In this interview, Chris Schlaeger, Vice President R&D for SUSE/Novell and long time KDE contributer, talks about Novells desktop plans, OpenOffice.org Integration, Groupwise support in KDE’s instant messanger Kopete, the Kontact groupware client as well as the integration of the Gecko rendering engine into Konqueror. He also shares interesting opinions and ideas for the future of KDE.
Novell/SUSE’s Chris Schlaeger talks about KDE
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSAlert.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Great interview, and I like the fact that SuSE (and a lead KDE developer) acknowledges something that I’ve been saying about Konqueror for a long time:
Konqueror is fast and nicely integrated, but Mozilla renders more pages better.
I was getting sick of all the fanboys talking about how much better Konqueor was and blah blah this, and blah blah that.
With the integration of the Gecko engine, Konqueror really can render pages just as well as Mozilla, but it took integrating part of Mozilla to do it
what does this mean for safari? (when’s the last time safari/webcore synched with khtml?) or is webcore its own entity far removed from khtml now?
while i agree with him on toolkits (isn’t QT and GTK enough?) i think there is still a place for khtml and gecko.
I’m surprised to see Novell playing both sides like this. So basically Novell will be surporters, to some degree or another, 2 desktop environments, 4 email clients (Win32, Java, Evolution, and KDE’s), etc. That seems like a lot of work but I guess that’s ok. The Linux Desktop will be interesting to see.
Shawn, i will agree in that i prefer Mozilla’s Gecko engine for rendering web pages better, but i will have to say KDE is the best thing going on the Linux desktop for a full featured environment, i dont always run Linux under KDE but i do have it for when i do want it (nice choices) xfce4 is what i mostly run, my 71 year old mother thinks KDE is a better desktop than windows, mom is a Linux/KDE user now but she also prefers Mozilla’s suite (browser & email client) i think Konqueror’s file manager is top notch and i use it when i get tired of mc in a terminal (more choices)…
i thought i read that KDE was going to implement mozilla’s Gecko rendering engine for Konquerer browser, would like to see that…
so to sup up my post/reply i wont throw stones at KDE…
“The next version will be a Novell Linux Desktop which will be based on the SUSE LINUX Enterprise Server 9 code base and combines the best of SUSE LINUX Desktop and Ximian Desktop”
OK but will we be getting a Ximian Desktop in SUSE Personal/Pro?
> i think there is still a place for khtml and gecko
and at this point, the KDE project generally agrees with that sentiment. KHTML is still being actively maintained and improved, and there’s no sign of that stopping. it’s more like the family is growing rather than we’re ditching one kid for another
Only the old QT libraries are *free*. SCO is a share holder in QT, until that problem is resolved I don’t care how well Konqueror renders anything.
Trolltech ought to either buy out SCO & kick em to the curb or KDE better find something else to replace QT with, i dont like the sound of sco being a share holder in Trolltech/QT
Chris Schlaeger is right.
And since QT has a bad license, QT will have to go. Too bad that KDE can’t use a different toolkit.
The head of Novell linux desktop, Nat Friedman, has ordered all developers who used to work for Suse and now work for Novell to start learning C# and gtk#.
As a long standing KDE & Konqueror fan I sincerely hope they do not impose the Gecko engine on users of Konqueror in SUSE 9.2 Professional, I would preffer being able to chose the rendering engine to use and this is something I would switch distributions over in a heart beat.
I’m not trying to be biased here, I use both Konqueror (KHTML) and FireFox, the former in Linux an the later in Windows and I know FireFox is a good web browser with lots of neat addons but I have made my decision and I preffer the KHTML engine because it was designed for KDE and clearly integrates better and starts up faster then Gecko. If its not optional then Konqueror would just become yet another version of Mozilla along side Netscape, FireFox, Galleon, Epiphany and a handfull of other web browsers.
KHTML has advantages of its own such as intigrated spell checking which I can guarantie FireFox (which I’m using right now) doesn’t have.
Qt doesn’t have a bad license, its available under the GPL and the QPL at the developers choice. Besides if anyone is going to make money off software that uses Qt but isn’t licensed under the GPL they should buy a license to support the company, besides if they are making money selling software they should be able to afford it.
Also it is possible to make money off GPLed software, the GPL doesn’t say that you have to give the software away for free is simply says that the source code should be available with the software at no extra cost.
http://www.gnu.org/licenses/gpl-faq.html#DoesTheGPLAllowMoney
Chris Schlaeger says there are too many Linux toolkits… and he’s right. So how can he justify trying to fully support both KDE and Gnome projects? Not just the base desktops, but all the various applications, a complete duplication of effort to re-implement everything twice. And they have to provide enterprise tech support for each, what a bloody waste.
Don’t get me wrong… I’m glad both projects exist, I think the competition they provoke is helpful to each project. I just don’t see the benefit that Novell gets from splitting its energy between the two. Doing it this way means that they implement projects the wrong way, for example netapplet being a notification area applet instead of a normal Gnome or KDE applet.
I’ve always found this to be an issue too but since Qt, GTK, wxWidgets and a lot of other toolkits out there each have their merits and are needed by the various popular Linux programs none of them can really be excluded.
SUSE might benifit from having a more open development process like Mandrake has in order to deal with the widening range of applications included in their distribution.
novell have allways sold servers. for a while they used windows as a “terminal” for their server solutions but then microsoft undercut them with windows network and active directory. basicly people could replace the novell servers with microsoft products and get a package deal.
what novell now is doing is to fight back by makeing linux as a terminal and allowing the sysadmin, or whoever makes the final call, to select what desktop to use.
as for useing gecko or khtml, i dont care as long as they both follow the standards set by w3c. but i would like to see gecko as a free standing library rather then me needing to install firefox to get it. then i may as well use it, right?
as for toolkits. sure there are a lot of them but some a nice to have around. like say you just want to make a quick gui for a script or something. then tcl/tk is your kit as its a scripting based gui rather then a compiled one. java is crossplatform so if you need it on serveral platforms then go ahead. but its overkill for a smaller project or a proof of concept, most compiled languages are.
as for them looking diffrent, personaly i dont care. sure it can be a pain for the disabled people as some apps dont work with screen readers are similar. but the solution to that isnt to take away choice in toolkits, but rather to make a toolkit agnostic system for said screen reader or similar.
> I’ve always found this to be an issue too but since Qt, GTK,
> wxWidgets and a lot of other toolkits out there each have
> their merits and are needed by the various popular Linux
> programs none of them can really be excluded.
Really? Can’t be excluded? Somehow I manage running only GTK with the exception to that is OpenOffice. I find it to be a very good user experience that way, and I imagine the same thing is true of people using only QT-based apps if they prefer KDE.
For Novell, dropping KHTML is a good start, but to finish they have to bite the bullet and completely ditch KDE. It will be hard for them to overcome the bitching by the fanboys, but trying to sell a desktop to enterprises that requires a Trolltech license to develop proprietary apps is just plain stupid. Or if they want to go that way, if they think they can sell KDE/QT, get rid of Ximian and focus their efforts.
There are too many toolkits, but two is not too many. Qt and GTK will both be around for some time, and that’s not too many. There’s room for at least a half a dozen, with two used most often. Beyond that half dozen are numerous obscure toolkits which are already fading away.
My guess is for the survivors: GTK, QT, Mozilla’s XUL/XPCOM, wxWidgets (formerly known as wxWindows), and at least two Java-oriented kits.
Going, going, gone: Motif, TK, Athena, direct use of xlib, and others that I’ve already forgotten. TK’s not bad, but it is dated, and I doubt that it can attract enough support to keep it up to date. X-only toolkits won’t cut it; native support for Windows is too important, and framebuffer support for embedded systems helps. Mac support could decide the order of the survivors.
We need to get over this idea that there can only be one solution. There will be multiple toolkits, multiple desktops, multiple operating systems, multiple databases, just as there will always be multiple programming languages. The concept of the One True Path has caused more misery throughout history than any other idea. Too few is just as bad as too many.
