OSNews.com: Interview with KDE and Gnome Developers

OSNews.com is featuring an interesting three-way interview (5 pages) with KDE's very own Waldo Bastian and Aaron J. Seigo as well as GNOME's Havoc Pennington. An interesting diversity of opinions on various UI and usability issues is presented. "On the Ok/Cancel issue: KDE has implemented things in such a way that allows all KDE applications to have the order of these buttons flipped with a change to a single line of code. It is exactly this sort of brilliant design that allows KDE to be so internally consistent. So the question often comes down to whether or not we should make a change, rather than if we can. Personally I think it is irresponsible to impose personal aesthetics on your users in a seemingly random fashion by disrupting the interface they know without very compelling reasons to do so."

Dot Categories: 

Comments

by Erik Kjær Pedersen (not verified)

I am actually running KDE on redhat 8.0, and it is clear to me that Havoc
Pennington is no friend of KDE. I use kbabel a lot, but it was useless since
it was not compiled with database support, and there were many other small
issues. I then compiled kde to /opt, which is fairly easy using konstruct. I am
now very happy with the system. Having the two versions side by side makes the
difference rather striking. I am convinced the real motivation was to reduce
KDE to make GNOME look better.

by Troels (not verified)

Hej Erik :)

I am using redhat 8.0 as well, and i do agree that KDE on redhat 8.0 is useless, but i don't think its fair to blame havoc personally, as i don't think he has anything to do with KDE really. (i could be wrong)

Their main KDE guy (bero) did leave the company shortly after the 8.0 release though, after what seems like an internal disagreement as he too found their KDE to be broken (at least that is what one could read on the net). He is now one of the main guys behind arklinux.

by Erik Kjær Pedersen (not verified)

I think it is pretty fair to blame Havoc personally as he is one of the main advocates of "Nullifying" the difference between KDE and GNOME which basically is done by reducing KDE. I now run redhat 8.0 with three versions of KDE: the one that came from redhat, one that is close to being the 3.1.1 release and last night I compiled HEAD from cvs. Once you have a clean KDE install it is very easy to use kmenuedit to add various programs to the menu such as gimp ooffice etc. I had to do a few other things to be able to read Danish newspapers again without getting the special Danish letters replaced by square boxes. When it is so relatively easy for me to do this, what reason could there possibly be for redhat to deliver the inferior version they did. It would have been so easy to take a standard version of KDE, maybe replace a few K's with red hats if they absolutely have to, maybe add a "standard" default redhat theme without removing the standard KDE themes and left it at that. They made it extra specially difficult to get to use kdm the way I am used to: A line in a hard to find config-file
DESKTOP="GNOME"
had to be replaced by
DISPLAYMANAGER="KDE"
How is that for user friendliness.

The point I am trying to make is that I am really happy with redhat 8.0 with my self compiled version of KDE, and I absolutely can not comprehend why they had to deliver what they did. It was totally different in redhat 7.3 where KDE was actually very good. At work the system administrator had to go through all kinds of hoops to get the menus made so that the around 60 users that were used to KDE on redhat could still find the kmail they are used to.

Erik

by Jens (not verified)

I'm having the same problem with swedish characters.
Do U know about any solution for the Redhat version ???

by Erik Kjær Pedersen (not verified)

Yes, in /etc/sysconfig there is a file called i18n.
Change the first line from LANG="en_US.UTF-8" to
LANG="en_US"

I could also have chosen da_DK, but I prefer English on the commandline and Danish in the gui-programs, something that will be very difficult once the monolingual redhat maintainers have nullified all differences in this world.

Erik

by jmk (not verified)

> I'm having the same problem with swedish characters.

