I just don't understand why they refuse to maintain it as a fully supporteds piece of KDE4.

Re:The interface doesn't need to be changed much

[karper]

Regarding the ctrl+v and rightclick-paste comment, it's quite simple to 'fix' this. There's an x clipboard and there's a system clipboard. You can sync them using klipper (the thing that looks like a pair of scissors in the kde system tray). If sync'd, you can copy by selection or copy by ctrl+c and pasting (ctrl+v) always pastes the last copied thing.

Your complaint is valid, though, in that middle-clicking pastes only what you selected and copied, not what you ctrl+c'd. My solution to this is simple: use select/paste. Use ctrl+c/v only if you plan on selecting multiple things which you don't plan on adding to your "clipboard".

Re:The interface doesn't need to be changed much

[The Master Control P]

If only there were some terms like "alpha" and "beta" and "rc" to describe software that's not yet ready for general use and only for developers.

Or in the case of kde 4.0, "early sub-pre alpha." I remember quite vividly at the time of launch that they weren't saying anything like "don't use this if you're not a developer." No, I remember reading article after article raving about how awesome it was, how innovative that was, how pretty it was, how it and all the new backends like Akonadi and Strigi and Nepomuk were gonna revolutionize the whole desktop experience. Then everybody absolutely hated it, I personally found it used 10x the memory to do the same things as before, it turned my screen into a slideshow, and it crashed every 10 minutes. The only unequivocally positive change I noticed was that kmail now remained responsive while waiting on the remote mail server!

The "4.0 was a dev-only release" bullcrap started to try and retcon away just how badly they'd screwed up. And I can understand... I wouldn't want to remember 4.0 either.

Re:The interface doesn't need to be changed much

[Kjella]

I guess you missed the part where Qt closed the bug as OUT OF SCOPE, as in not a Qt problem but a Linux problem. But I see the anything that breaks with the holy gospel gets modded down, so...

Re:The interface doesn't need to be changed much

[Omestes]

In GNOME, You Edit your Preferences.

Not any more.

More on multi-agent based AI

[elucido]

So I mentioned this in my previous post and I recognize some people don't know why or don't understand why this would be useful. So I'll give some examples of what agent based AI can do for those who don't know and how it could be implemented.

To implement multi-agent based AI on linux first there would need to be a backend or a framework of some sort that would allow scripting languages such as python, ruby, and perl to connect to it. The framework or backend would have to be written in C for certain intense data processing tasks. The front end should allow programmers of all sort to write their own scripts in their favorite scripting languages to create robots. These robots should have the ability to automate system processes.

For example I decide I need to do research on artificial intelligence because I don't know what it is, so I should be able to tell the robot to search Google, to find X amount of articles on artificial intelligence which meet certain criteria. This could be done using regular expressions. But of course this isn't all that I need to do. I have a to-do list for this specific robot related to the topic of AI, to download certain files from the net and install them, to then load up and use certain files to process certain data. All of this should be automated completely and should happen in the backround and it all should be related to the topic of AI.

The news robot on the other hand I would program to act as an RSS feed, this robot would look not just at specific websites such as slashdot, but for specific articles on slashdot and present those articles along with research on certain keywords or buzzwords it thinks or suspects I know little about or wont understand.

The log analyzer robot could analyze logs for me and highlight any potential redflags, and then if it finds them run through an automated process that I determine is best for dealing with these redflags.

Each robot would be assigned to a task. Each robot should have the ability to do what the user could do, and it should be simple to show the robot or program the robot into doing it a number of very highly complex tasks.

