Also, there is very little different between the amount of research, usability study, design and development that went into OS X as went into KDE 4, Gnome Shell, Unity or Windows 7

Completely unbased claim. Win7 has desktop similar to any previous version of Windows with some features added (except the reworked menu for which there is already a third-party plug-in for restoring the old outlook), KDE4 is a perverted desktop, departing from previous efforts, Gnome Shell is not a desktop so far and in any way GS and Unity both employ a paradigm, designed for for netbooks, not desktop computers.

Anixx wrote:OS X is a completely different operating system. It was not a new version of Mac OS. Apple was in a very difficult situation: their OS was quickly becoming outdated. It did not support the new network, web, document formats. Most importantly, Microsoft canceled their shipment of MS Office and IE for Mac OS, which meant definite death of the OS (no browser, no office suite). In this situation Apple decided to employ the already existing open-source operating system on their hardware with its modern network stack, libraries and applications

Except that it was still a new version of Mac OS - rebuilt from the ground up (on top of Nextstep). OS X is Mac OS version 10 - all of their releases since have been numbered 10.x - version 10 because their previous version was version 9; called OS X because X is the roman numeral for 10 and it sounds cooler. Apple's continuing of the version numbering with OX X - Mac OS 10 - is no different from KDE's with KDE4.
Yes it was a completely new operating system, but it was still a Mac OS and used a number of the same concepts and designs that existed in previous versions of Mac OS, while at the same time introducing many new concepts and designs. KDE4 is no different.
KDE4 is a completely new code base and desktop experience, but it does still retain a lot of the same concepts and designs as existed in previous versions of KDE. Should we name it KDE IV to distinguish from previous versions of KDE the way Apple named it OS X rather than OS 10? I don't think that's really necessary and there are plenty of people who agree.
In most software, major number changes mean big changes - often times not backwards-compatible - which usually means you can't upgrade in place. You need to make changes (either in workflow habits, or in configuration, or in tools used with that software, or in using new APIs/recompiling, etc) and adapt to the new software. There are numerous examples of this: Python (2->3, and other languages that have undergone major revisions), Firefox, QT (3->4, one of the primary instigators of the change to KDE4), JBoss, GCC, etc. Projects realize that they need to go through major changes - throw away old code/features that can/need no longer be maintained, add new concepts and designs, refresh the code base, implement new paradigms - and they number these changes with a major version number change so that people know that it is something new and different, even though it is the same project.

Anixx wrote:

Also, there is very little different between the amount of research, usability study, design and development that went into OS X as went into KDE 4, Gnome Shell, Unity or Windows 7

Completely unbased claim. Win7 has desktop similar to any previous version of Windows with some features added (except the reworked menu for which there is already a third-party plug-in for restoring the old outlook), KDE4 is a perverted desktop, departing from previous efforts, Gnome Shell is not a desktop so far and in any way GS and Unity both employ a paradigm, designed for for netbooks, not desktop computers.

I didn't say they revamped anything in Windows 7 like they have in the other projects. I said that they put a lot of research, usability study, etc. into it. However, while it is largely the same interface, they made numerous tweaks and UI changes - the new taskbar, the compositing effects, snap, peek, libraries, etc. Compared to other Windows operating systems, they made some radical changes in Vista and 7, but they are plagued by a commitment to backwards-compatibility which means that radical changes don't really change much at all so these radical changes were nothing compared to the changes made in OS X or any of the other projects mentioned.

Except that it was still a new version of Mac OS - rebuilt from the ground up (on top of Nextstep).

No.

all of their releases since have been numbered 10.x

This does mean nothing, other than for marketing purposes.

but it was still a Mac OS

No. It is (Mac) OS X.

called OS X because X is the roman numeral for 10

X may stand for Unix/Posix as well. It is traditional ending for many Unix systems (QNX, AIX, Minix, Linux, Xenix, HP UX etc)

KDE4 is no different.

The difference I already pointed out: Apple had no choice other than to switch the operating system they used because of Microsoft's policy. KDE team had no such necessity.

but it does still retain a lot of the same concepts and designs as existed in previous versions of KDE.