Ah, swedes. They're *scandinavian* characters, not swedish :)

Greetz a Finn

by Some Dude (not verified)

Wrong dude.. there is quite a some diff between swedish, danish and norwegian charsets ;-P

/SD

by Maynard (not verified)

I think Redhat's method of having the menus 'autogenerated' is better. Every time you install an rpm, it appears somewhere in the menu if it is an app that you would click on to start it. I think this is an issue of having the menus done better.
I think the point of a unified theme is NOT to make 2 desktops the same. It is to make all application look at least consistent to some point, regardless which framework they use, GNOME or KDE. When you launch Nautilus under KDE using Bluecurve, It has a similar look to KDE apps. That, is a good thing.

by David Johnson (not verified)

It has a similar look to KDE apps. That, is a good thing.

It's more than just the "look". If that's all it were, then simply use Bluecurve, Keramik/Geramik, Qinx/GTK-QNX, etc. But that's not the case. The issues go much deeper. They go even deeper than just autogenerating root menus. This whole thread brings up many issues.

Trying to make KDE just like GNOME, or GNOME just like KDE, is not feasible now, or in the near future. This is an example of Redhat operating under the paradigm of "we know better than you." They are behaving as if they know better than the KDE developers, GNOME developers, KDE users, GNOME users, and *you*. That's the Microsoft way of doing things, and it has no place in the world of Free and Open Source development. I use Free Software so that I can be in complete and full control of my own system, but Redhat doesn't want me to have that freedom.

When KDE built from scratch (via konstruct) works better than the default Redhat packages, then you have to blame Redhat. No two ways about it.

by Joe (not verified)

"They are behaving as if they know better than the KDE developers, GNOME developers, KDE users, GNOME users, and *you*. "

No. What I see here is you trying to make it into something that it isn't.

I'm a KDE user, and I like what Redhat has done. I like the simpler menus, and I like the fact that, by default, gtk and qt apps look similar.

I use GNOME as well, and I don't mind what they're done to GNOME or KDE.

You think Redhat knows better than the KDE developers? Perhaps it goes both ways? Perhaps KDE developers think they know more than Redhat? Regardless, it's not an issue of who is right or wrong, it's about opinions and agendas that don't necessarily match up. KDE is trying to build a gui with EVERYTHING included, while Redhat is a distributer who is not as interested in KDE's agenda as it is in providing a solid experience for it's users.

I really *HATE* it when people try to speak for everyone else like this. You sound as if you're trying to say that *YOU* know what is best for others. Leave Redhat alone. Leave us people who like Bluecurve alone (although I prefer Keramik). You aren't going to convince us that red is an ugly color, and you won't convince them that Bluecurve is a crappy theme. You can't convince us that simpler menus suck and menus with everything in them is better. You can't convinceus that Evolution sucks compared to Kmail, and you can't convince us that Mozilla is not as good as Konqueror. People will always have different opinions and experiences with things, and have different ways of going about things. Just because someone else's way isn't your way, it doesn't mean it's the wrong way.

I realize that some of the KDE purists don't agree with Redhat, and that's fine, there will always be people who take these things personal, but the fact is that for every person who doesn't like Redhat's KDE, there is someone who does.

So please, give it a rest. What they have done makes the computer more usable for ME, and that's what counts. Not what you think I should be using.

by Eric Laffoon (not verified)

> I realize that some of the KDE purists don't agree with Redhat, and that's fine, there will always be people who take these things personal, but the fact is that for every person who doesn't like Redhat's KDE, there is someone who does.

> So please, give it a rest. What they have done makes the computer more usable for ME, and that's what counts. Not what you think I should be using.

Would that it were that simple. I guess you're a KDE purist if you don't think you should display Red Hat only in your booth at a trade show when you've gotten no support from Red Hat but have from Mandrake and SuSE? I know I felt like a KDE purist when I read that Havoc issued a bug report on KDE's single click default. I'm in complete agreement with Aaron that it's a far more rational default setting.

I think you should compile KDE from source and run it and then compare to the default install before being able to say you like the changes. How can you have a point of reference? Have you tried the KDE print system on cups? Oops! You're running RH.

