Slashdot Mirror
What's Coming In KDE 4.4
buzzboy writes "If you're wondering what the folks over at KDE have been cooking up for the next major release, KDE 4.4, well, quite a bit as it turns out. In a lengthy interview, KDE core developer and spokesperson for the project Sebastian Kugler details the myriad changes that are coming with the 4.4 release — the fifth major release since KDE 4.0 debuted to much criticism nearly two years ago. The project has closed about 18,000 bugs over the past six months and the pace of development is snowballing. The 'heavy-lifting' in libraries and frameworks for 4.0 is now starting to pay off. Perhaps the biggest change is in the development of a semantic desktop. According to Kugler, 'If you tag an image in your image viewer, the tag becomes visible in your desktop search. That's how it should be, right?' There is also a picture gallery of KDE 4.4 (svn) screenshots so you can see what it will look like."
I work on the "System Activity" thing (pops up if you press ctrl-esc. Like Task Manager). It's hard to get feedback about it.
So if you're a KDE user and use this, let me know what you think, how you find it, suggest any improvements/features etc. UI designers, code documenters etc also welcome to give feedback :-)
I often see people posting about how KDE/Gnome never listen to UI designers, Usability people, etc. But I've personally never had any feedback or bug reports about that sort of thing, ever. So do feel free to file such bugs - us developers are listening.
they'll run out of version numbers in the 4.x series before the series reaches its full potential. I'm really looking forward to using 4.4 but, since it will be the first release that really starts developing the ideas that KDE wanted to implement in the 4 series, the .4 increment seems a bit high. Still, 4.3 already does what Windows 7 and OSX only hint at moving towards so 4.4 will be interesting.
So if this is the future...where's my jet pack?
Also, the semantic desktop concept is shaping up nicely. I was weary of enabling nepomuksearch with strigi, because in the early 4.x releases they were extremely buggy. Then I went ahead with 4.3.3 (on Arch), and now strigi seem to work fine. It uses minimal resources, indexing is automatically switched off when you switch to powersaving mode (useful on a laptop), otherwise CPU usage is barely noticable. It still uses a shitload of memory, but with KDE 4.x you have plenty to spare. I have 2 Gb in my laptop, and without nepomuk/strigi memory usage after startup is 15%. That includes all the daemons necessary for a modern desktop (including cups), 2 desktops with different wallpapers and widgets, wicd. After running it for days without reboot, memory usage stabilized around 30% including ktorrent running in the background. After I started using nepomuk, that number icreased by around 20% - still pretty lean considering what it does. Which reminds me, nepomuk (on my setting at least) works in dolphin (just start typing in the searchbar), not in the normal Find files option accessible from KMenu.
Barring the advent of AI, or at least uncannily clever automated systems, there really isn't much alternative to manual tagging(outside of a bunch of specific, though admittedly useful, special cases like facial recognition tagging for images, or origin tagging that makes it easy to distinguish between "files I received as email attachments" and "files I downloaded from the web" and the like).
In my (admittedly lay) opinion, what makes "semantic desktop" 'desktop' is the fact that there is some consideration given to making tagging data meaningfully useful across applications and the desktop environment, rather than just inside one particular application(as has historically been the case with MP3 player programs, photo organizers, and similar). Demanding that it solve this problem and somehow automatically tag the untagged seems unfairly dismissive of a useful incremental step.
What would be helpful, in terms of increasing the amount of already tagged stuff; but would open up a giant can of worms, partly in terms of legacy dealings, but mostly in terms of inadvertent information leakage or malicious tag pollution, would be making the transfer of tags along with the file they describe from system to system easy. ID3 tagging, for instance, is largely manual; but works quite well because, for any given MP3, the odds are pretty decent that somebody has bothered to do the legwork properly and the tags migrate with the file(because they are inside it) and files with proper tags are considered more desirable than those without. Were that state of affairs more general, there would be a lot more tagged files without much more work by any particular individual. Of course, given that many tags are far more sensitive or subjective than ID3 tags tend to be, some rather deep thinking would be in order to make that actually work properly.
GNOME 3 isn't supposed to be out until September 2010. That's quite a long time for a software project. And once it's released, it'll suffer from the problems that the early KDE 4.x releases did. A barely-stable release won't be available until September 2011 at the earliest, but more likely by September 2012.
Oh please. The new gnome-shell is already usable and mostly stable. Just because the KDE team made their transition into a new release a massive trainwreck doesn't mean every software project will follow in their footsteps.