It ratains less concepts and designs common with KDE3 than say, LXDE or Gnome2.

I said that they put a lot of research, usability study, etc. into it.

You so often repeat this that I have no choice other than to ask to point me to the research results.

Compared to other Windows operating systems, they made some radical changes in Vista and 7

Most of their changes are just new features which can be switched off.

but they are plagued by a commitment to backwards-compatibility

If Linux was not plagued by vandalist's spirit of breaking everything and disrespect towards other's labor, Linux already would be the OS of the majority and Qt would be the standard toolkit for nearly any Linux applications.

Windows retains the old interface not because it is plagued by "backwards compatibility" but because it was a success and the users like it. Theoretically they can completely rework the interface without breaking any compatibility nor API. Even the changes they made in Vista attracted so much criticism that this release is considered a fail and the majority of users still use XP.

No. It is (Mac) OS X...

...X may stand for Unix/Posix as well. It is traditional ending for many Unix systems (QNX, AIX, Minix, Linux, Xenix, HP UX etc)

Mac OS X is the version of Mac OS and a continuation of 7,8,9..Otherwise it would be called Mac OS X 1.x instead of 10,x.

The difference I already pointed out: Apple had no choice other than to switch the operating system they used because of Microsoft's policy. KDE team had no such necessity.

In order to take advantage of Qt4 technologies and the new license, a rewrite was needed. If/when Trinity is finished being ported to Qt4, its potential won't be half that of KDE 4's.

Most of their (Microsoft's) changes are just new features which can be switched off.... Windows retains the old interface not because it is plagued by "backwards compatibility" but because it was a success and the users like it. Theoretically they can completely rework the interface without breaking any compatibility nor API. Even the changes they made in Vista attracted so much criticism that this release is considered a fail and the majority of users still use XP

Windows 3.x to 4.x was a major change in they way people interacted with their computers - much more drastic than both Mac OS 9 to OS X and KDE 3.x to KDE SC 4.x. and you can't just turn it off. Now, most people would probably rather eat charcoal than to go back to the 3.x interface.

Another major change was Windows 4.x to Windows 5.x (Me to XP) which can also be considered a rewrite as it was a completely different OS to begin with - and that did break backward compatibility. I had games that wouldn't work even with the compatibility layer enabled.

Mac OS X is the version of Mac OS and a continuation of 7,8,9..Otherwise it would be called Mac OS X 1.x instead of 10,x.

How to call it decided the marketing department. This does not affect the fact that it is another operating system (even belonging to a different family of operating systems).

In order to take advantage of Qt4 technologies

What is the advantage?

and the new license,

Qt3 has the same license as Qt4. And not anything in KDE was rewritten thus no benefit in legal terms.

If/when Trinity is finished being ported to Qt4, its potential won't be half that of KDE 4's.

How did you decided that?

Windows 3.x to 4.x was a major change in they way people interacted with their computers - much more drastic than both Mac OS 9 to OS X and KDE 3.x to KDE SC 4.x. and you can't just turn it off.

As you know, any Windows version after Win 3.1 up to Vista offered desktop of the Win 3.1 style with Program Manager and Winfile. The modern Windows versions (at least XP) can run such old applications as those from Windows 1

Another major change was Windows 4.x to Windows 5.x (Me to XP)

You meant 9x to NT?

and that did break backward compatibility.

For the majority of applications it did not break the backwards compatibility (and as you know there is a compatibility mode in Windows). The core of Windows API (Win32k) is essentially the same for 18 years since 1993.

Linux is not for users so the whole question is moot. If it were, the development strategy would look quite different. There is no agreed upon stable platform on which bugs are fixed. Rather, there is a continuous effort to add new features whether it be to the kernel, graphics system, ui and so on. New versions of these packages are then released with seemingly limited testing. Interest in correcting problems in former releases is largely lost. In the end, potential users don't have any way of determining if linux will work with their hardware except to give it a try. Installing a new OS under such conditions is usually a tremendous waste of time.

Qt3 has the same license as Qt4. And not anything in KDE was rewritten thus no benefit in legal terms.