Here's what I know. Quanta 3.0 was released Sept. 30th last year. Yet RH has shipped it with 3.0pr1, a notably buggy quick "get something out the door" release. 3.0pr2 was much better behaved and out two weeks before 3.0. Our 3.0 release has docs, plugins and a lot of bug fixes. I seem to remember that RH 8 came out in October or November. Did they intentionally ship a pre release? I don't know. Would it have taken them long to update it if they weren't really mucking around with it? No! Would I have released pr1 if I'd known RH would ship it? NO!

More than cosmetic changes have been made by RH. I would say that Apple is much more friendly and courteous to KDE as they are using khtml. It seems to me that since RH can point to my software as one of the great features they have they could do things differently... send patches, interact with developers, ask about a final release date if an app is on the margins of thier release and generally demonstrate a little more good will toward the developer community.

Lest we forget it was not that long ago that Linus Torvolds chewed RH out for shipping a beta compiler with broken C++ support. Note that KDE is largely C++ and Gnome is not, so perhaps there was some humor in this for RH but as it turned out some of the kernel is C++ and kernel compiles were broken. Along with Linus being upset so were gcc maintainers. It seems RH has a knack with developers.

by Maynard (not verified)

Redhat has a strict policy of which packages are released in its distros. Sometimes they ship things which are broken in ways they know. Most of these are stuff a normal user would not care much for. If you are a power user, you probably can get most of the stuff you need elsewhere. Corporations are not loking for moving targets. Redhat has made some promise that software built for their releases will be compatible for something like 5 years and all teh releases between. i suppose they are trying to make their releases future proofquite a bit. Right now there is no assurance that software developed for other distro's will continue to do the same.

Redhat is actually giving Linux much needed exposure, so don't just slate them, encourage them. You can live off their success too, even if you do not use them.

by free market (not verified)

Please people, give this a rest. Redhat is doing what they feel is in the best interest for their company and their share holders. If they don't do the right thing, then sales will reflect as such, and they can either change or face losing market share. This is just simple economics at work. If you don't like what they have done, just use another distro or remove the default KDE on Redhat and install from source. You can also buy enough the company, get on the board and vote. They're not anti-kde, they're not pro-gnome. They are pro money making, plain and simple.

Cheers

by free market (not verified)

Whoops ... didn't finish my thought here ...

You can also buy enough stocks in the company thereby getting a vote on the board. Then you can vote for someone who makes decisions that you agree with.

by Eric Laffoon (not verified)

> If they don't do the right thing, then sales will reflect as such, and they can either change or face losing market share. This is just simple economics at work.

They've already lost a lot of market share. How do you think Mandrake got so big and knocked down so many numbers. Now explain to me why they're still on top of SuSE and Mandrake when they adopted journaling file systems last, are the only one that can't resize FAT partitions on install (generally requiring extra software purchases) and are several releases behind the other installing cups which totally rocks. They also don't have the great KDE print system and I'm sure others could go on from here.

Do you know any high tech company who has put marketing in front of technology and succeeded? If so then your hypothesis has holes in it. What have we learned? If you can keep shoveling something down somebody's throat without actually gagging them they won't have a reason to look at the menu. Simple economics gives way to name recognition, puppy dogs and fear of change. If you build a better mouse trap will the world really beat a path to you door? How about your web site? So much for theories.

> You can also buy enough stocks in the company thereby getting a vote on the board. Then you can vote for someone who makes decisions that you agree with.

How naive are you?! How much stock to be a voting member? How much does your vote count? Look, I'm pro business and pro open source, but the reality is once your company goes public your are not serving your vision or your roots, you are serving your stock holders. Stock holders are notorious for two things, short term perspective and profit at all costs.

Very few companies today manage to implement a set of standards the way Thomas Watson did with IBM. No matter how you applauded Bob Young in "Under the Radar" you'll note he's not at the helm today.

by wlsb (not verified)

"are the only one that can't resize FAT partitions on install (generally requiring extra software purchases)"

considering the market their after atm is the server market, this isn't important. What release of Solaris do you think will introduce this feature? the release that finally kills it, maybe. I mean an "Advanced Server" with pre-existing FAT partitions probably isn't that advanced after all, even by Windows standards.