The problem with using computers is most of the stuff we do each day is just routine. Most of us fit into certain patterns. Robots would allow us to save time, we can leave the computer on all day or all night and it will do a number of boring clicks and boring tasks that take up a great deal of time. This saves time and increases productivity.

two weeks without KDE... and not missing it

[C0vardeAn0nim0]

before someone mods me '-1 flamebait', let say a few things:

1- NOT a gnome fanboy. i dislike gnome in all it's incarnations, always did.
2- i use windowmaker. always have, always will
3- i only had parts of KDE installed to use some of it's applications from inside wmaker (mostly K3B, koppete, ktorrent and dolphin)

now, in the last two weeks i apt-get purged all things KDE4 from my system (kept only pana, a fork of amarok 1.4). the reason is that newer versions of KDE were starting to interfere with my way of doing things. what tipped the scale was keyboard configuration.

you see, i don't use graphical login managers, i log from good old fashion console, then type "startx" by hand. i consider this a must, since i use debian unstable, so breakeage of x.org because of updated kernel, ati drivers, etc sometimes happens. this means i have keyboard with swapped ctrl and caps lock, as well as locale (pt_BR) configured on the console. with wmaker i don't even need a keyboard section on xorg.conf, it just goes with what's configured on the console. that is, until you fire up a KDE app and it loads all those libraries. other thing that i had configured manually was CPU frequency management, so i don't run the risk of overheating the notebook when doing something CPU intensive on the console. i use userspace governor with kpowernowd and it works just fine.

keyboard becomes all messed up, KDE insisted in changing the frequency governor to wathever it damn well pleased, not to mention taht the load time for all those libraries was atrocious, i had to wait some 20 to 30 seconds until kopete, bluetooth applet and power applet loaded.

after i ditched everything, now i'm using XFE as file manager, pidgin for IM, gnome's bluetooth applet, xfburn and qbittorrent (a qt app. it doesn't load all the KDE libs like ktorrent). the result is faster load times for the GUI, less anoyance and no loss of functionality.

if the KDE guys make their environment behave better when a KDE app is loaded from some other window manager, maybe i'll give it another shot. until there, it'll stay out of my computer. i have better things to do with my time than fight against misbehaved apps that try to wrestle control of my system out of me.

Re:two weeks without KDE... and not missing it

[Linzer]

if the KDE guys make their environment behave better when a KDE app is loaded from some other window manager, maybe i'll give it another shot.

Now, this is a problem right there. If you load a KDE app in a different window manager, then it isn't "their environment" anymore. Making KDE apps behave nicely in other environments is definitely not KDE people's priority. The same goes for GNOME, to be honest: I've had some unpleasant time trying to setup applications with GConf while not using GNOME.

While I am can totally understand your case and sympathize with it, I think you're right that KDE is not for you. It's meant to be much more exclusive. It's designed for KDE users, if you will (no irony intended).

Re:Where are the GUI designers going to realise...

[Microlith]

Good, stick with your mid-90s and earlier window manager. The rest of us will enjoy the capabilities afforded us by our hardware.

Giving KDE a new chance.

[Balinares]

Maybe it's time to be cautiously optimistic again.

When Unity came out, I gave it its 21 days[*]. After that time, I was still not very happy with it, so I figured that after using Gnome 2 for a while, it was time to give KDE another chance.

Well, I'm glad I did. There are still little niggles here and there if you look up close, but as a whole, things work pretty darn well. They've finally managed to return to that KDE sort of state from the 3.5 days, where multitudes of little features activate as needed to support your workflow and otherwise stay the fuck out of the way. Klipper is still so freaking convenient that I miss it sorely wherever I don't have it (the Gnome equivalent, Glipper, unfortunately didn't work very well for me). Also, Chromium now natively supports the KDE password storage thing. Quassel is like a smoother X-Chat with less bugs.

All in all I've been somewhat pleasantly surprised, and I think I may keep it after its 21 days. There are still things that annoy me -- their overthought Akonadi thing, for instance; seriously, guys, I shouldn't need an RDBMS to freaking read mails -- but much fewer so than I feared. Maybe it's time to be hopeful again for that Linux desktop thing we've been hearing about.

[*] When trying out a new tech, you've got to give it at least three weeks of real use, it is said; otherwise you can't necessarily tell if it sucks or if it's just different from what you're used to, and thus, uncomfortable at first.