IMHO in a long term the KDE must die. Yes, at this moment there are too many toolkits under linux and it is not a good thing for the average users. And if this desktop environments will more advanced the paralell run of the Gnome and KDE architecture will require more and more resources (CPU,RAM, etc).
The windows with the one and only win32 API (and the .NET platform) give more solid user interface for users and programming interface for developers.
And the WIN32 API or .NET is freely available for the developers for any purpose (for commercial or non-commercial applications) the Qt license is far more restrictive. Yes, you can buy Qt commercial license, but IMHO the Qt is not too advanced development environment if compared to Visual Studio, and there is not exists any useable development environment on top of Qt (Kylix was good but Kylix seems to dead). And IMHO it is the biggest problem of linux. Without professional development environments the windows will win the race because will far more and advanced applications on windows then linux.
There will be multiple toolkits, multiple desktops, multiple operating systems, multiple databases, just as there will always be multiple programming languages. The concept of the One True Path has caused more misery throughout history than any other idea. Too few is just as bad as too many
IMHO it is not true. The best thing IMHO the only one API. It is better for developers (at this moment there is not any widget set which is exists on all desktop linux) and far better for users.
From Chris Schlaeger…
Linux is plagued by too many toolkits. We’ve got Tcl/TK, Java, Motif, Athena Widgets or the old X toolkit, GTK, and Qt, and all of them look and feel totally different. Applications written in those toolkits do not follow the same standards and guidelines and are a mess to use. Especially if you have them side by side or you need to use them frequently.
Tcl/Tk, Motif, Athena, and raw Xlib are all anklebiters so it’s irrelevant whether they stay or “go”. Java is a platform of its own and isn’t really used that much on the client.
If he really wants to get to the crux of the matter then he’s got to say that either gtk+ or QT has to go. Not in the literal sense, but in the sense of relevance.
I’d be curious to know how many ex-Suse employees are working on the desktop and how many ex-Ximian employees are working on the desktop. Is Novell going to stand for all of this duplication of effort. I think it’s crazy if they don’t focus all of their desktop developers on one DE and just prove the other DE as “extras” on the DVD.
LC is exactly right, but unfortunately the FOSS advocates are too busy screaming “choice” to see the forest through the trees.
“IMHO in a long term the KDE must die. Yes, at this moment there are too many toolkits under linux and it is not a good thing for the average users.”
Average users don’t care…average users want to be able to surf the web..write docs..print them..email them..etc..etc. Having 1,2,3 or 500 toolkits has nothing to do with this.
” And if this desktop environments will more advanced the paralell run of the Gnome and KDE architecture will require more and more resources (CPU,RAM, etc).”
uhm…Running KDE and GNOME at the same time?…the only way this is even remotely a concern is running KDE programs in GNOME or GNOME programs in KDE…while there are real concerns here, you don’t mention any of them.
“The windows with the one and only win32 API (and the .NET platform) give more solid user interface for users and programming interface for developers. ”
hahaaahaha
“IMHO the Qt is not too advanced development environment if compared to Visual Studio,”
ROFL..oh man..thats great…I feel no need to actually comment here.
“Without professional development environments the windows will win the race because will far more and advanced applications on windows then linux.”
You mean like eclipse..or kdevelop..or monodevelop. Anyway, since when did a dev environment mean anything? I use VIM…others use emacs..are we somehow not able to write “advanced applications”? Oh, and btw…windows having more advanced apps then linux?….I don’t think I really need to comment on that one either.
”
IMHO it is not true. The best thing IMHO the only one API. It is better for developers (at this moment there is not any widget set which is exists on all desktop linux) and far better for users.”
Only one thing is never best..go read history. Also, QT or GTK…what is so complicated about that?..When was the last time anyone saw a desktop linux that did not ship with both QT and GTK?
[quote]
Really? Can’t be excluded? Somehow I manage running only GTK with the exception to that is OpenOffice.
[/quote]
I was reffering to distributions not users. If Novell/SUSE were to drop all but one desktop environment despite having a lot of users from each they would end up losing a lot of their users to another Linux distribution.
[quote]
For Novell, dropping KHTML is a good start, but to finish they have to bite the bullet and completely ditch KDE. It will be hard for them to overcome the bitching by the fanboys, but trying to sell a desktop to enterprises that requires a Trolltech license to develop proprietary apps is just plain stupid. Or if they want to go that way, if they think they can sell KDE/QT, get rid of Ximian and focus their efforts.
[/quote]
Thats rediculous, the companies that use SUSE Linux don’t have to pay for a Qt License unless they actually plan on developing commercial software using Qt that isn’t permitted by either the GPL or QPL, in which case they should be able to afford a commercial license anyway and thats what keeps companies like Trolltech alive and making good products.
I also disagree with your statement about dropping KHTML, there is a good reason to provide users with a choice between KHTML and Gecko and that is that dropping one would result in a complete browser monopoly in Linux, also KHTML imo integrates a lot better with KDE and starts up faster and completely removing KHTML from konqueror would just turn it into a redundant copy of FireFox.
As for dropping all desktop environments but one that would cause them to lose a lot of users as I’ve already said.
Average users don’t care…average users want to be able to surf the web..write docs..print them..email them..etc..etc. Having 1,2,3 or 500 toolkits has nothing to do with this.
The users are almost irrelevant to this discussion. It’s all about the developers and what dependencies are required for their apps. But as user, it would be relevant to me that I have to load half of the KDE libs and all of QT to use that one KDE app. What a memory waste.
“IMHO the Qt is not too advanced development environment if compared to Visual Studio,”
ROFL..oh man..thats great…I feel no need to actually comment here.
Uhmm, QT is just a toolkit, not an IDE. Kdevelop has made great strides in their rewrite, but it’s still not as good as VS.NET. Eclipse is great, but it requires Java…… Monodevelop is way too alpha to do anything serious. I love Vim, but I would consider it going back to the dark ages to write a java app when you have Eclipse or IDEA at your disposal.
Only one thing is never best..go read history. Also, QT or GTK…what is so complicated about that?..When was the last time anyone saw a desktop linux that did not ship with both QT and GTK?
Yeah, but most people want a unified DE and also don’t necessarily install both desktops. You want that one KDE app, your going to end up apt-getting all of KDE.
I dunno, in my opinion, the dual license of Qt is it’s greatest strength. Buying the commercial license buys you an important thing. Support. Companies like support. Also, gnome is a mix of gpl and lgpl licenses. Linking your app with GPL sources forces you to release your source code for your app too, something I wouldn’t think many companies would be too hot about. As for SCO having shares in Trolltech (Qt), so what? Can they help that? At any rate, it’s a miniscule amount, not a majority. Educate yourself before condemning: http://www.trolltech.com/newsroom/investors.html
Linux has had both gnome and kde for some time now, and seems to be doing just fine. There’s room for both.
“The users are almost irrelevant to this discussion.”
I agree somewhat..for the dev area..ya they don’t matter..I was pointing that out to LC. However, to dump KDE or GNOME…that would piss many users off.
“But as user, it would be relevant to me that I have to load half of the KDE libs and all of QT to use that one KDE app. What a memory waste. ”
Then don’t…there are GNOME apps for almost everything..they may not be as good as the KDE app (k3b), but they are there. What is easier? Improving apps on both sides..competing to make the best one…or killing one off entirely and somehow forcing two communities with different ideas to join. Also..the memory thing is a lame argument these days. Memory is very cheap.
“Uhmm, QT is just a toolkit, not an IDE. ”
Yep..he compared them..not me. The QT API beats the crap out of anything MS has to offer and I have used both. Linux has IDE’s, they may not be as good as VS to some people..but they are also not required. I don’t like them..some do..its a preference, not an edict from god.
“Yeah, but most people want a unified DE and also don’t necessarily install both desktops. You want that one KDE app, your going to end up apt-getting all of KDE. ”
Most KDE apps depend on kdelibs and kdebase..maybe arts. All of that is not loaded when you start it either..so this is doing what? Filling up your 200 gig hd?