Gnome hasn't stagnated, it's a mature and stable desktop environment. Because of this, Gnome is often the preferred choice for enterprise desktops. KDE is obsessed with shiny objects and web 2.0 garbage, which is why it has taken so very long for the 4.X series to be merely usable.
I'm not claiming that KDE is down for the count, just that the desktop environments seem to have taken a different development strategy as of late.
4.2 wasn't bad, and I actually *like* 4.3, I can easily set it up to do what I want/need easily.
I *hate* it. At least the 3D stuff can be turned off now, but there's still a noticeable lag with keyboard input randomly. I mean, seriously. I have a 2x2,4 GHz processor and you tell me you can't display the key I pressed under 0,1 seconds?
Oh, and please don't try to find and animate every possible program on the run dialog until I actually finished typing the relevant part.
I'm still waiting for the damn site to load, so lets all just read the KDE 4.4 Feature Plan instead.
# cat
Damn, my RAM is full of llamas.
4.2 wasn't bad, and I actually *like* 4.3, I can easily set it up to do what I want/need easily.
My only worry is that... with 4.4 out, are we going to be subjected to KDE5.0 soon?
I called it some time ago
4.4 = all 1st party tools pretty much finished, 3rd party tools there but not polished
4.5 = 3rd party tools good to go
then somebody will release dbus/kross/plasma malware and they will realise that the whole DE has to be redone from a security perspective! ....
5.0 = an entire re-write with some concept of security and threading (kross for example runs in the same thread as the parent app)
5.1
6.0 = port to qt 5?
IranAir Flight 655 never forget!
I DO NOT want semantic tags. The reason is simple: they are LOST when you copy or do anything with the files. If you have important info about the file: put it in the filename. Or inside the file (exif tags for images, ID3 tags for mp3s, etc). Or in a txt file with the same name next to it. The rest is no better than putting varnish on a turd: it works only as long as you don't get too close.
Non-Linux Penguins ?
Wrong. I built and installed GNOME 3 this past weekend, and it's nowhere near being comparable to even KDE 4.0. It's nowhere near as stable as you're making it out to be.
And you're wrong about it being the "preferred choice for enterprise desktops". That had nothing to do with quality or usability. That just had to do with it having a more corporate-friendly licensing scheme than KDE did, mainly because of Qt. So vendors like Sun and HP tried to adopt it. But since Nokia's acquisition of Trolltech, and their subsequent relicensing of Qt, that's no longer an issue.
And finally, you're wrong yet again about the direction KDE is taking. It's not about "shiny objects" or "Web 2.0", it's about them making better use of the hardware resources of modern systems. While some demos have been put together showing OpenGL-accelerated effects, the main benefit has been with the desktop being more responsive as a whole, due to better hardware acceleration.
KDE 4.3 is extremely responsive, while GNOME 2.28 on the same system is clearly less responsive.
Maybe we should devise a method to attach arbitrary metadata directly to a file in a way that requires little to no modification of existing unix programs while also making it accessable to any program that can deal with files and directories.
GNOME has stagnated, and is on its deathbed. I mean, they're at release 2.28 for fuck's sake, and they only release every March and September. GNOME 2.0 was released in 2002! That was nearly 10 years ago! It's not an active project. It's just barely maintained.
Ubuntu crowd would disagree. Project can only be declared dead if it has no users.
Slow release cycles are fine - as long as software delivers what users do expect. Though I'm personally Gnome hater, some folks are pretty happy about Ubuntu and its default UI.
On the other hand, the KDE devs have shown that they're willing to innovate and provide a much richer and useful desktop environment. Even if there are bumps along the way, it's clear that they're lightyears ahead of GNOME. And at this point, it doesn't look like GNOME has any chance of catching up.
"Much richer" in what way? And how do you define "useful desktop environment?"
Those are silly questions to ask of "desktop environment" whose sole purpose in life is to allow its user to browse for a file and to start application. Applications are the meat of the desktop, not the desktop environment.
N.B. Likewise most of early complains about KDE4 were "where the hell the applications for it?" - KDE4 might have been ready already in KDE 4.0 times, yet applications using KDE 4.0 were few and unstable.
All hope abandon ye who enter here.
Not really a "distro" problem for me as I'm a FreeBSD user. I chose to install 3.x and 4.x simultaneously.
After putting a lot of effort into 4.0 for a week, I said "fuck it", and went back to 3. The same happened with 4.1.