The licensing was different on the platform specific versions of Qt. Qt3 was only offered under the GPL for Qt/X11, Qt/Mac and Qt/Embedded. Qt 4 became available under the GPL on all supported platforms including Qt/Windows. This allowed KDE to be ported to all platforms supported by Qt without having to buy a commercial license.

What is the advantage?

Running cross-platform natively using the same code-base.
This is how I understand it, KDE has to be abstracted from the underlying operating systems in order for it and KDE apps to be portable to other operating systems and run natively - allowing KDE developers and third party developers to work with a single code base for all platforms. From what I've read, this would hardly be possible, if at all, to do by merely porting KDE 3 to Qt4 without a major rewrite to take advantage of all Qt4 has to offer.

As you know, any Windows version after Win 3.1 up to Vista offered desktop of the Win 3.1 style with Program Manager and Winfile

No, I didn't know that. They must have hidden that option very well or I just never noticed it.

You meant 9x to NT?

Yes

cmrntnnr wrote:Linux is not for users so the whole question is moot. If it were, the development strategy would look quite different. There is no agreed upon stable platform on which bugs are fixed. Rather, there is a continuous effort to add new features whether it be to the kernel, graphics system, ui and so on. New versions of these packages are then released with seemingly limited testing. Interest in correcting problems in former releases is largely lost. In the end, potential users don't have any way of determining if linux will work with their hardware except to give it a try. Installing a new OS under such conditions is usually a tremendous waste of time.

The real problem isn't the developers but that most users want the latest and greatest, which is understandable as humans are curious creatures.

There are rolling releases, semi-rolling releases, biannual releases, long term support releases and other stable releases to choose from. Therefore, you shouldn't really make a blanket statement that minimal testing goes into new versions of apps. It depends on the user's distro as to how much testing goes into it and if/when and which version should be included in the repos.

Debian goes two years between stable releases. This gives them plenty of time to fix bugs as the packages trickle down from the unstable branch to the testing branch and finally make it into the stable branch. But, by the time the next stable release comes out, you'll be using outdated software.

I use a semi-rolling release. Pardus is still using KDE 4.5.5 but it also contains numerous backports and fixes which improve the stability and user experience significantly. It "feels" like I'm running 4.6 but KDE 4.6 won't be rolled out until it's been tested to Pardus's satisfaction.

cmrntnnr wrote:Linux is not for users so the whole question is moot. If it were, the development strategy would look quite different. There is no agreed upon stable platform on which bugs are fixed. Rather, there is a continuous effort to add new features whether it be to the kernel, graphics system, ui and so on. New versions of these packages are then released with seemingly limited testing. Interest in correcting problems in former releases is largely lost. In the end, potential users don't have any way of determining if linux will work with their hardware except to give it a try. Installing a new OS under such conditions is usually a tremendous waste of time.

Please remember that most work is being done by enthusiasts, most of them have regular jobs or are studying.

Since most distributions have live CD:s an install isn't needed to determine if ones hardware is supported, a test run from a live session will tell you that. That is something you don't get with Windows for example, plug and pray springs to mind .

This allowed KDE to be ported to all platforms supported by Qt without having to buy a commercial license.

So the reason for total rewrite is to have another license?

By the way, you're incorrect, KDE4 still has parts of KDE3 (such as Konqueror).

And of course something published under GPL can be ported to any platform without GPL violation.

Anixx wrote:

This allowed KDE to be ported to all platforms supported by Qt without having to buy a commercial license.

So the reason for total rewrite is to have another license?

Not the only reason but part of the reason for the rewrite was to be able to take advantage of the technology and the new licensing.

By the way, you're incorrect, KDE4 still has parts of KDE3 (such as Konqueror).

I don't remember saying that. You probably inferred it from something else I said.

And of course something published under GPL can be ported to any platform without GPL violation.

But the native versions such as Qt/Windows couldn't because there were no GPL versions of them then. Yes Qt/X11 was ported to Windows but that's not at all the same thing as the native GPL'ed Qt/windows.

Why do I feel like I'm repeating myself?

part of the reason for the rewrite was to be able to take advantage of the technology and the new licensing.

This has no sense because parts of KDE3 still in KDE4 as I said.