Do you really think these annoyances really worth ditching a choice between DEs?
Thank the powers that be that not EVERYONE thinks Qt has a bad license. People who don’t actually know things should stay on the sidelines.
Speaking of which, I’m seeing a few people say KDE needs to go. I have to guess that these people A) Haven’t used KDE for more than a few minutes, B) Have certainly never actually tried programming for KDE vs. programming for GNOME, and therefor don’t know just how powerful, feature-complete and elegant Qt+KDE is (and GNOME isn’t ugly, it’s just that KDE’s very good), or C) These people are newbies / zealots who are just vouching for their flavor of the minute.
“IMHO in a long term the KDE must die. Yes, at this moment there are too many toolkits under linux and it is not a good thing for the average users.”
And why not Gnome instead?
There is a very good reason why both KDE and Gnome should both continue to exist and that is that each one is geared towards a different audience. Gnome seems to appeal to those who are happy with a faily simple straightforward desktop and don’t mind some of the hard coded settings while KDE is geared to those who want more control over the desktop environment and like more tightly integrated applications.
My preference is KDE since I personally find it feels more professional and because I’ve also found Qt to be the best toolkit for my needs, but that doesn’t mean that every other desktop environment doesn’t deserve to exist.
Do us all a favor and do some real research before posting any more slander towards Linux or any of its applications or toolkits. I respect that people like and say good things about Windows and in return I expect others to respect that I like and say good things about Linux, without that kind of mutual understanding there would be no point in being able to comment on articles because it would always result in fights.
QT has a bad license
The GPL is a bad licence? I mean I perfer BSD but I’ll give credit where it’s due and the GPL has served us well. GCC being the prime example, especially where NeXT and ObjectiveC is concerned.
The current .NET toolkit is a silly toy compared to Qt. There is no more room in the world for non-layout-managed toolkits. UIs are just too much of a pain to to program without layout management.
“But as user, it would be relevant to me that I have to load half of the KDE libs and all of QT to use that one KDE app. What a memory waste.”
I preffer just to use the Qt libs when I can to write my programs to avoid this but personally I’ve run KDE apps in other desktop environments and the only think I noticed was an extra second or two for Konqueror to open in IceWM, this is the result of optimizations for KDE if I’m not mistaken and its a worthwhile price to pay since these programs were intended to be used in KDE anyway.
“. Kdevelop has made great strides in their rewrite, but it’s still not as good as VS.NET.”
That depends on what you are used to, I started out writing my programs using JCreator in Windows, then I started using Qt Designer after dropping Java and learning C++ and later on I had to use Visual Basic in school. From my standpoint Visual Basic was primitive in comparisson, it had poor syntax highlighting even compared to KWrite which was nothing more then a plain text editor for KDE and its selection and control over widgets was not as extensive as what Qt Designer had save some rare exceptions. The only advantage I saw it having was a button to compile/run the program within easy reach and even that paled in comparisson to what I was used to in Qt Designer and even the free version of JCreator (A java ide for Windows). KDevelop IMO beats even the latest Visual Studio, especially in support for programming languages and in my opinion its lacking only one feature compared Visual Studio, a visual form designer which can be found in Qt Designer and iirc an upcomming version of KDevelop is supposed to feature Qt Designer integration into the application itself.
Rainier has a very good point I forgot to bring up in my last post, MS still has not implemented layout managers in Visual Studio and without those you either have to make your application windows stuck on one size or spend a long time making sure your widgets are the right size and in the right place every time the window is resized.
Also..the memory thing is a lame argument these days. Memory is very cheap.
There’s a difference between how cheap memory is and what is actually in the average box. Most joe-average users don’t ever mess with memory and those 256 mems that come in the cheapy boxes are barely adequate for the modern Gnome/KDE desktop.
Most KDE apps depend on kdelibs and kdebase..maybe arts. All of that is not loaded when you start it either..so this is doing what? Filling up your 200 gig hd?
Do you really think these annoyances really worth ditching a choice between DEs?
maybe not all of kdebase and kdelibs is loaded into memory at app start, but because of the very nature of KDE (not as many libs as Gnome) a good chunk of it is.
I don’t think you can ‘ditch’ either desktop since it’s open source. I was looking at it from Novell’s perspective. Should have developers they employee duplicating effort. At some point in time, I think you have to have a “standard” that everybody has on their drive. It doesn’t mean one or the other will go away, but developers will know that the vast majority of users will have the desktop.
My question to you is would it be better if we had 4 “dominant” desktops that had roughly equal marketshare. Would that make linux on the desktop better in the long run? Or is 2 major desktops some kind of magical equilibrium?
The only reason that Gnome even exists is because of historical events that no longer exist (the QT license), so if the QT license issue had never existed and KDE was dominant, would we be better off overall? I think so. It’s all good to have some anklebiter DEs around for the competition, but I think having one standard is a good thing too.
You seem to be confused. QT is not a development environment it is a toolkit. QT apps can be made in many ways, on windows it has plugins for versions of both Borland and Visual Studio. On Linux there is Kdevelop. Of course you don’t need an IDE to make QT applications. Of course you can use QT Designer on Linux, Windows, and Mac. Also there is more than one toolkit/API on Windows. QT, GTK, and WxWidgets are all available on Windows.
Toolkit liscenses have nothing to do with corporations adopting a certain desktop over another. There is no law saying that if your company uses KDE on their computers that any applications they develop for those desktops need to use QT. Also as long as they don’t distribute their apps, they can use the GPL’d QT and not have to give out the source. So for internal applications, liscenses are a non issue.
“There’s a difference between how cheap memory is and what is actually in the average box. Most joe-average users don’t ever mess with memory and those 256 mems that come in the cheapy boxes are barely adequate for the modern Gnome/KDE desktop. ”
256 should be fine…I have a laptop with 256 and its just find.
“maybe not all of kdebase and kdelibs is loaded into memory at app start, but because of the very nature of KDE (not as many libs as Gnome) a good chunk of it is. ”
My understanding is that they are metapackages…I know they are modular. Just because they have grouped them all up, does not mean that KDE programs link to all of them.
“My question to you is would it be better if we had 4 “dominant” desktops that had roughly equal marketshare. Would that make linux on the desktop better in the long run? Or is 2 major desktops some kind of magical equilibrium? ”
Sure, I am hoping that XFCE gains a big enough following to start entering the GNOME/KDE circles. That would be excellent. There is nothing magic about 2…there is something horrible about 1. You mentioned a standard earlier..I am very pro-standards. Look at freedesktop.org. With their help such things as menu and notification are being standardized..hal..dbus…cairo..these things are standards across the desktops and will be in both. Having standards should not cut out choice and does not have to. This is one reason why the LSB is so important.
“The only reason that Gnome even exists is because of historical events that no longer exist (the QT license), so if the QT license issue had never existed and KDE was dominant, would we be better off overall?”
No..in no way..GNOME has brought us a lot and it would awful for all that to be erased. There are huge differences in opinions between the GNOME and KDE camp..those would not just go away.
Btw..I am mainly a GNOME dev…I have developed for KDE/QT, but I am not a KDE guy
You mentioned a standard earlier..I am very pro-standards. Look at freedesktop.org. With their help such things as menu and notification are being standardized..hal..dbus…cairo..these things are standards across the desktops and will be in both. Having standards should not cut out choice and does not have to. This is one reason why the LSB is so important.
So what about competition for HAL, DBus, cario, etc…Why are these standards so important, but having a standard toolkit somehow cuts down on “choice”?
Having one standard toolkit doesn’t cut down on DE choice. You brought up XFCE. I love XFCE as a lightweight DE and it uses the same tookit as Gnome.
There is no more room in the world for non-layout-managed toolkits.
It is not true. The layout management is only one way to solve the form resize problems. And IMHO not the best way, easier to create nice user interface without layout management with anchors property. I’am basically Kylix (and C++/wxWidgets) programmer and I really like the Delphi/Kylix widget handling systems witch are very similar to windows.forms. And I created small applications with Java and GTK layout managers but IMHO this layout managers are a little bit wooden.