"and are several releases behind the other installing cups which totally rocks."

again, totally good for desktop use, you'll get more use out of lpd or lpr in the server market. Workstation usage is an afterthought for RH atm, so that's probably why it doesn't have all the good stuff that other distros might.

yes, KDE on Red Hat sucks. Gnome might be considered better, but considering they're just trying to provide a GUI, and not trying to provide a great desktop OS, I can't see where it's such a problem. I'm sure if CDE would have been "free as in beer", Red Hat would still be shipping it, except as the default environment. as far as RH being a few steps behind, maybe that's the spin they're after?

by free market (not verified)

"How naive are you?! How much stock to be a voting member? "

Actually, it doesn't always take a lot to get a vote (of couse that depends on your definition of a lot). Each company is different. I don't know what it takes for Redhat. If it's a lot of stocks, you can always get a group of people together and buy stock that way. Regardless, all I was saying is that there are many ways to combat a company that is not doing what you like. I was also trying to point out that Redhat doesn't have any hidden adjenda to kill KDE. They just care about doing what they feel is in the best interests for their share holders and their customers. If the majority of their base is happy, they make money.

What is the problem with making profit, as long as it's done within the law? That is the goal of buisness. Without profit, there is no product. Share holders want profit. To get profit, you need to send some product to market that outsells the competition. If that product is superior or not is irrelevent. What is relevant is what the buyers want. Obviously Redhat feels that what it is selling is what their customers want, and there is nothing wrong with that.

Now companies like IBM are different. They have huge sums of money, which allows them to throw large portions of that into pure research for tomorow. Not all companies have this ability. Research costs a lot of money and is not alway fruitful. Yes, IBM will be around for years to come because they have the ability to do so, but I don't think comparing IBM with Redhat is totally fair.

by Maynard (not verified)

Eh, say loudly MICROSOFT. Yes, marketing will win over technology any day. Some people place too much importance in things like resizing partitions on the fly. Heck, even XP does not do that. XP lets you only create 2 types of partition, FAT32 and NTFS. Some of this is unnecesary hassle for the users. Minimize the choices and the choices actualy become useful for 99% of the users out there. Most people I know do not care if their partition are ReiserFS or EXT3 or whatever. I do not know of too many profitable Linux companiews too besides Redhat. Mandrake has a place on the enthusiast desktop, like BSD has now too, but one day, when it comes to the real deal, I see Redhat forging forward and leaving them because they have a more clear goal. SuSE has a clearer goal and will probably compete with Redhat, if not only because of their German connection. But I would not contract Mandrake right now for a service contract. Their future looks shaky indeed.

by Navindra Umanee (not verified)

They did make both desktops the same, ON PURPOSE.

That, is stupid.

Interoperability is *another* matter. You do not have to make both desktops the same to get good interoperability and integration between the apps.

And btw, Debian already did the menu thing many years ago.

by Maynard (not verified)

"They did make both desktops the same, ON PURPOSE."

They did, but for reasons of consistency and branding. If the KDE people do not like it, they should not have released their desktop under the GPL. You can't claim to be GPL and say that, 'you can do whatever you like with teh source', then when someone does exactly that you turn around and say, 'don't do that to our source'

It was not so much about integration, but about getting things to look similar. From my experience, any app woulod look totally out of place under teh other desktop. Stop just bashing Redhat here. Go and bash Mandrake while you are at it too, because they are doing it with 9.1 too. Yes, they will have 'Galaxy' and they will have a modified KDM too.

You have to know that Redhat is not simply downloading apps and putting them on some CD's, they are working on them too to make the sum of these otherwise loosely related programs a whole thing. They do not want to duplicate functionality. They choose one tool for the job, and get rid of all the others. Unlike other distro's.

by Erik Kjær Pedersen (not verified)

I am not discussing their right to do it, but how they did it. Historically KDE
spent a lot of time building foundations, while GNOME built good programs
that were quite separate. Now that KDE is about to reap the benefits of a real
solid foundation Pennington is trying to convince the KDE team that they should
throw out, or replace this foundation for the sake of integration.