Re:Giving KDE a new chance.

[sarhjinian]

I'd agree. Interestingly, I'm finding this with GNOME 3: it's surviving the "three week" test pretty well so far. I think it's the "interface gets the hell out of the way" factor, too: you end up working with apps and documents, not fussing with settings.**

The problem, if you can call it that, is that the distro of choice for GNOME3 (Fedora 15) makes it a little hard to get going out of the box. It's not by any means insurmountable, but it's a little harder than it should be as some things are missing entirely (an Office suite really ought to come preinstalled) and playing "find the repo/RPM" is a lot harder than "It's probably already there, and if not it's trivial to find a PPA" of Ubuntu.

I'm interested to see what, if anything, the Linux Mint folks will make of GNOME 3, and it's unfortunate that Ubuntu isn't going this route. It really is a good DE, and it would benefit from Canonical's (former, traditional) user interface polish.

** I find myself fussing with settings a lot in KDE, and more often than I'd like in Ubuntu 11.04.

Re:Giving KDE a new chance.

[Sipper]

I've been using KDE4 since about KDE 4.2.2, and have just upgraded to KDE 4.6.3. For the most part I really like it; it's got a long list of features and a large feature-rich set of applications, and I enjoy the Qt backend, too. There are three major exceptions to the things I like, and one minor exception.

Tthe major exceptions:
1) Strigi
2) Nepomuk
3) Akonadi

Strigi and Nepomuk are what turn most people off to KDE4, because these both cause performance problems. Strigi is a Desktop file indexer; Nepomuk deals with metatagging of files for tagging or rating. Strigi immediately wants to index files on the system as soon as you first log in, and that heavy immediate I/O load makes the first impressions of KDE4 to be poor. Nepomuk is a more consistent performance problem -- last I used it, even selecting a hundred files in Krusader took multiple seconds (and on my old Desktop, more like half a minute to a full minute). The system pretty much has no choice but to perform badly in file operations because inotify has to watch the entire system for file moves. None of these performance problems are discussed in the documentation. >:-| Discussing it with developers is extremely frustrating; it starts with "get used to it, Nepomuk and Strigi are not going away", and ending with dropping the job of documenting the performance problems onto some user's lap.

Thankfully, both Strigi file indexing and Nepomuk metatagging can be disabled within the "Desktop Search" settings. That fixes the performance problems most of the time. Last I checked, disabling these does not clear out the database, though -- on my Desktop system the Virtuoso SQL database grew to be > 2 GB as stored on the filesystem, and I had to clear that out manually. (And as I mentioned, performance was abysmal.)

Akonadi is the storage for personal information (names, email addresses, phone numbers, etc) which uses SQL for storage: and by default it wants to use MySQL server for this. Using a 30 MB MySQL server instance required just in order to hold a few contacts is overkill. There are now several Akonadi SQL back-ends such as Sqlite3, however the configuration defaults to using MySQL and changing the back-end is done on an individual basis within each user's home directory (in ~/.config/akonadai/akonadaiconnectionrc) -- and as far as I know there is no way to change the system-wide default. >:-| That is not pleasing.

See: http://api.kde.org/kdesupport-api/kdesupport-apidocs/akonadi/html/classAkonadi_1_1DataStore.html

So the combination ends up by default being a large SQL database file index in Virtuoso for Strigi and Nepomuk, and a separate SQL database in MySQL for Akonadi.

The last minor issue I have with KDE4 have to do with plasmoid configurability, such as the clock, network monitoring, temperature monitoring, etc. There are always details I would like to see which I cannot get to be shown, such as being able to manually choose rendering font and font size, colors, and whether or not to show a number value as well as a graph. The plasmoid configuration options change between KDE4 versions, sometimes adding features and sometimes removing them. For instance, in the previous version of KDE4 the Digital Clock plasmoid was one of the few plasmoids that allowed changing not only the font used to display the clock, but also the font size... however now in KDE 4.6.3, the font size options have been removed for that plasmoid. ? Weird

I consider KDE4 a great achievement, and I enjoy using it. At the same time, I still don't consider KDE 4.6.3 to be on-par with where KDE 3.5 left off -- KDE 3.5 was a well-oiled machine, and KDE4 is getting close but isn't quite at that same level yet.

