Is KDE 4 The Desktop Answer?

I just read an article about this in ComputerWorld Australia. The article is an interview and talks about some of what will be new in KDE 4. Having used KDE for close to 10 years now, I am clearly a fan but I am not sure KDE 2, 3, 4 or 27 is the answer.

More accurately, I am not sure KDE, Gnome or whatever is the answer to getting everyone using Linux systems. Yes, KDE today is very nice from a user point of view and KDE 4 will be better. Also, each version of KDE "does more" for the developer so there is less work to do. I think there are two problems:

  1. Politics
  2. Intertia

The politics is obvious: the mainstream answer, years ago, was an IBM mainframe. More recently it has been Mircrosoft Windows. Good or bad, what is popular is generally the right political choice. The same is true with intertia.

What will get people to change is not having "something new that does the same thing" but having something people want which isn't available elsewhere. This is why each model year a new VCR had a new bell and whistle. Or each new PDA had more memory, more pixels in the display, color, more speed, ... My black and white HandEra PDA did what I wanted--it remembered some stuff for me and it would record audio. Today, my now somewhat outdated Palm Zire 72 does the same. It does a lot more but it was the current "bottom of the line" that had what I really wanted--the audio record feature.

Linux on the web server side has won because it offered lower cost than UNIX-based solutions and more reliability than Microsoft-based solutions. Other than the high end were there are vendors with integrated hardware/software solutions, the remaining non-Linux web servers can be attributed to politics or inertia. But, that is far from desktops.

On desktops, applications become the issue. When Linux started its movie career (the water in the movie The Titanic) there was a lot less desktop than there is today. It has since gone on to do the rendering for Shrek and others. The transition of the needed software to the Linux desktop was done because it was a wise investment--it saved a lot of hardware and software money.

The difference is that for a a general purpose desktop you have multiple vendors (including Microsoft) who offer applications to general consumers. If the consumers are buying Windows-based systems anyway, there is no reason to want to invest in creating a Linux-based version of their application. Why, for example, port Adobe Illustrator to Linux if it only means you will sell the same number of copies as you do now and have to support another platform?

In much of the world, desktops transitioning to Linux is moving faster than in the US. This is likely a result of less political influence, less inertia (because there is more desktop computing growth) and less money available. If these international desktops don't need Adobe Illustrator, for example, they are not going to encourage Adobe to take Linux seriously. On the other hand, most need an office suite and the advent of OpenOffice.org users is putting pressure on Microsoft. If the most popular "Windows-based office suite" was not a Microsoft product, I feel confident you would have seen a Linux version by now.

What about other applications? Call them niche applications. If, for example, a workstation is used for doing computer aided design (CAD) as well as normal office functions, just making a CAD application available for Linux could result in a lot of conversions. The "trick" is for the CAD vendor to offer something in or related to their Linux version that isn't available elsewhere. This is where vendors need to look at, for example, what KDE 4 will do that isn't being offered on other platforms. Or, as KDE is now cross-platform (Linux, MacOS and Windows), maybe the "trick" is that if a vendor has a KDE version of their application, that is all they need.

In any case, KDE 4 will be released sometime in 2007. I am betting that some applications vendors that are working on versions of their software to work with KDE 4. That is, vendors new to the Linux scene

fyl - Mon, 2006-12-18 10:22.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

GNOME 2.16

Already GNOME-2.16 is here to take care of all Linux's
DE needs.KDE 4 is another crap next to M$ Windows

Prakash Jose Kokkattu (not verified) - Sat, 2007-01-06 00:50.

Kde is not crappy. If you

Kde is not crappy. If you think that you're crappy
Kde was present on linux before gnome. Gnome is based on gtk that has lots of performance problems.
Kde is based on qt which is very elegant and fast.

Cyril (not verified) - Fri, 2007-09-07 10:53.

New Software Concept

This is not actually a "new".