“So what about competition for HAL, DBus, cario, etc…Why are these standards so important, but having a standard toolkit somehow cuts down on “choice”? ”
If those projects need to be forked then so be it..but what I was mentioning was that they were developed as standards. They are not KDE or GNOME specific and both KDE and GNOME people worked on them. If anything, even a standard, somehow gets too set in its ways or is doing the wrong thing then a fork can do nothing but good. Either they will merge again later down the line (ala GCC), one will die (ala xfree/xorg) or they will both thrive (ala kde/gnome). GTK+ was created for a reason..QT is still alive for a reason. Dismissing those reasons is stupid. Standardizing on one of those toolkits would just throw out the differences and force everyone to “just deal with it.” While this may be a good idea for a company, telling opensource devs to “just deal with it” isn’t going to go over well.
“Having one standard toolkit doesn’t cut down on DE choice. You brought up XFCE. I love XFCE as a lightweight DE and it uses the same tookit as Gnome. ”
Whats the difference between a toolkit and DE libs then? So its fine if GNOME and KDE use QT, but utterly different DE libs? If having the toolkit being the same is so important why not where the menus are? The default UI?
Qt doesn’t have a bad license, its available under the GPL and the QPL at the developers choice. Besides if anyone is going to make money off software that uses Qt but isn’t licensed under the GPL they should buy a license to support the company, besides if they are making money selling software they should be able to afford it.
But Qt is too primitive (too low level) to use instead of Visual Studio or Delphi and too expensive to base widget set of other commercial developer environment. The Kylix tired it but it is failed.
Also it is possible to make money off GPLed software, the GPL doesn’t say that you have to give the software away for free is simply says that the source code should be available with the software at no extra cost.
But if you give the source code nobody will pay for your software. Let see the reality with a professioanl developer’s eye: on the windows side exists many professional developer tools, this tools are not too expensive and easier/faster/cehaper to create a program with this tools then linux alternatives. And the other side there is a Qt with about VS prof’s price or you must give your program under GPL. Witch do you prever if you must create application with many (>100) forms, database with many (>50) tables and tight deadline ? Qt or Visual Studio ?
IMHO the base libraries of the system (glibc, widgets, desktop environment libs) must be free for any purpose. And it is the main reason wherefore KDE must die. Without commercial applications the linux desktop will only the hobbi-os fans playfield. And IMHO if glibc or libstdc++ released under GPL instead of LGPL at this moment the linux is only a hobby-os like ReactOS, Sybilla and others.
LC…you should have left this in Lumbergh’s hands
…he is a lot smarter then you are
Having a layout manager goes beyond proper resizing. It just makes programming a lot easier. Consider:
1) With a layout manager, you don’t have to manually place widgets. Most programmers aren’t artists, and having the toolkit automatically place widgets, thus having them automatically composed, is a big help.
2) With a layout manager, it is much easier to make your apps font-sensitive and language-independent. GNOME handles larger fonts perfectly, and KDE handles them nearly perfectly. Many Windows XP apps, however, have problems with layout when the font-resolution is larger than the standard 96 dpi.
a) First of all comparing Qt to Visual Studio is comparing apples to oranges. Qt is a library Visual Studio is an IDE.
If you really want to compare Qt to something then use the MFC and Qt beats the MFC by miles.
Second professional tools. I don't see any lack. You can get UML tools (Argo, Poseidon, Omodo), Database managers, IDEs (Kdevelop, Eclipse, Netbeasn and lots of others)
Source Control and so on
GTK+ was created for a reason..QT is still alive for a reason. Dismissing those reasons is stupid. Standardizing on one of those toolkits would just throw out the differences and force everyone to “just deal with it.” While this may be a good idea for a company, telling opensource devs to “just deal with it” isn’t going to go over well.
IIRC, Gtk+ was created for (surprise, surprise) Gimp and de Icaza needed a toolkit for Gnome – which was started because QT had license issues for the FOSS folks back then. You’re right that you can’t get people to really change now, even if the Gnome project was started because of a historical incident that has now been rectified. But we’re talking about Novell here, and how they’re going to approach the desktop issue. I just don’t see why they would put resources into 2 different DEs.
Whats the difference between a toolkit and DE libs then? So its fine if GNOME and KDE use QT, but utterly different DE libs? If having the toolkit being the same is so important why not where the menus are? The default UI?
I tend to look at things from a programmers perspective since I’m a programmer. KDE seems to be much more of a unified framework than Gnome, even though I’ll admit to not using the Gnome API much at all…some gtk+ though. The Gnome folks are actually going to be folding back the GUI stuff of Gnome back into gtk+, while the KDE folks add a lot of extra functionality into the KDE libs proper on top of GUI, including stuff like Kparts, Dcop.
I think that Gnome has the right idea in folding back the GUI stuff into gtk+ and KDE has the right idea of more developer-friendly, unified framework that allows easy inter-process communication and embedding.
I think if you do it right, then you could have a framework where the desktop components are modularized enough where people can put together DEs that differentiate themselves, but that can also easily interop with apps that are targetted at the framework.
I guess my point is that Xlib is somewhat of a toolkit and is somewhat of a standard…don’t see many people using framebuffer stuff, and what I would like to see is a higher-level gui framework standard, but where distros and developers can distinguish themselves, but still easily interop.
Of course this is all shades of gray and freedesktop is getting things moving in the right direction, especially since I think its a given that KDE will adopt D-Bus as its interprocess communcations for 4.x. I’ll also give the KDE folks props for changing an interop system that already works well for the greater good of Unix desktops.
“Gtk+ was created for (surprise, surprise) Gimp and de Icaza needed a toolkit for Gnome – which was started because QT had license issues for the FOSS folks back then.”
yep
“But we’re talking about Novell here, and how they’re going to approach the desktop issue. I just don’t see why they would put resources into 2 different DEs. ”
I agree. Novell should not split its attention like that. They will still have to ship both desktops, but they don’t have to be a driving force behind both of them.
“I tend to look at things from a programmers perspective since I’m a programmer. KDE seems to be much more of a unified framework than Gnome, even though I’ll admit to not using the Gnome API much at all…some gtk+ though. The Gnome folks are actually going to be folding back the GUI stuff of Gnome back into gtk+, while the KDE folks add a lot of extra functionality into the KDE libs proper on top of GUI, including stuff like Kparts, Dcop.”
As do I, and KDE does have a much better framework. Hopefully this will drive the GNOME devs to improve theirs. Ya, they are trying to get rid of libgnomeui et. al., but I was just pointing out that there is nothing special about using the same toolkit if everything else is different. (mainly libs)
I think the fact that this discussion emerged from LC’s posts has made it a bit more difficult. I think we actually agree on most of the things here. I also think that Novell should focus more, but I want the choice in DE on OpenSource desktops in general to remain and possibly more to rise up. If the main DE(s) is better then the challengers, it will stay on top.
> when’s the last time safari/webcore synched with khtml?
Dunno if this happens.
> is webcore its own entity far removed from khtml now
It’s basically becoming more and more an own entity. The reason are a team of full-time developers on one side and hobbyist on the others.
Why a lot of work? The load is shared (at least for the OSS clients) between many, SUSE is only one contributor of many.
“But Qt is too primitive (too low level) to use instead of Visual Studio or Delphi and too expensive to base widget set of other commercial developer environment.”
Qt is not low level, its very easy to use and very powerfull as far as any toolkit goes, also it is not too expensive if you are writing free software you can pick a license that is compatible with the GPL or QPL and you can use Qt for Free, commercial licenses are no more expensive then commercial licenses for Kylix or Visual Studio. You clearly have no idea what you are talking about and you are going around in circles.
“But if you give the source code nobody will pay for your software.”