Re:Or stop fucking wasting space.

[Hultis]

Jevons paradox. I'll just leave that here and let you think about how it works with increasingly fast hardware, increasing hard drive space and the obvious parallell to increasing screen real estate.

Kde 4 is so last decade

[Billly Gates]

I want a new cell phone interface. One where key functionalty is removed and only one app can be shown at a time with strange mouse gestures that take up the whole screen to shuffle between apps with no buttons focused on single tasking.

Re:Or stop fucking wasting space.

[fnj]

He may be a Gnome guy and simply doesn't realize that the KDE structure isn't designed by totalitarian bastards who KNOW what is BEST FOR YOU and damn sure won't be caught dead giving you the CONTROLS to actually TUNE IT.

Re:Or stop fucking wasting space.

[walshy007]

For crying out loud, look at the goddamn KDE 4.7 beta 1 screenshots in the article! LOOK AT HOW MUCH WASTED SPACE THERE IS! In the screenshot of Dolphin, look at how shitting massive the icons are! If they were half the size, you could get twice as many shown at once, and still be able to see the thumbnail image just fine.

Everyone that has been using kde since the pre-dolphin era uses konqueror for their local file storage browsing, dolphin is horrendous in comparison.

Those of us that use kde day to day likely don't encounter most of the suckier new items, simply because we keep on doing it the older way. (most useful aspect of konqueror imho, browsing sftp like it were local and copy/pasting etc like normal)

Re:Where are the GUI designers going to realise...

[Artifakt]

If you can really draw a bow with 170 lbs of pull, then it doesn't really matter if you miss once or twice, unless the opponent is starting out at much less than your normal working ranges for a bow.:
1. Even a limb hit will probably result in instant crippling at that pull. Bows like that will shoot through cinderblocks. At ranges of less than 100 yards, you aren't really doing archery any more, because the arrow doesn't perceptibly arch, it's as straight a shot as with a rifle.
2. You can always rip the maniac in half with your bare hands if he gets too close.

Re:Where are the GUI designers going to realise...

[muuh-gnu]

> Good, stick with your mid-90s and earlier window manager.

Why changing something that works, for the worse?

> The rest of us will enjoy the capabilities afforded us by our hardware.

I wouldnt mind if all those fabulos capabilities you allegedly "enjoy", were for the better, but they arent. They are mostly a superficial, never ending designer circlejerkoff in fight for winning the useless "oh shiny, now wheres my starbucks app" crowd. But you cant win that crowd for more than a year, because they change trends faster then you change underwear. You fool yourself by thinking that every time you completely jettison the old, working configuration, for a completely new design, you are improving something, but you arent. Youve just entered the fashion zone, without realizing it.

Re:steady as she goes

[Osgeld]

no it didnt fuckwit, it created a widget, now widgets and icons do not play by the same rule so if you dare have a fucking document on there at the same time they will stack on top of eachother, if you mouse over one it has a box expand to 3 times the size of the icon and you have to dance around these fucking idiot wigets with little semitransparent boxes flashing all over the screen

god fucking damit thats so much better than a fucking icon, wouldnt you say so dipshit?

Re:steady as she goes

[SomeKDEUser]

Plus, you are full of crap. I just did that: dragged an icon from the menu to the desktop (not the folderview, the desktop) and it created the link. And in fact, there wasn't even a box.

The box thing, I remember from KDE 4.0-4.1. Two years ago. So there you are, insulting people spending countless hours of their lives to give a better desktop to the world, and you can't even fact-check.