One of the biggest frustrations many of us have with Windows software is the total inflexibility of it and it's monolythic nature. Software should be subservient to the user, not the other way around. These days the user is increasingly forced to go through unnecessary steps, and to conform to interfaces that are "carved in stone". The programmer determines what features will be included and what methods the user will conform to. Designes the menu system, speed keys, dialogs, etc. Personally this enrages me as a user...... Software should serve the user, not rule him.... This is probably the single most compelling aspect of Firefox / Thunderbird. The fact that plug in's are available to do most anything you want, and to allow you to configure it in many respects........ To in fact build from a basic framework the program I want........not necessarily what the programmer had in mind for me.
Apple Computer was working on this software paradigm under the name of "Opendoc" when Steve Jobs took control and scrapped the project. The idea was that a basic application..... let's say "page" was the beginning point. It would have the ability to do basic word processing, and contain the basic but configurable menu structure. The user would then purchase from various sources the building blocks for the functionality he wanted. It would be launched in whatever iteration you needed for a particular task so in effect you would custom configure a spreadsheet program for example with only those functions you actually needed and configured in the layout you found comfortable..... In another iteration perhaps you might use a set of graphical tools for photo editing, or a word processor configured for a particular project, and the bundled modules would be launched in each case together as the user determined was best suited for his or her needs.
This software paradigm would make Linux a very compelling environment for the power user, or for business user who wishes for example to provide an employee with the necessary and simple tools just for a specific job. A simple scripting language mated to a macro function which would automatically generate scripts which could then be edited.... as for example was used in Hypercard ....... but linked to actual productivity applications could greatly empower the user.

This is my somewhat radical vision of where Linux needs to go....... make it flexible, make it simple, make it powerful, and make software SUBMIT to the will of the user rather than the reverse.

Howard

Howard (not verified) - Wed, 2006-12-20 10:20.

Of course its the Applications

Intertia? => I assume that was a typo for inertia.

I've said this many times. If you can develop a secure "Windows application layer" that is seemless for end-users to install Windows apps on Linux (including games), then you WILL turn the tide on Microsoft. (About 2/3rds of their money come from Windows and their specific apps for Windows: Like MS Office, Outlook, games, etc).

And no. Wine, CrossOver, and Cedega aren't seamless...They're more like a lottery. Or more like rolling dice. (You won't know the problems you will encounter until you try your apps on them).

I guess the other alternative is wait for ReactOS (Windows clone under GPL)...Maybe development from that may lead to further improvements in Wine?

As for KDE4...I hope that does encourage the way for devs of popular apps to be ported to Linux. But I doubt it...As it depends on the marketshare. Bigger marketshare of a platform, higher probability of money to be made...More likely to spend money on development on that platform.

An - ymous (not verified) - M - , 2006-12-18 13:51.

Bullshite

It's not the apps, it's the marketing. Linux already *has* an equavalent of nearly everything that's on Windows or Mac OS X. (Major exceptions: games, (and I mean, top-tier, big name, blockbuster type games), niche market apps)

Why is Windows so prevalent? Well, it certainly isn't because it is so *loved*. Apple OS is "loved" by its users. Linux is "loved". But Windows dominates.

Why?

MARKETING.

Oh, and the inertia that comes from many years of successful marketing.

Oh, and "marketing" can be extended to include the Embrace, Extend, Extinguish philosophy of Windows, as EEE is certainly not a technical concept but rather a business concept.

When we start to see Linux ads -- cool, funny, hip ads, mind you -- on major television programs and pasted all over city buses, magazines, websites, taxicabs, billboards, etc., then we might see some uptick in the upatake of Linux as a consumer OS.

Ask Aunt Tille: "What is Linux?" and you are likely to get a dumb stare. Ask her "What is Microsoft?" or "What is Windows?" and she's likely to answer "Oh, its your computer!" or similar. Why?

MARKETING.

QED

JDS (not verified) - Tue, 2006-12-19 13:44.