Either way if people don’t want to pay for your software they won’t, if its proprietary all they’ll do is pirate a copy if they really want it that bad and that happens all the time with kids burning copies of games for each other or downloading a copy of MS Office off DirectConnect or some other P2P network. Your argument carries no weight since people often pay for good GPLed software anyway just to support the developers.
“IMHO the base libraries of the system (glibc, widgets, desktop environment libs) must be free for any purpose. And it is the main reason wherefore KDE must die. Without commercial applications the linux desktop will only the hobbi-os fans playfield. And IMHO if glibc or libstdc++ released under GPL instead of LGPL at this moment the linux is only a hobby-os like ReactOS, Sybilla and others.”
First off you’re going in circles again, there is nothing wrong with Qt’s licenses as I’ve explained over and over again, its available under the GPL and QPL and that should be good enough for any piece of software thats given away for free anyway, also it is easy to make money selling GPLed software — TheOpenCD and OpenOffice.org for example are all over eBay and I posted a link to the GPL FAQ which stated that GPLed software can be sold as long as the source is available at no extra cost. If a company wants to sell proprietary software that uses the Qt toolkit all they have to do is buy a license and if they are making money selling that proprietary software they shouldn’t have any trouble paying for the license.
Your statements about Linux being a hobby OS are rediculous, major companies like IBM are using it and its user base is constantly growing. You just said that in the hopes of starting a flamewar but as you can see no one here took the bait.
I highly doubt that Novell would even care about having to buy some QT license in order to develop commercial QT apps. I don’t really understand what all the fuss is about. The QPL license is a DEV license, not a user license. In other words, Novell’s customers will be able to run the commercial QT apps with the free GPL’ed QT.
I had a look at freshmeat concerning toolkits.
Remarks:
* ‘stuff’ means projects that are implemented very often (editors, file managers, etc.) or bindings or extensions for OpenGL, etc.)
* ‘advanced’ means the opposite of ‘stuff’
* for comparison: The search found 876 projects for GTK, and 266 projects for QT
* wxWidgets is the same as wxWindows, just the name changed.
* these numbers include some false positives (for example, ’emulate the look of Motif’ in a project description)
FLTK:
* 24 projects found including the toolkit entry
* 3 projects seem advanced and interesting, the rest seems to be stuff.
* 9 projects updated in 2004
* Mac ports seem to rather easy
FOX:
* 24 projects found including the toolkit entry
* 11 projects updated in 2004
* no project seems advanced
wxWidgets
* 4 projects found
* all updated in 2004
* 1 seems to be advanced
wxWindows
* 32 projects found
* 13 projects updated in 2004
* 3 or 4 seem to be advanced
Motif/Lesstif
* 53 projects found
* 16 projects updated in 2004, includes some false positives
* 3 or 4 projects seem advanced, the rest seems to be stuff
Tcl/TK
* 311 projects found
* more than 50 projects updated in 2004
* probably some advanced projects
Athena
* 13 projects found
* no project updated in 2004
* none seems advanced
Looks to me, Tcl/TK is pretty much alive and it won’t change soon. While Athena appears to be dead already, Motif/Lesstif seems to be dying, and ROX seems to be unable to start through. It’s a shame for the 3 or 4 advanced projects written for Motif. They should start using FLTK or wxWidgets/wxWindows, soon.
Please keep in mind, that I have no closer experience in using one of the above.
A small correction: Motif/Lesstif based projects are mostly written in C, thus GTK might be a better choice than FTLK, or wxWidgets. Sorry, didn’t know that.
For moving an advanced Motif/lesstif projects to a more modern toolkit I would recommend using Qt. Because it has the right tools to do the migration and you don’t have to do a wholescale rewrite as you must do with GTK. Or like the trolls put it:
http://www.trolltech.com/products/qt/migrate/motif.html
Trolltech offers a Motif to Qt migration tool that comes free with Qt. It allows Motif and Qt code to co-exist in the same application, allowing for hybrid applications while a gradual migration takes place, eliminating the need for a complete and immediate application re-write.
If the development of GNUstep continues as scheduled, then we will have something real powerful thing…a interface framework compatible with MacOS X!!!
But too bad that the DE and toolkit war is nearly dominted by two big giants….GTK+/GNOME and Qt/KDE….I always wanted more more diversity…but alas….the reality goes in the direction I never wanted.
I am going to try out GNUstep anyway.
I was getting sick of all the fanboys talking about how much better Konqueor was and blah blah this, and blah blah that.
Quite right. I really wish Mozilla was as fast as Konqueror, but I wish Konqueror could render some pages better.
It’s a real problem poser, but you just need to look at why Apple chose KHTML. They could have easily chosen Gecko and it would have probably worked better initially, but they felt KHTML would be faster and give them competitive advantage. They couldn’t just have Safari being yet another Gecko browser.
There are a lot of Safari KHTML patches, but they’ve been difficult to integrate as Apple just haven’t documented most of them, so I don’t blame the KHTML developers at all. It will take time to integrate them.
But what would happen now if someone bought out Trolltech? This is one of the reasons I prefer GNOME to KDE (the other is that I just think it’s nicer to look at and use
. Is there really anyone who could pick up the GPL QT code and continue developing it without delay? And how much code could be lost that hasn’t yet been released as GPL?
This is assuming the new owners didn’t want to continue releasing GPL version of course, but it does seem like a real risk?
For Novell, dropping KHTML is a good start, but to finish they have to bite the bullet and completely ditch KDE. It will be hard for them to overcome the bitching by the fanboys, but trying to sell a desktop to enterprises that requires a Trolltech license to develop proprietary apps is just plain stupid. Or if they want to go that way, if they think they can sell KDE/QT, get rid of Ximian and focus their efforts.
Well, I’ve documented my reasons as to why KDE is the better option (for enterprises, not for people who post here), and I won’t dwell on them here. Unfortunately, because of the amount of rubbish around no one actually does any reading or thinking.
But, you are right even though you’ve been ‘reported’. Look at what companies like Xandros are doing. You can drop a Xandros distro straight on to a Windows or a UNIX/NFS network (I’ve tried it on a few) and it works absolutely brilliantly. Why Konqueror the file browser doesn’t just look like XFM, I have no idea. It’s a good tool and a good bit of technology Konqueror, but some of the things done to it by default are totally braindead.
Suse/Novell have a good distro, but they’re spending too much time on chasing too many options. At the moment the smaller distros like Xandros are wiping the floor with what the ‘enterprise’ companies like Sun, Novell and Red Hat to an extent are doing. And everyone said these companies didn’t matter and it would be done by Novell. I bet that’s what IBM thought about Microsoft and Windows as well.
It’s a question that often comes up, but no one does any reading about.
But what would happen now if someone bought out Trolltech? This is one of the reasons I prefer GNOME to KDE
The last version of Qt would be released under, probably a BSD license. It’s set in stone (Trolltech Buyout Scenario):
http://kdemyths.urbanlizard.com/viewMyth.php?mythID=13
The KDE Free Qt Foundation was set up to answer and resolve these questions.
and at this point, the KDE project generally agrees with that sentiment. KHTML is still being actively maintained and improved, and there’s no sign of that stopping. it’s more like the family is growing rather than we’re ditching one kid for another
That’s fine from a community perspective Aaron, but Novell need to pick a direction. A community decision and a decsion an organisation makes are two separate things – they can’t just maintain several tracks of development.
The windows with the one and only win32 API (and the .NET platform) give more solid user interface for users and programming interface for developers.
Since you obviously haven’t used either and everyone I see here believes the hype of .Net without having found out the practical realities, your comments are totally irrelevant.
but IMHO the Qt is not too advanced development environment if compared to Visual Studio
Come back when you’ve done some development. Spend four or five days working out how you are going to perform layout management and to place your widgets with Visual Studio (I’ve seen it happen on many, many commercial projects), and then tell me what’s more advanced.
Uhmm, QT is just a toolkit, not an IDE.
Qt has a very good IDE (Designer), and other tools around Qt to help you get things done. It is not just a toolkit.