I missed 4.2, and ended up with 4.3 on an Ubuntu Live CD I was experimenting with. My first thought was "Wow, they did some nice tweaks to this to make it play nice with Ubuntu. I wonder what it's like on FreeBSD?"
I went back and installed it on FreeBSD and it was just as nice as it was on Ubuntu.
I went back and found some 4.2 releases, and they didn't seem so bad either. My old 4.1 release still wasn't pleasant though.
Self proclaimed typo king, and inventor of the bear destroying coffee table (patent not pending).
Well, I dunno. Training some kind of Bayesian algorithm seems to work well enough for spam filtering.
I think the real problem is knowing the significance of some piece of information to us will be *in the future*. As user interfaces become more "semantic", I doubt they will be as usable as a stable way of organizing data.
We can take a lesson from physical filing systems. There are really only three methods of file organization that make sense: alphabetical, chronological and by physical size (this sucker won't fit in the cabinet). I once worked for a guy who insisted on organizing the company files by category: Hot, medium, cool, cold, interested in product X, interested in product Y, vendors of A, vendors of B, related to project M. The problem is that these categories aren't stable or mutually exclusive. Some files would languish in "Hot" long after the prospects lost interest. Sometimes a cold prospect would become hot, but its file wasn't in cold because it fell into some other category. Every time you had to file something, you were faced with dilemma as to which of the many possible categories might apply.
What was worse is that he would come in on weekends and reorganize the files. Accounting went *nuts* because invoices would disappear from files and they had to guess where he might have put it. This demonstrates that the semantic significance of an artifact depends on its context (e.g., sales vs. accounts receivable, or current self vs. future self).
One function of a record keeping system is to impose some stability on information in a world that is constantly changing world. Semantic metadata is most welcome, but it's no good as a stable organizing principle for information.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
I'm using Kate all the time and I have mostly 10 to 20 or more documents open. A tab panel on the side is really good with the 16:9 screens that now everybody has anyway. But if you think it's wasted space, why you don't click on "Documents" to hide it?
Kate is one of the very best text editors. Big thanks to the KDE4 developers.
The vertical text is very space save, too (which is good). In Kile there is like 4 panels to open with the vertical text, very use full. But what I don't understand is, why can't I move the panels to a different location, like to the sides. In Kile is the output panel on the bottom side, but I wish I could move it to the right side.
http://www.mueller-public.de - My site http://www.anr-institute.com/ - Advanced Natural Research Institute
``I'm a happy KDE 3.5 user. They screwed KDE 4 BY DESIGN, so I suppose the bugs they'll fix now again are just some bugs of fancy unneeded things, and not UI problems with important base components ... ''
With so many people expressing that sentiment, I'm curious: has KDE 3 seen any significant development since the release of KDE 4? You'd think there would be a lot of people even among the developers who thought they way KDE 4 was handles was a disaster, and continue to improve on KDE 3 instead (perhaps in a differently named fork). Has this actually happened?
Please correct me if I got my facts wrong.
See, nobody forces you to update.
Sometimes, in the life of an app, the devs find that it is not possible to better it without making some radical changes. This only means that the older version becomes the older version instead of being the current version.
It stays the same app.
For the developper, this is the occasion to correct major design mistakes, rethink how the apps should work, etc. This is crucial so the set of functionnality of the app continues to live.
Of course some users are unhappy.
But I would never give up my KDE 4.4svn for the latest KDE3. It is _so_ much better. Alt-F2 was great, it is now insanely great (unit conversion on the fly? an actual calculator? remote controlling amarok and the volume? live search of wikipedia?).
I have been a KDE user since 1.0. And never a transition was so ambitious, and yielded such insanely great results. And yes, never a transition had been so painful. But even in the 4.0 days, it was clear that it would be worth it. From the start, it was clear that it would be more integrated, more beautiful, and yes, more powerful.
Simply, it took time to add the features back, but this time around, not only do you get the features, but they also are not a drag on the devs, who can then bring their apps to the next level.
Is it perfect yet? No. Is it the best DE bar none? yes. (Sorry mac-addicts).
Which is exactly why there should be ALPHA_XX, BETA_XX and PROD labels. We use them for our releases, and they tell my customers what to expect. Alpha means I'm developing functionality and you should only use it to test what I'm giving you, not to support your testing. Beta means that I want you to exercise this and it should be looked at as feature complete but there will be bugs, Prod means that we've tested all spec'd functionality to the level of confidence required by our customers.