I'll just have to support

I'll just have to support other answers made before me. Linux does not have all apps I need. I need Microcal Origin, I need Matlab, sometimes I need Mathcad (well, rearely, but still), Multisim or Microcap. Well, lets admit, Linux has almost nothing to offer in area of CAD - at least, it seems so. Although most of mentioned software has a Linux-clone (except Mathcad - still haven't heard of any clone), other don't seem to be very promising. Well, you can use them if you really want to, but you need some determination. Currently, Windows offers better software in means of it's capabilities. For example, SciLab just sucks when compared to Matlab. Although it has virtually everything Matlab has, I just could not bear it's "speed" when rendering a 3D plot of a huge array, not to mention the way it was rendered. Overall speed of SciLab is discouraging... So, I'm forced to use Windows (not because of MatLab in particular, but in general - it's easier for my nerves).

And don't forget hardware support! Sometimes it just refuses to work without some tweaking, if works at all.

And of course, the diversity is both a feature and a pain in the neck. First - you don't know where to start. Then - lots of incompatibilities and dependencies. Some standarts (LSB) sure might help, and I really hope it will. Well, I don't mind there are so many distros, but managing all the software packages by software vendors - I think, it's really unattractive for them. Why to bother because of a tiny market, which is more expensive to support (every distro needs it's own package - so fot Linux you need to maintain several versions of the same software instead of one) and is less profitable? At the moment Linux has everything for software development, for networking, for casual users (browse internet and check e-mail), but no games and lots of other things, some of which don't even have clones!

maksrules (not verified) - Sat, 2007-01-06 19:28.

Re: I'll just have to support

Why whining about Matlab? It is available on Linux. It is in fact probably developed on a sane platform (read: some UNIX/Linux), and then ported to Windows. Hello scientific computation people is the most heavier UNIX/Linux users in the buisness.

--
MatlabOnUnixIn92

An - ymous (not verified) - Sun, 2007-01-07 05:45.

No, it is the apps

Sorry, but the Linux scene is in a bit of a mess. No, I am not a Windows supporter. I am posting this from my Toshiba laptop, running Kubuntu Edgy. The sound does not work properly. The Nvidia drivers does not work properly. Whenever the kernel is upgraded, all the lengthy setup to get the NVidia drivers working, is down the tube, and it is back to downloading every dependency dev package to build the NVidia driver. And that usually breaks something.

And when I want to update Scribus or Inkscape, I have to install "experimental" packages, because the repository don't have the newest versions, and usually - it breaks something. And you should have expected it, because it was "experimental". But the Windows one, I can get the latest version of Scribus or Inkscape - and it just works. And this is applications started of for Linux!!! Linux killer apps, and they run better under Windows!!!

The shite is that it is the same old forking issues as with the propreitary UNIX. This one won't work on that one, this package will break that package, this one don't have the latest libraries for that app - so do without or live with the old version while Windows has the latest version - and this for a Linux app.

I have been running Mandrake, SuSE, RedHat, PCLinux, Ubuntu, Fedora, even SCO OpenLinux. What is clear is that you cannot just download a package and install it. There is different package formats. And sometimes you need to run alien to convert from RPM to deb files, and then it breaks something. Then some tell you you must use KLIK or Autopackage, and that breaks some other stuff.

Linux is not going to make it to the desktop if this situation continues. I can't event get the latest OpenOffice, Scribus, Inkscape, ALSA, NVidia drivers to run properly without breaking something somewhere. The state of apps in Linux - compatibility with Linux apps - is a mess. A complete and utter mess unlike we have witnessed before. We need - literally we need - one desktop distribution to rule them all. One package format to rule them all. Just so that we can get some sanity. Every second asshole thinks he can write a better package, make a better distribution, and causes just more of a mess. We should kill 95% of the distributions and get focus on making one, just one, work with the apps that we would like to use. Where the up to date packages actually works, without needing to download the kernel headers, the source package for half the libs in the distribution so that we can do a config, make install, just to see it fail yet again, or breaking something else.