It would certainly be nice to have gimp and open office better integrated in
KDE, but without being a programmer it seems to me the gimp and open office
people could obtain that goal if they really wanted. It does not have to happen
by ripping out KDE's foundation.

It seems to me my original point remains: That I get a much nicer (and much
more bugfree) version of KDE on redhat 8.0 by doing my own compiling than
the one delivered by redhat

Erik

by Strider (not verified)

There is alot to be said for this statement. Gnome apps lack consistancy and integration between EACH-OTHER... is it really hard to understand why they lack integration with KDE?

So Redhat wants to make KDE/Gnome look like eachother, act like eachother, work like eachother. Fine! But lets not be so ignorant as to believe that they are NOT pro-Gnome (even if they are not anti-KDE.) Gnome has a huge investment in gnome that they don't want to see fail. They have a large number of Gnome developers under their payroll, and not ONE KDE developer. The single biggest "complaint" by end users of Redhat 8 was that there was no feedback when an icon or button was clicked. Something like a third of their total complaints were concerned with this! But did they enable icon/button feedback in KDE, NO! Why? because a thrid of their users would have started using KDE bcause "it was just like Gnome but better." Redhat's response to users that complained about this missing "feature" was... "it will be available in the Redhat 81."!!!?! Of couse we all know that it was available on Redhat 8 (if you where willing to use KDE and turn it on!)

Strid...

by Navindra Umanee (not verified)

Did I set my lawyers on Red Hat? No.

Look at what I wrote again.

I said that for interoperability and consistency of KDE apps within GNOME and vice-versa, *you don't need to make both desktops look and act the same*. I said that making the desktops look and act the same by default was *STUPID*.

Get the distinction?

by Erik Kjær Pedersen (not verified)

I think this thread goes along way to explain redhats behaviour

http://mail.gnome.org/archives/desktop-devel-list/2003-March/msg00026.html

The urge to "nullify" the difference is obviously stronger when behind!

Erik

by foo (not verified)

The most meaningful fact I could extract from the article was:

Waldo and Aaron from KDE are a LOT better looking than Havoc from GNOME! What a nerd that Havoc is!

Plus 'Waldo' is such a cool name. "Where's Waldo". Go KDE!

by Zkhizo (not verified)

But Waldo's photo seems to be from like 20 years ago, or so.

by Datschge (not verified)

So where are the swarms of girls joining KDE? ;)

by YouKnowWho (not verified)

They moved to Gnome after seeing Havoc's picture... :D

by reihal (not verified)

"Plus 'Waldo' is such a cool name."

I think "Havoc" is the c00lest hacker name. I am sure any system administrator would love to install his software.

by Jody (not verified)

Havoc is really nicer...

by KDE User (not verified)

Havoc didn't say "KDE" one single time.
That in itself does says something, though.

by KDE and GNOME User (not verified)

Envious? Havoc's answers appeared to be carefully written to avoid emotional influences or responses.

Havoc was never promoting either desktop or making comparisonal statements. He answered the questions in the spirit of cooperation between the two desktops, which I thought was the point of the interview.

by KDE (not verified)

Havoc hasn't answered clearly on the questions asked to him, he always babbeled around the questions hyping his own stupid visions how a Desktop has to look like - which IMO are wrong visions.

by Joe (not verified)

When asked about Windows XP:

Waldo Bastian: I am not familar with Windows XP.

What conclusions can we draw from this one? Would you say Waldo is envious of Windows XP and therefore refuses to test it or acknowledge that it exists?

.. or, perhaps were are making something out of nothing, like a bunch of immature children?

I'm really dissapointed at the overzealous attitudes that exist here.

by Tim Jansen (not verified)

Maybe just that it costs a couple of hundred euros/dollars and he has no need for it?

by Joe (not verified)

I'm sure he knows someone with Windows XP, and if he didn't want to buy it he could just use theirs to see what it's all about. Perhaps he doesn't want to know how KDE stacks up against Windows XP so he doesn't use it?