But the native versions such as Qt/Windows couldn't because there were no GPL versions of them then. Yes Qt/X11 was ported to Windows but that's not at all the same thing as the native GPL'ed Qt/windows.

If you mean that to have advantage of Qt4 on Windows KDE should be ported to Qt4, I can agree but this does not require total rewrite.

Anixx wrote:

part of the reason for the rewrite was to be able to take advantage of the technology and the new licensing.

This has no sense because parts of KDE3 still in KDE4 as I said.

Yes, there are applications that were in KDE3 that are still in KDE4, however nothing moved from KDE3 to KDE4 without code changes - either because of the port from Qt3 to Qt4 or the changes in the KDE libraries.

Anixx wrote:

But the native versions such as Qt/Windows couldn't because there were no GPL versions of them then. Yes Qt/X11 was ported to Windows but that's not at all the same thing as the native GPL'ed Qt/windows.

If you mean that to have advantage of Qt4 on Windows KDE should be ported to Qt4, I can agree but this does not require total rewrite.

Portability isn't the only advantage of Qt4. The toolkit itself underwent a major overhaul and broke source and binary compatibility with Qt3 - meaning projects wishing to upgrade from Qt3 to Qt4 needed to make code changes as part of the upgrade.
Porting to Qt4 isn't the only reason for the massive rewrite which was KDE4. These reasons have been articulated a few times in the forums as well as in various blog posts. One of the big reasons was the unwieldy codebase which made adding seemingly trivial features a monstrous undertaking, and implementing new designs and concepts simply infeasible. That was actually a more important consideration in the decision to rewrite everything for KDE4 - it would be easier to rewrite everything than to try and refactor the existing code.

however nothing moved from KDE3 to KDE4 without code changes

And what? How does it affect licensing?

One of the big reasons was the unwieldy codebase which made adding seemingly trivial features a monstrous undertaking, and implementing new designs and concepts simply infeasible. That was actually a more important consideration in the decision to rewrite everything for KDE4 - it would be easier to rewrite everything than to try and refactor the existing code.

So the code was rewritten just because the devs wanted to change the desktop's concept.

Anixx wrote:

part of the reason for the rewrite was to be able to take advantage of the technology and the new licensing.

This has no sense because parts of KDE3 still in KDE4 as I said.

But the native versions such as Qt/Windows couldn't because there were no GPL versions of them then. Yes Qt/X11 was ported to Windows but that's not at all the same thing as the native GPL'ed Qt/windows.

If you mean that to have advantage of Qt4 on Windows KDE should be ported to Qt4, I can agree but this does not require total rewrite.

KDE 3 uses its own libraries and frameworks along with Qt's. Simply porting it to Qt4 won't take care of the portability problems all by itself as KDE 3 was not written to be cross-platform. This is why KDE 4 introduced new frameworks such as Solid, Akonadi, Phonon (which was shifted to upstream Qt), and most importantly, in my opinion, Plasma as it does not dependent on X or any single windowing system.


As you know, the KDE 3 desktop is made up of the Kdesktop shell, Kicker, and Superkaramba . These components are heavily dependent on the X11 windowing system because KDE was only released on the X11 based systems. These three desktop components of KDE 3 need to be replaced or rewritten in order to allow KDE desktop components to run natively on other systems.

KDE3/Qt4 would still require X11, the native themes would have to be emulated, it won't integrate very well e.g. icons not showing up in system trays or docks, and you won't be able to maintain a single code base for all platforms. Platform integration won't happen at the library level, which will lead to maintaining several source trees.

As Airdrik and I have said, portability wasn't the only reason for the rewrite but they had to do a rewrite to take advantage of Qt4 anyway. So while they were at it, they chucked all the unwieldy code that had accumulated and replaced it with a much cleaner and more flexible slate.

So the code was rewritten just because the devs wanted to change the desktop's concept.

Now you're being purposely obtuse. It's very clear what we are saying here. There were many reasons which have been discussed ad nauseum here.

Dead is just an expression. Even if 1% of the world is using linux as a desktop, that means 67,000,000 people are using linux. That's enough. I think.