As I say, I am posting this from a Kubuntu laptop, so I am not technically challenged, or against Linux. Just getting tired to update to see the bluetooth drivers suddenly dying, or OpenGL stopping, or having to revert to the NV driver because the kernel update rendered the graphics driver dead.

And no - Linux does not have the "equivalent" of Windows apps. And Wine is not there yet. Linux seem to be making one step forward, and three steps back with all the senseless diversity. You have myriads of desktop managers out there. Apps written and depending on some of them, so if you started out with GNOME or KDE, you will need to load the other one anyway to get all things running - hopefully.

So in short - desktop Linux is not happening. Get over it. If it must happen, then it needs to get this dependency hell and millions of forks and incompatibilities sorted out and fast - because every week we see a new distribution coming out, fixing something and breaking others - and just adding to the mess in general. We don't need Oracle Linux, or ten different distributions that need different packages each. We need a package format that will install once and work on all. The Linux Standards Base seems to be just a word - and no action. Lets put pressure on packagers and app developers. Don't make the package for Fedora, or Debian, or Ubuntu, or SuSE. Make it for LSB version X. And force the distributions to be compatible with LSB version X. And if it must, make a package for RPM and deb. But make them standard, the same, same libraries, dependencies, so that we can work of one codebase, not ten million.

TK (not verified) - Sat, 2007-01-06 04:48.

Latest and greatest

Respectable TK,

What you've described makes me wonder: why do you need to run the newest versions of programs the very moment their source code is released? And then expect said programs to run perfectly off the shelf on the distro of your choice? In fact, this way you're doing the job of distribution maintainers. If there is a new version of some package, they will try to compile and run it for you. If there are any problems, they will try to get rid of them. And you should get a nicely packaged program/library/whatever after a while. Just be patient.

And if you don't like to be patient? Try another approach - use the unstable version of your favourite distro ([K]Ubuntu Fawn, Debian Sid, etc). Or switch to another distro, like Gentoo or Arch, where packages are compiled on your machine and you are much less likely going to see dependency problems. If you're adventurous, there are plenty of distributions to satisfy your needs.

But probably you'd rather take the "easiest" approach. So you'd just pick some distro and expect all programs to run correctly on it. You've written you've encountered quite a bit of programs giving you headache in your specific configuration. But how many of those issues have you reported in the respective bug tracking systems? Using this magazine's comment system to complain is not the right thing to do if you really want the situation to improve. Remember, talk is cheap. You've already received a lot for free by using Linux. Maybe it's time to return the favor and join the side of people that improve Linux instead of complain?

And one more thing, there will never be only one "true" Linux as long as there is the source code accessible. Which can be modified, improved, compiled and distributed the way the people like (especially if it's licensed under GPL).

Regards,
SirYes

PS. And while you're still using Kubuntu, maybe it's time to [re-]read its Code of Conduct? You know, the on saying about being considerate, respectful and collaborative?

PS2. I have already contributed a number of bug reports, ideas, and feature requests (and some patches too) to several projects, like Firefox, OpenOffice, GNOME, GTK, ALSA, Ubuntu, Gentoo, Debian (to name the most well-known). I received a lot of positive feedback and most of the problems were already resolved. My overall experience was very positive, if one knows how to approach the right people the right way. And what about your experience?

SirYes (not verified) - Sat, 2007-01-06 21:08.

Respectable SirYes

Yes I submitted bugs to various apps - mostly for SuSE and Ubuntu.

Why do I need the latest package of an app? Well - for example - the Scribus that came with Kubuntu - had a horrible flaw with exporting to pdf. It crashed. The same with the Inkscape - it would sometimes refuse to open its own files - some error in the svg handling. Also - its layer management was horrible - and the new package fixed that beautifully - and the work I was doing really really needed me to work well with the layers.

Now what I am trying to say - is that I have been contributing to get things fixed. What becomes really really annoying - is when nearly everything seems to be broken in some way and needs fixing of some sort. Eventually you don't really know anymore what is broken and what is fixed, what broke what and what fixed what.