Thanks for the info, Morty. Hopefully, some of the project developers reads it, and starts wondering whether it makes sense to develop a project for several years just to see it die because the toolkit isn’t distributed, anymore.
The new version of Windows.Forms does widget Layout using
tables, boxes and flows.
“The last version of Qt would be released under, probably a BSD license. It’s set in stone (Trolltech Buyout Scenario):”
Thats not set in stone, Qt to date has had the GPL as one of its licenses so anyone is free to fork it and continue to provide it under the gpl or lgpl. Since there are a lot of KDE & Qt developers who use the GPL/QPled version in Linux there would be no shortage of qualified and willing people to fork or continue the product should a later version of Qt be licensed differently or be discontinued.
>> But what would happen now if someone bought out Trolltech?
> The last version of Qt would be released under, probably a BSD license. It’s set in stone (Trolltech Buyout Scenario):
In this shortage the statement is wrong. There is no “Trolltech Buyout Scenario”. A buyout alone has no effect at all. Only if the new owner decides to not continue to release new GPLed Qt versions, and this condition is independent from who is the owner or if there was a buyout, than the Qt Foundation kicks in.
“the smaller distros like Xandros are wiping the floor with what the ‘enterprise’ companies like Sun, Novell and Red Hat to an extent are doing.”
Thats a matter of personal opinion. I personally hated Xandros because there was absolutely no easy way to set Konqueror back to the default web browser and for that matter it was chaulk full of bugs that caused the entire desktop to crash so I had to restart the X Server. I payed for the Deluxe edition and all I got was a lousy, bug ridden disgrace to KDE and Linux which crashed more often then Windows from the simplest things such as trying to create a desktop icon or use a program that would work in any other distribution but would bring the entire X server down in Xandros.
Not only did they charge me money for something that should never have been labeled as a stable release but they also wanted more money for the simplest new or updated free software titles leaving the majority of users who were cleaned out by the purchase of the deluxe edition stuck with without a lot of necessary software and updates.
To top it all off Xandros networks would fail to install all packages if just one of them one could not be downloaded but would never say which one failed so you had to go through the entire list eliminating the bad package though a process if trial and error or just give up and forget about what you wanted to install.
SUSE, Mandrake and all the larger distributions on the other hand have much better package handling systems, have a lot more resources to put together a professional product and do a much *MUCH* better job of catching bugs and fixing them.
The new version of Windows.Forms does widget Layout using tables, boxes and flows.
There will be layout management coming in starting with Whidbey (VS .Net 2005) and the next iterative versions of .Net next year. It’s a big step forward for Windows.Forms granted, but not as complete as other toolkits.
The question is, do we really have to wait for Microsoft to produce this stuff first?
Since there are a lot of KDE & Qt developers who use the GPL/QPled version in Linux there would be no shortage of qualified and willing people to fork or continue the product should a later version of Qt be licensed differently or be discontinued.
Yer, well so what? The link provided was to look at what would happen with the Trolltech track of development, not anything else.
In this shortage the statement is wrong. There is no “Trolltech Buyout Scenario”. A buyout alone has no effect at all.
Given the amount of crazy people around, I think the words “Trolltech Buyout Scenario” speak for themselves and are what is implied.
Thats not set in stone
It is. See the link provided.
> Given the amount of crazy people around, I think the words “Trolltech Buyout Scenario” speak for themselves and are what is implied.
You hope that Trolltech will be bought so that Qt falls under BSD license? But that will not happen for above reason given.
For an open community distro like Fedora, Debian etc I can see the utility in supporting KDE and GNOME. For a corporate product this is retarded. Novell Linux will not see adoption if they try to ship a Mandrake-like “kitchen sink” distro. Remember most businesses know nothing of KDE or GNOME, and the choice is meaningless to them. They just want something that works. If one of their employees insists on an alternative, let them get the RPMs and run them…I am not disputing having the option at a post-install time.
Just remember new adoptees of linux will take what they are given and don’t care about the desktop war fought on OSAlert and Slashdot on a day by day basis.
But GNUStep is even uglier than fvwm. It fails the eyecandy test and is therefore DOA.
You hope that Trolltech will be bought so that Qt falls under BSD license? But that will not happen for above reason given.
No, I definitely don’t want Trolltech to be bought out (since the employees own Trolltech, that’s highly doubtful). It is essential for Qt to continue to have a working business model and be funded through it, and I like the situation as it is now. It also means that that funding can be used to produce even more quality free software
.
The crazy people I was referring to were those who run around screaming “what if they are bought out”, “Trolltech controls KDE”, “Trolltech is controlled by SCO”, and so on and so forth. It was necessary for the KDE people to dispel a lot of those myths. Of course, some of them are so stupid that no response can be given.
Thats a matter of personal opinion. I personally hated Xandros because there was absolutely no easy way to set Konqueror back to the default web browser
That’s a personal choice and not very relevant. I like Konqueror as well, but Xandros chose Mozilla – which still browses the web quite nicely. I was referring to the whole package from a functional point of view.
which crashed more often then Windows from the simplest things such as trying to create a desktop icon or use a program that would work in any other distribution but would bring the entire X server down in Xandros.
I never experienced that, but there are some annoying bugs in XFM. I don’t know why they didn’t just contribute and make Konqueror better and share the bug-hunting workload with the community rather than maintaining their own, but I guess Xandros have a lot to learn about that aspect of open source software.
they also wanted more money for the simplest new or updated free software titles leaving the majority of users who were cleaned out by the purchase of the deluxe edition stuck with without a lot of necessary software and updates.
Quite right, and they’re going to have to learn fast or make it a very bad selling point.
The point I was making was that Xandros is probably the best distro around for simply dropping straight on to a Unix/Windows network and having it work. Functionally speaking Xandros excels at this, but they’re going to have to remedy the rest of it quickly.
To top it all off Xandros networks would fail to install all packages if just one of them one could not be downloaded
I had that problem, and it seemed to happen sporadically depending on the reliability of their servers. Since Xandros is based on Debian and effectively uses apt-get, they really need to pull their finger out regarding this aspect. The tools are there.
SUSE, Mandrake and all the larger distributions on the other hand have much better package handling systems,
That may be so, but out of the box Suse and definitely Mandrake just aren’t well put together and integrated enough for you to drop them straight on to a network as a Windows replacement. Suse is not that bad, but functionally speaking, that’s what Xandros excels at.
With the negative aspects of Xandros that you’ve quite rightly mentioned, they’re going to have to pull their fingers out and solve them. But, that’s the competition everyone is up against though.
Since you obviously haven’t used either and everyone I see here believes the hype of .Net without having found out the practical realities, your comments are totally irrelevant.
I wrote about 200 000 lines of code in the last 10 years with different development environments (Cobol, Clipper, Turbo C/Pascal, Visual Basic, Delphi, GNU C++/wxWindows). And tired different other environtments (Java, C#, GTK, etc). In this environments I didn't created big programs, only small codes for testing/learning purpose. The layout managers are not too complex things, only a little bit sullen.
Come back when you’ve done some development.
I created about 500 or 1000 form with Delphi, and it’s layout management system is very similar to windows.forms.
Qt has a very good IDE (Designer)
IMHO you don’t know what is IDE mean :-).
IDE = Intergrated Development Environment = project manager + editor + integreated compiler + integrated debugger. The QT designer = a relative good GUI editor + poor project manager + not too good text editor. If you see the KDevelop under linux or the Visual C++, or the Eclipse, etc, this environments are IDE-s. The Qt designer is not.
Also as long as they don’t distribute their apps, they can use the GPL’d QT and not have to give out the source. So for internal applications, liscenses are a non issue.
But they do have to GPL their code up front, and deal with the consequences if they later want to distribute the apps to non-employees or if it is decided that the GPL applies to distribution to employees, or if they decide the app is so cool they should package and sell it.
If a company wants to sell proprietary software that uses the Qt toolkit all they have to do is buy a license and if they are making money selling that proprietary software they shouldn’t have any trouble paying for the license.