I don't believe this, I'm just using similar logic to the guy that suggested that Havoc is jealous just because he doesn't actually say KDE. One could also come up with many reasons why Havoc didn't say KDE.

If you believe Havoc is jealous of KDE merely because he didn't say KDE, then you must think something similar of Waldo for never having used Windows XP. Note that *I* don't this way, but if I used KDE User's logic, I would.

IMO lots of people including Havoc are hunting a phantom.
Computers will never be as easy as a telephone.
Why? Because they can do a lot more. And people - even
novices - at some point want them to do more.
If one colleague with a "more complex" desktop can change
his background image / print landscape instead of portrait /
whatever - the other colleague (novice) who sees that will ask:
Oh! You can do that? Is that difficult? Can I do it too?
This happens many times every week in my company.
If I told them: Sorry - that doesn't work with your OS they will
at some time get the impression that their OS is inferior and
they want to change. Why did Corel Draw succeed and is now
ironically know to be uses by a lot of novices. Do they use all the
functions? No. Corel came, put in more functions than any
other drawing application and succeeded. It got rave review even by
novice computer mags. Nobody said: Hey there are to many functions.
This is how the market works. This is how advertising works.
Do we really use all the bells and whistles in todays products - no.
For example I never use the "record every week" function of my VCR.
But nevertheless it's there. If you can get a product with more
options / functions / features and one with less at the same price
the first one wins everytime. Furthermore lots of people use
perhaps only one or two "advanced" functions but everone of them
uses a different one. That's why there are so many functions.
I never use the multiple desktops in KDE for example. But others
would certainly cry out loud if they removed it. So even if I think
it's useless I do not want to force my opinion on everybody else.
But this function doesn't bother me. It can be removed / turned off, so
it's OK for me.
So the challenge IMO is now to organize many functions in user-friendly simple interface. This is no contradiction. This is what KDE is all about. Lots of OpenSource apps have the functions but a horrible interface. Dumbing down the
interface shows that the GNOME nerds didn't really understand
what good interface design is about. KDE is THE hope of the opensource
community to acquire something which made M$ what they are:
Interface Design. Keep up the good work.

It seems to me that all this "less is more" attitude in Gnome2 is just a lame excuse for not being able to keep up with KDE. But good luck for every1, nonetheless.

Yeah, they removed a lot of already existing features just to be able to excuse themselves for not being able to keep up with KDE...

I don't like Gnome 1 or 2, but it still annoys me whenever i see such silly cheap shots at it. And i am happy that KDE and GNOME seems to have a very different focus. I know a bunch of people who prefer GNOMES focus, and i know a bunch of people who prefer KDE. Its a matter of personal taste really. But just because one prefers the KDE way doesn't mean that one has to disrespect gnome and start trashing it whenever possible in public.

I'll make it in private next time, then. ;-)

by Tim Gollnik (not verified)

As it repeatedly comes to the question if there should be more or less options to configure the desktop I see clearly the missing point, and this point is also applicable to features. If an application becomes bloated can be seen in different ways: bloated in the way that every application has its own and independent options concerning the same thing, or bloated in the way that the applications code is unmaintainable: there are a million places where you have to introduce, debug or change replicated or even different code concerning the same thing.

These are typical problems for programs written with mostly not or not easily reusable code, written in plain c.

Both do NOT apply to KDE.

What we see is this:

In a first attempt someone implements a new feature. Maybe it is not very well reasoned in terms of reusability, but the initial idea is good. The next time the same author or someone with more insight to the overall very good structure of the KDE-API has an idea to implement a class or to improve and extend an existing class which gives a more general solution.

At this time every application takes advantage from this, and, what's even better, this improvement comes to the complete desktop. In a lot of cases even without recompiling the individual application. In a consistent fashion.

Now neiter the capabilities to configure nor the code are bloated.

And every improvement to this is easy to implement. At one place.