For example - fixing the latest Scribus updates some libs. Same with the other apps. Now fix the ALSA drivers - which eventually - no one seem to be able to fix at this point for the laptop Intel HDA Connexant chipset. All that works was to turn ACPI off on the kernel... But to try and fix the ALSA drivers, you need the source packages, the dev headers, lib headers etc etc etc. Suddenly a 1MB download to fix becomes 60MB in all the other dependencies. The error has been there since Ubuntu Dapper, and still no fix. Imagine that - I WANT to use Linux on my spanking new Toshiba Notebook and use GLX with the NVidia 79000GS chipset, listen to the music over the harman/kardon speakers, and work on my apps. Not possible. The sound is broken - since the start. Had the GLX / Compiz working - but first had to compile the NVidia drivers because it does not work with the kernel, and everytime the kernel is updated, the download must include the kernel headers, lib headers, I have to redo the whole setup thing, which always breaks the xorg.conf file which I have to edit by hand to get the drivers to work... etc etc etc.

Now.... respectable SirYes
It is not about how I behave or contribute - I do. It is not about how I hate MS, I do. It is not about my determination to get rid of Windows, I am trying to.

The question posed was if KDE4 is the answer. And the answer is flat out "NO".
At the moment dekstop Linux is not even an answer for experienced power users. We need to form a possy, storm the devs, and hold them to make their apps at least, first and foremost, compatible to a base LSB distribution. Make a distribution that NO-ONE distributes - a standard LSB baseline distribution. And then, make a RPM and deb package for that base distribution that works. And then keep the distributions accountable to AT LEAST work with that base package. If they want to tweak and hone the app to be faster, use newer faster lib, great. Put it in the repository... but first - let whatever comes out first run against a base LSB distribution, and make the distributions compatible with those packages.

Otherwise, forget about ever getting to the desktop. The mess is just to big and diverse to think it will actually impress people who are not interested in the techie mumbo jumbo. If people go out and big mouth waffle about how Linux is so competitive with Windows on the desktop, and I as a VERY experienced technical users, have difficulty in just trying to get the latest OpenOffice, Scribus and Inkscape - that is available for Windows to run on Linux, then don't even discuss this over anything more than a strong cup of coffee very early in the morning.

TK (not verified) - Sun, 2007-01-07 04:06.

Unfortunately, it's not just marketing

While I agree with your sentiment, it's not just marketing. There is a scarcity of really high quality applications in some very important fields. While Linux has thousands of applications, many of these applications don't offer a robust enough feature set to really be utilized in a professional setting.

Please do not misunderstand me, I utilize and love Linux, but the strongest Linux applications are relegated to a few areas: Softare Development and Server Management. In all other areas, the applications are still evolving. This then leads to a problem. When I need to do actual work, I find myself often having to flip over to the Windows partition of my hard drive. If I want to do some easy technical computing work utilizing MatLab, it's to Windows. The Linux version is finicky at best and often does not work with my SLED installation. If I want to work in CAD, it's off to Windows for the AutoDesk programs. If I want to do image refinement or layout, then I utilized Adobe products. For statistics, I use SPSS.

While you will rightly hollar that there are OpenSource equivalents of many of these programs and I might be able to do much of this work in Linux, you would be wrong. At this time, R, the GIMP, and OpenOffice don't quite offer me the seamless integration that I get from my Windows applications. Further, they tend to be less stable. I am therefore, much more productive when I utilize the equivlaent software in Windows. For the time being, I am willing to pay for those differences.

There is also a different problem. Many of the the OpenSource equivalents are pretty good, but the examples I have listed are the best applications available and many of them are not available for Linux. When they are (as in MatLab), they often don't work quite right. Until these particular problems are addressed, I don't expect Linux to be a serious desktop contender and it will remain my hobby. When it comes to a desktop operating system, it is about applications. The operating system should be stable and stay out of the way.

Rob Oakes (not verified) - Thu, 2006-12-21 21:41.