They do have to buy that Qt license up front, before they start developing the app. It wouldn’t be a big problem for a startup with a good business plan, or an existing business, but for someone just starting out in their basement, who isn’t sure if the app, toolkit, or the self is going to be up to the task, it is a big up front expense.
But they do have to GPL their code up front, and deal with the consequences if they later want to distribute the apps to non-employees or if it is decided that the GPL applies to distribution to employees, or if they decide the app is so cool they should package and sell it.
And they can’t then buy a commercial Qt license and go from there, even if they own all of the code they developed under the GPL. The commercial Qt license explicity states that code must be developed under the license from the start.
Well they could use BSD or LGPL for their library code and then release an application as GPL and later a newly written closed source.
Depending on the applications size a rewrite might be necessary after the “prototype” release anyway.
They do have to buy that Qt license up front, before they start developing the app. It wouldn’t be a big problem for a startup with a good business plan, or an existing business, but for someone just starting out in their basement, who isn’t sure if the app, toolkit, or the self is going to be up to the task, it is a big up front expense.
While not applicable to all situations that is a valid point, some people would have to pay upfront for their licenses if they don’t have one before they start writing commercial software. Despite the cost however some people should still be able to afford the cost and there are people in the Qt-Interest mailing list who will compile your applications for you with their comercial license for a few dollars so they make a few dollars and you get a compiled binary of your program that you can distribute under any license at any price.
There is also still the option of selling GPLed software until you have enough money for a Qt license and if Trolltech were just to give Qt away free for commercial use they wouldn’t be making enough money to pay their employees and there wouldn’t be a product. If you want good software you have to accept the fact that its not always going to be free, that would be like expecting grocery stores to give food away for free or banks to loan money with absolutely no interest.
And they can’t then buy a commercial Qt license and go from there, even if they own all of the code they developed under the GPL. The commercial Qt license explicity states that code must be developed under the license from the start.
You are mistaken, if you have a commercial Qt license you can distribute and/or sell your software under whatever license you want, AND since you own the copyrights to your own code you can always change the license your software is under at any time.
I really don’t think we should be including wxWidgets in our discription of “too many toolkits”. Even though wxWidgets is working on an X11 version that will be a true tookit WITH ITS OWN WIDGETS wxWidgets today is mainly being used as a cross platform MFC style WRAPPER around existing toolkits (Usually Win32 and .NET on Microsoft platforms and GTK+ on Linux/*nix based platforms).
This is a GOOD THING. I have always seen GTK+ with its pure C source code and proprietary software friendly LGPL licensing is kind of being the Win32 API of GUI linux and wxWidgets of being its MFC equivalant for MS Visual Studio afficianados that want to do cross platform programming. Even though I use KDE more than GNOME (Though this will probably change with my next Linux upgrade now that gnome has improved greatly over tyhe version I have now) I have always considered QT to be somewhat of an interloper in the
linux world. KDE should have stuck with the idea of developing their own LGPL toolkit which was what they were in the process of doing when the licensing controversy started.
Weather the fanboys and “free software” Ideologists like it or not proprietary software and shareware will one day come to Linux just as it has to all other operating environments
and in the long run it will be the GUI desktop and tookit that is the most friendly toward such software in its licensing policies that is going to win the desktop battle.
Right now that is GNOME/GTK+/wxWidgets, NOT Kde/QT.
Well they could use BSD or LGPL for their library code and then release an application as GPL and later a newly written closed source.
I don’t think they can for two reasons:
1. Since they are using a GPL Qt, all their code must also be GPL. Since they own the copyright on their code, they could theoretically dual license it later except …
2. Last time I saw it (Trolltech has hidden the text of it somewhere on their website), the Commercial Qt license explicitly required the developer to develop the entire app from the beginning using the commercial version, therefore they must purchase it up front. As NTWS01 points out, they could then release under the GPL a later date if they so desired.
There is also still the option of selling GPLed software until you have enough money for a Qt license
I think this would fall under the scenario I outlined in my reply to Kevin, unless you are talking about selling two different apps with no code in common. I agree with the rest of your paragraph from which I excerpted.
You are mistaken, if you have a commercial Qt license you can distribute and/or sell your software under whatever license you want, AND since you own the copyrights to your own code you can always change the license your software is under at any time.
Yes, you can do the former, but you would be violating the commercial Qt license by doing the later.
I wish I could find that license online again to refer to the clause that states this. Believe me, if it weren’t so, I would currently be working on a Qt app using the GPL version, with intent of buying the commercial version later when I was sure I had a viable product.
wxWidgets
* 4 projects found
* all updated in 2004
* 1 seems to be advanced
wxWindows
* 32 projects found
* 13 projects updated in 2004
* 3 or 4 seem to be advanced
It should be noted wxWindows is the former name of wxWidgets thus projects of wxWindows count under the name wxWidgets, too.
I created about 500 or 1000 form with Delphi, and it’s layout management system is very similar to windows.forms.
Wow, really? You’ve used Delphi and you think Windows.Forms has a layout management system? Sorry, but Windows.Forms and .Net won’t get one until next year.
You still come up with gems like this as well:
The windows with the one and only win32 API (and the .NET platform) give more solid user interface for users and programming interface for developers.
which is quite clearly rubbish. It could have come from a Microsoft press release (probably did).
IMHO you don’t know what is IDE mean
In my humble opinion, you’ve done very well to program what you have whilst writing terribly poor English. Can anyone understand your code comments? Do you have any?
IDE = Intergrated Development Environment = project manager + editor + integreated compiler + integrated debugger.
An IDE doesn’t come with a project manager. You need separate software for that. I haven’t got the foggiest what you’ve been using your IDE for. The compiler and debuggers are usually separate components, but are usually (but not always) able to be used from the IDE itself.
The QT designer = a relative good GUI editor + poor project manager + not too good text editor.
Qt Designer’s a text editor? Wow. It wasn’t designed to be a project manager, and neither was Eclipse or Visual Studio or……..
Last time I saw it (Trolltech has hidden the text of it somewhere on their website), the Commercial Qt license explicitly required the developer to develop the entire app from the beginning using the commercial version, therefore they must purchase it up front. As NTWS01 points out, they could then release under the GPL a later date if they so desired.
I’m not calling you a liar but I have never heard of this at all before and I’m in the Qt-Interest mailing list plus I’ve been using Qt for over a year. I think you are mistaken since that sounds absolutely absurd and since people in the Qt-Interest list know I’m using the free version of Qt and Qt-Designer and yet have still offered to compile my code with their commmercial editions (for a small price of course) if I wanted to sell my programs. I have heard of cases where software written by one person under a GPL licensed version of Qt was compiled with a commercially licensed copy of Qt and Trolltech didn’t prosecute those people plus I sincerely doubt that they’ll be able to tell from a compiled binary which license Qt was being used under when the program was written as long as it was compiled with a commercially licensed copy.
If you find that Qt License Agreement I’d be more then happy to read over whatever parts you think imply that programs have to be entirely written with a commercially licensed copy in order to be sold legally as long as that license is on the Trolltech web site so I know its authentic.
[quote]
IDE = Intergrated Development Environment = project manager + editor + integreated compiler + integrated debugger.
[/quote]
Actually an IDE doesn’t have to have an integrated compiler or debugger. As for the project manager Qt designer has that as well as a form designer and an editor with equal if not better syntax highlighting than I’ve seen in any IDE from Microsoft.
[quote]
The QT designer = a relative good GUI editor + poor project manager + not too good text editor.
Qt Designer’s a text editor? Wow. It wasn’t designed to be a project manager, and neither was Eclipse or Visual Studio or……..
[/quote]
Qt Designer can handle C++ source files and does a good job of it for that matter and as I’ve said the syntax highlighting is at least as good as anything I’ve seen from Microsoft if not much better. Calling it a text editor is hardly giving it the credit it deserves since it goes above and beyond what a good text editor does.
You’ve used Delphi and you think Windows.Forms has a layout management system? Sorry, but Windows.Forms and .Net won’t get one until next year
I use Visual Studio Express beta2 with .NET 2 beta. The 1.1 layout handling system similar to Delphi (without layout managers, but alignment / docking and the anchors property).