THIS is the reason for the rapid development of KDE.

It's quite amazing, and it's unbeatable in terms of productivity. The last point is the reason why a quite small number of maintainers produces such a good and fast evolving desktop. Look at the safari developers: they come from nautilus, eazel and mozilla, and now they use khtml and kjs.

It's not a flame war against GNOME, but where would we ALL be if all the GNOME developers could only take a look at C++ and the Qt/KDE-APIs. It's not difficult. It's easy, if you made the first step.

I fear it's only the GPL (of Qt) instead of the LGPL that lets a lot of companies put money into the GNOME desktop, they want a free platform for developing commercial software. But if they would precisely calculate the total costs, they would see that they NEVER get their money back. The costs of Qt developer licences are VERY small compared to the gain of productivity. It's simply a wrong decision.

Don't get me wrong: I don't want to offend GNOME developers. I wish you could see how fine and productive KDE is. You can write better code in less time. And you instantly profit from the work of all the other KDE developers, as they profit from your code. This is essential in free software.

Best regards and good luck

Tim

P.S.: The key words in respect to KDE aren't bloated or complicated, they are: symmetrical, powerful, consistent, simple. KDE: Kan Do Everything.. it's a pleasure to work with KDE.
Thank you all for the good work.

by Troels (not verified)

I'm sorry, but really, QT is dirt cheap. I could understand companies not wanting to use QT just to develop on Windows alone, but adobe seems to think this is a good option too. But if you are doing something cross platform, then i don't really see any alternatives. (you could use java, but that would alienate users on all platforms as its look and feel doesn't match the target platform, which is why i personally consider mass market GUI java appliations impossible)

Sure i know there are alternatives to QT, like wxwindows, but is it as feature rich or as well documented? It sure wasnt the last time i looked at it.

And finally, good developers are expensive. Any software that can improve their productivity is a good thing. I personally feel that QT is a bargain, and so do many companies, at least those using C++, but i would never wan't to work for a company that write GUI software in C anyway :) (Unless it is really well paid, then i could take the pain ;)

by Lauris (not verified)

Why is software so expensive? Because everyone thinks, that if you are
commerical entity, they you can pay like hell.
I was working in IT startup in 3rd world for a for a while. The total
funding to operate with, was ... well, let's say below 100000E.
I wanted to use Qt because of its features, and boss did even agree on
that - until we discovered the Qt license was per-person. Then the case
was closed, and instead we used wxWindows.
And why we didn't chose open-source model? Because we didn't have the
necessary marketing money to compete on open field. Neither were we able
to compete with brain-hunters. Open source program is no bigger than
core developers - buy them out, and you have the control.
So do not think the world is so uniform that 3000E is uniformly dirt
cheap everywhere.

by Bob (not verified)

Why is software so expensive? Not because everyone thinks that you can pay like hell. That theory went out the window around the beginning of 2000.

Software costs so much because of basic supply vs. demand. No, shipping unit #2 doesn't cost any more or less than shipping unit #1 (for the most part). What we're paying for is the fact that software developers are expensive. It takes a lot of them to develop the comercial software and then mainain it. With Qt, you are also looking at a rather small audience. The number of people that want to use Qt are much fewer than, say, the number of people who want to use the latest Whiz-bang game. Fewer people buying the product means that the costs incurred in the development of that product must be spread out over a fewer number of licenses.

In addition to that, the demand for Qt is relatively high. When demand is high, you can command a premium under the assumption that people will be willing to pay for what you have to offer.

by KDE User (not verified)

> Qt license was per-person

This is not my experience. They can give you other deals too.

by David Johnson (not verified)

The price of Qt for a single development platform is the same as that single development platform. The price for it for more than one platform is *considerably* less than the individual systems. In other words, why can you afford $1500 for a quality development platform, but can not afford $1500 for quality tools?