Open source apps (and I'm a F/LOSS advocate) tend to have devs that treat their users like crap when it comes to this type of simple to implement, easy communication. I shouldn't have to dig into your dev or user mailing list to determine if I should pick up a version. That's what labels are FOR.
My Babylon
I actually liked 4.2 too. The only thing I kept "back" from KDE3 was Amarok (which is semi-independent of KDE releases, but somewhat connected.) Amarok 2.0 was awful - horrible error handling for streams, and inconsistent collection performance (although much of that is likely just crappy Gentoo packaging.)
Gentoo continuing to have KDE 4.x masked as unstable even well into the KDE4.3 release cycle is why all of my new machines run Kubuntu, and I have even started doing reinstalls of machines with existing Gentoo installs. It's just ridiculous that the only KDE release marked stable by a distro (3.5.x) has been unsupported by upstream for months.
KDE 4.3 on Kubuntu systems is excellent, I LOVE it.
retrorocket.o not found, launch anyway?
GNOME has stagnated, and is on its deathbed. I mean, they're at release 2.28 for fuck's sake, and they only release every March and September. GNOME 2.0 was released in 2002! That was nearly 10 years ago! It's not an active project. It's just barely maintained.
Are you seriously making your argument from a version number? I wonder what you'd say about Emacs, then...
(By the way, just to remind everyone, GNOME uses the oldschool versioning scheme with even numbers for stable releases, and odd ones for betas; so 2.28 is the 14th stable release of 2.x branch, not 28th).
In any case, the claim that GNOME "has stagnated" and "is barely maintained" is trivially debunked by looking at overview of changes for every release. There are definitely fewer of them than there used to be, but there are still quite a few; and, on the whole, I find GNOME desktop today to be much more thought out and polished compared to either version of KDE, without looking dated. I would imagine that developers similarly appreciate API stability.
Ultimately, you've got to wonder why most distros today, especially "enterprise" commercialized ones, go with GNOME, and have been doing that for several years now.
Windows 2000 was famously released with 65,000 bugs (or items, at least) still in its bug tracker: http://slashdot.org/articles/00/02/11/1840225.shtml
And yet it's almost universally considered one of the best Windows releases ever.
I'm not saying anything about how buggy KDE 4.3 is, but simply counting the number of bugs doesn't really indicate much.
Comment of the year
What does Gnome not do that you want it to?
People always complain about this, but what are these awesome features I should be using that Gnome won't allow? I'd honestly love to know, 'cuz there must be all kinds of great workflow-enhancing, appearance-improving shit out there that I just don't know about and that Gnome won't permit me to use.
The only time I've run in to this was the spatial folder view crap, but that was trivial to turn off even then and IIRC they saw the light very soon after and made it a checkbox in an options dialogue.
As far as I am concerned kate is the best text editor in unix today. It's syntax highlighting is far and away the best (the only thing that compares is scite, which is designed to demo the same editor class kate is based off), it does everything. I can open up stuff on an ftp server and edit it as if it were local, I can edit any text file with any extension and get correct highlighting, I can do all my building and testing in the terminal.
And, the loveley vertical text makes options easy to see and not space consuming. The only thing I wish is that gnome would get their act together and make something as complete as kate so I could have everything looking gtk.
Ze Atomic Device! It iz Ztolen!
although I've recently been tending towards gnome, I really love a lot about KDE. I just wish they would forget about certain features for now and focus on stability and quality every-day features.
Specifically the "semantic desktop" I've used kde for years and never used it. Why the hell would I waste time tagging all my files? I have a sensible directory hierarchy which works just fine. I never find myself spending hours searching for stuff on my computer, because I know where all the things I need are, because I use them all the time. If I didn't know where something was that would imply I never use it, in which case, why am I spending time to tag things I never use? Just in case I might need it?
What I do need is for firefox to pick up on my application preferences (what opens up a zip, etc), for drag and drop to be snappy and accurate and always work, for ark to not suck so hard, for my folderviews on my desktop to always be up to date, look good, not pile up icons in weird ways, etc, etc.
I like that kde is very forward thinking in their features, but sometimes I'd like them to live a little more in the present. If you had an awesome super-intelligent automatic tagger that would let me search with vague queries and get exactly what I want, that'd be great, but spending your time on a dressed up database that tracks all kinds of stuff I have to put in by hand is a waste of everybody's time.
Ze Atomic Device! It iz Ztolen!
+1, I'm to used to dealing with integrators as my customers that I forget the end user view of it sometimes.
My Babylon
No, the API wasn't NEW. For Python developers, and probably others, it was NON-EXISTENT for a long time.