In my humble opinion, you’ve done very well to program what you have whilst writing terribly poor English. Can anyone understand your code comments? Do you have any ?
Yes, my english is very poor, because I am hungarian and I don't have enough time learning English. I can read documentation without any problem, but I write english text very seldom. The communication in english is not easy for me, especially if I a little bit tired. But I wrote my programs in C++, Delphi/Kylix, etc languages not in english.
An IDE doesn’t come with a project manager. You need separate software for that. I haven’t got the foggiest what you’ve been using your IDE for. The compiler and debuggers are usually separate components, but are usually (but not always) able to be used from the IDE itself.
I tired/used the following integreated environments: Turbo C++/Pascal, Borland C++ from 3.3 to 4.0, M$ Visual C++, KDevelop, Code Crusader, Eclipse, Borland CBuilderX.
In this environments the project management (you can add/remove source files to your project), the editor (in the modern systems with code completition), the debugger and the compiler are integrated. Yes, mostly this systems are used separated debuggers and compilers but you can compile your program from IDE, you can start debugging, add watch, call single step, etc.
I’m on your side.
Thank you.
Things would be better with a standard toolkit.
I agree. At this moment the applications uses widget sets. My favorite image manipulating tool, dictionary program (startdict), web browser(mozilla) are GTK based, but my favorite IDE (KDevelop), text editor (Kate), web editor (Quanta) are Qt based, my daily used RAD tool (Kylix3) is based wine + Qt2. IMHO it is not a too good thing. IMHO only one but coherent and powerful desktop environment is better then 3-4 small and incompatible system.I don't want only one windows manager, only one panel application but IMHO the common widget set, component architekture, sound server, etc can be good idea. And this things must be totally free (with LGPL or any similar license).
It should be noted wxWindows is the former name of wxWidgets thus projects of wxWindows count under the name wxWidgets, too.
I know. That was the reason I wrote under my remarks:
wxWidgets is the same as wxWindows, just the name changed.
However, I admit I should have spend more time doing the post, correcting grammar, and format it.
And maybe I shouldn’t have listed wxWidgets at all, like somebody suggested. It’s indeed just a wrapper, not a toolkit. It is, however, a valid alternative to FTLK if speed is not as important as native platform look.
> 1. Since they are using a GPL Qt, all their code must also be GPL
I wrote that the application had to be GPL when being distributed, no argue about that.
What I meant was that the developer could licence his/her library code under a licence that allows to be included in the GPLed application as well as the later close source application.
So large portions of the free version could also be part of the non-free version.
It would be similar to writing a non-free KDE application: large portions of it (kdelibs) would have been developed using Qt/Free but the application itself would have to be developed using Qt/non-free.
The following post to qt-interest quotes the relevant section of the Qt Commercial License
http://lists.trolltech.com/qt-interest/2003-05/msg00900.html#msg009…
NOTE: Qt Free Edition is licensed under the terms of the GPL and not under this Agreement. If Licensee has, at any time, developed all (or any portions of) the Application(s) using Trolltech’s publicly licensed Qt Free Edition, Licensee must comply with Trolltech’s requirements (see http://www.trolltech.com/developer/download/qt-x11.html) and license such Application(s) (or any portions derived there from) under the terms of the Free Software Foundation’s GNU General Public License (the “GPL”) a copy of which is located at http://www.gnu.org/copyleft/gpl.html#SEC1 (i.e., any Product(s) and/or parts, components, portions thereof developed using GPL licensed software, including Qt Free Edition, must be licensed under the terms of the GPL, and the GPL-based source code must be made available upon request)
There are also a number of discussions on that list as to whether this is legally enforceable or just how Trolltech would ever know. I don’t know if it is still in the license. I emailed Trolltech to ask if they would post/email me the text of the current license. If they do, I will post it here.
There is still a Q in the FAQ here: http://www.trolltech.com/developer/faqs/license_gpl.html#q14 that appears to be based on the section of the license quoted above. It says:
Q: Can we use the Free Edition while developing our non-free application and then purchase commercial licenses when we start to sell it?
A: No. Our commercial license agreements only apply to software that was developed with Qt under the commercial license agreement. They do not apply to code that was developed with the Qt Free Edition prior to the agreement. Any software developed with Qt without a commercial license agreement must be released as Free/Open Source software.
actually, Mandrake found itself having to do something similar too – there’s a net applet in 10.1, and it had to be done in the same way as there’s no compatibility between GNOME and KDE for panel applets currently. Using a panel applet is the “right” way under GNOME HIG and that’s recognised, but it’s not really workable to write two separate applets for KDE and GNOME, or one applet which can run in two completely different ways. I think freedesktop is going to unify this pretty soon, though, at which point the problem goes away.
“Gnome seems to appeal to those who are happy with a faily simple straightforward desktop and don’t mind some of the hard coded settings while KDE is geared to those who want more control over the desktop environment and like more tightly integrated applications.”
whoa, nelly! Integration is one of the most important goals for GNOME. If you’ve used a pure GNOME 2.6 or 2.8 desktop one of the things that strikes you is just how integrated the core *GNOME* apps are (GTK+ apps are a different thing entirely; same difference between KDE apps and QT apps) and how similar they are in appearance and interface.
also, very little in GNOME is hardcoded; the focus is to make defaults that just work and *remove configuration from the GUI*. The idea is that configuration should be needed as little as possible; if needed it should be available and work well, but not necessarily obviously accessible to the GUI-thwacking end user. gconf is the mechanism for this. It’s a system that works well for enterprise; if some type of configuration setting that isn’t default works better for an enterprise situation, it can be edited in gconf and propagated out to everyone’s desktop easily. I guess for power users on single machines or home networks who really love to tweak configuration KDE is maybe better, if they don’t want to hack around in gconf. But I run GNOME on my main machine and I find the “sane defaults” theory *really works* – I haven’t tweaked a thing on the system and I’m very happy with how the desktop works.
“That may be so, but out of the box Suse and definitely Mandrake just aren’t well put together and integrated enough for you to drop them straight on to a network as a Windows replacement.”
this is a vague statement with virtually no actual meaning. Please refine.
Although those posts were not current I’ll accept that you are correct relating to the version of Qt that was out at the time however I should point out that one of the posts you were refering was asking if the person could purchase a commercial license after developing the software but before deciding to sell it so he may have also been implying that it was compiled with the Free Edition. Also since that old license doesn’t make much sense I’m hoping Trolltech might have changed it. If you could turn up some more current information that would be good.
Also even if Trolltech still implements this they would have no way to telling whether a program was developed with a Free or Commercial license as long as the software was linked against the commercial edition rather then the free edition when it was compiled.
Personally though I don’t blame them for having such a license, they need to make money somehow to support their product lines and if they didn’t have this then major companies could just get away with purchasing one license for compiling the product and have all the rest of their developers use the non-commercial edition. I also think Trolltech would not prosecute individual developers over this as long as the software was compiled with a commercial edition, it seems more logical for it to be there simply to insure that the big companies behave themselves since they would be much more concerned about what they could stand to lose if they broke the rules.
As for me I use Qt because its a good product, if I were to fuss over whether something was truly Free (as in speech) then I wouldn’t bother with a lot of Linux distribution which are now adopting less community oriented development process and I wouldn’t dual boot Windows XP.
You brought up some good arguments, I don’t think we’ll see completely eye to eye on this since there is an element of opinion and preference involved but I do see your arguments as valid and I appreciate you showing the evidence you did to back up your statements.
Thanks for the discussion. I too like Qt and haven’t completely ruled out purchasing Qt up front to work on this program I have in mind. I also understand why Trolltech would include this clause. I would argue that if they aren’t going to enforce it (or can’t practically) why include it in the first place?
The Trolltech sales people responded and enclosed a PDF copy of their current license (version 2.8). The note I quoted previously is still there in section 10 of the license.