Let's look at it another way. How much does Qt cost in comparison to a years wages of a developer? I don't know what developers are paid in the third world, but I still suspect there's room for a professional company to provide employees with quality tools.

by stephane PETITHOMME (not verified)

Hi all,

I am working in Malaysia for a Local company. Not exactly a very poor country, but neither very rich.
Here the averadge monthly income for a IT engineer is 2000-2500 RM per month for a beginner (or In Euro 500-625). A 5 year experiance one may get 3000+ (750 Euro). Well that means that a dual QT licence (3000Euro almost) represent 4 to 6 month of labor. This is much better if you consider that the real need is more for a Windows Qt only.

Well for a typical project like that one we handle (1 year, team of 3-4 people), that represent a Huge extra cost... That we can not afford.
I will say you need at least to plan using Qt for more that 3-4 project to get it affordable, which is not that simple as you want to maintain last version. The maintenace cost is not also that cheap.

Only solution: Take quite reasonnable effort to split the job in the team and design is such a way that only a few person uses Qt (Best is only one). Typical idea are: Client server solution where Server is not using Qt and client is Qt base... Just one constraint: don't use Qt network class and make your own! One this is done you get you freedom.

Further more, we are force to plan development in such a way that the server is developped first, using a fake client (text mode in our case) to delay the acquisition of Qt to the last moment. Difficult to spent som much money too ahead of the customer sell.
The ideal case would be to get the customer before finishing the product, and make it baring the Qt cost.

Regards

Stephane

by ac (not verified)

Or you can contact Trolltech and ask for a better deal. I am sure they are quite reasonable people.

by Paul. (not verified)

I would agree with you that the initial costs of Qt are acceptable. However there are some problems. If you want to port to an embedded device, you start to pay royalties for each program shipped. No matter how small, it's too much. These licenses have to be managed, and controlled which adds costs. Also Qt is very unreasonable for negotiating this point. They want to tap into any revenue stream that they can, and suck on it like a parasite.

Your argument about java not matching the look-and-feel of the target platform is not a good argument when all Qt controls are created from the ground up by Trolltech.

Okay, let's compare wxWindows with Qt. Let me agree with you that wxWindows is not as feature rich or as well documented for argument's sake. How many feature differences are there? How much are those features worth? What about the documentation? Same questions. IMO, it's not the cost of Qt, it's the headaches that comes with the costs. Consider that the Qt licenses cannot be rapidly transferred from one named user to another. What if someone wants to do some work at home? They breech their agreement with Trolltech, and the application would need to become open source.

So I ask: what are we looking for?

1. No run-time royalties
2. Concurrent licenses for the development tools.

If Trolltech did these two things, the choice would be so perfectly clear to use Qt, there would be no other choice. But we are dealing with Trolltech, whose marketing & sales people are about as dumb as rocks, while their technical people are very intelligent. And this will ultimatly kill Trolltech.

Say anything bad about wxWindows or Fltk, but just remember there is no cost to upgrade to the next version nor to add new developers to a project. And if you don't like the features or documentation, volenteer to help out.

by Anton Velev (not verified)

There is a very serious problem with making commercial apps for kde, and it's directly related with QT licensing policy. Let's say for example that company decides to purchase QT commercial licenses, will they use their licenses for making commercial kde apps? The answer is no! It's because when having money invested for a technology you have to make profit (else you die), and on the other hand we know that people buy QT because of it's ability to compile on multiple platforms. So everyone can guess that decision maker with comm QT will decide to not use KDE but only native QT because if do so will have 100 times more users (Win and MacOS X about 99% market, KDE 1%). Alternatively that decision maker may decide making separate branch for kde using kde's features but on a commercial world this kind of effort seems unreasnoble (many new lines that use cool kde features for only 1% of user base).

I think that KDE should spend time on porting just the kdelibs on the other qt platforms, and QT has financial interest i think to support that effort. And QT does so because of the large LGPLd kde codebase of extremely cool features (kioslaves, khtml, arts, dcop, kparts, etc. as well as various components and widgets) that can be great advantage for QT if having so for all their platforms, then the market situation would change in positive way for both QT and KDE.