OCT
4
2006

Qt 4.2 Released

Trolltech has announced the release of version 4.2 of Qt. The main features of this release are CSS-like desktop stylesheets, a new graphics view class, Qt/Mac look-and-feel improvements including the ability to host Carbon widgets inside Qt widgets and tighter cross-desktop integration. See the Qt 4.2 intro for a detailed list. The source can be downloaded for X11, Windows or Mac.

Comments

you mean: there is no release schedule -yet-. there will be one eventually. we're just not at the place where we can make one that we can reasonably commit to; we do have a decent idea of what generally needs addressing before we can start on that release schedule, so it's not completely unknown.

but yes, we are allowing ourselves the time needed. we usually do =)


By Aaron Seigo at Fri, 2006/10/06 - 5:00am

Well, this is a bit OT but I don't know where else to ask.

I'm looking for presentations (video's, slides, odt, ...) from aKademy, I did a gazillion of Google searches but I can't find anything except a video from a presentation by Aaron Seigo. aKademy website also doesn't seem to have them.

Anyone know if/where they are available?

Thanks a bunch!

Darkelve


By Darkelve at Thu, 2006/10/05 - 5:00am

High-quality videos from all presentations held during the Akademy 2006 Contributors Conference will be made available next week. Slides of many of the presentations are already available from the program pages on the Akademy 2006 website.


By Eike Hein at Thu, 2006/10/05 - 5:00am

Thanks for that info, I'm already eager to watch them :)


By Derek R. at Thu, 2006/10/05 - 5:00am

Thanks!


By Darkelve at Fri, 2006/10/06 - 5:00am

In related news, Qt Developer Days is this month. October 11 & 12 in Munich
and October 26 & 27 in San Jose. There are lots of good talks planned. San Jose's keynote speaker will be Bjarne Stroustrup.


By Brandybuck at Thu, 2006/10/05 - 5:00am

Congratulations for QT 4.2.

This slashdot article shows QT is $3300/- per seat!!!! (http://ask.slashdot.org/article.pl?sid=06/10/04/0452244&threshold=-1).

Is it true? Isn't it way tooooooooooo expensive for commercial software development for KDE? How much you guys think its reasonable?

Why Windows and Apple is so dominant because they have the backing of the commercial software sector.


By AC at Fri, 2006/10/06 - 5:00am

Actually, this is completely silly, as an argument. Let us take an extreme case.

You live in a country where your developers are paid 600$ a month, so 7200$ per year. So, excluding taxes, all in all, yours devs, with Qt cost you 11000$ a year.

Say you are a really small team, three devs. so 33000$ a year. so basically, to cover expenses, you need to sell the whooping amount of 2750$ a month. that would be 55 units of a crappy 50 $ shareware a month. Worldwide. Counting the windows market.

If you think you cannot do that, you'd better give up right away...


By hmmm at Fri, 2006/10/06 - 5:00am

Well, I am working for a software company creating engineering software. The problem: Specialized software mostly gets just a few, say maximum 30 customers. This can only be done by offering commercial support and expensive licences. If I run a shareware product where I can expect thousands selled copies or if am one of the major global software companies developing a mainstream (!!) application, then it is true, that it on't hurt. But not with us: Why should we increase the cost of our products by 10% per licence if we can find better solutions. It was often discussed in our team and with each release it is discussed again and again. There are the Qt/KDE enthusiasts, then there are those agreeing with them that Qt is a nice and comprehensive platform, and then we focus on our targets and stick with wxWidgets.
The reasons for wxWidgets: (1) It is free (of charge), (2) It allows proprietry software development, (3) it is completely cross platform, (4) Missing features are developed by us or are provided by other toolkits.


By sebastian at Fri, 2006/10/06 - 5:00am

>> Why should we increase the cost of our products by 10% per licence

Because you will save you more then 10% of the license costs on developing with Qt, compared with wxWidgets, where you have to develop lots of missing features yourself and need to shop for additional toolkits..


By AC at Fri, 2006/10/06 - 5:00am

That's not true in general. See, toolkits are appearing and disappearing over the years. At least for us, it is more important to have as much toolkit independent code as possible and to be able to migrate to a different toolkit within short time. The advanced features which are usually missing in other toolkits (eg. HTML/XML browsing, container classes, etc.) are such portions of code and it is unnecessary to save man power at the first time by delegating this to Qt when we know that we have to implement it anyways through another (independent) solution as soon as we use another toolkit.

Most of our developers' time goes into the algorithms and keeping the modularization uptodate, not into the GUI related stuff. Therefore, there won't be a saving of 10% time using Qt. Even the new graphics features of Qt which we certainly require are not that appealing. Why? They are basic stuff, we would need to extend them anyways. And then it is much easier to create OpenGL widgets being implemented from scratch where just keyboard and mouse controls are toolkit dependent. And this is something where wx and Qt don't differ in complexity.

I don't want to offend Trolltech's price policy or discredit Qt. I just want to point out that there are projects and companies where the calculation above is just not suitable. If you produce a software where the major code portions may go into the GUI related stuff, then Qt would be a wise choice. For companies like us, where the GUI is just a small portion of the software it may be not.


By sebastian at Fri, 2006/10/06 - 5:00am

Please do yourself the favor and look at the Trolltech homepage for the real prices. Qt comes in different editions, priced for number of supported plattforms: http://www.trolltech.com/products/qt/licenses/pricing

Also Trolltech offers a 65% (!) discount for small companies and startups.


By Nine at Fri, 2006/10/06 - 5:00am

I just talked with a customer yesterday who rolled their own OpenGL widget toolkit, because they were told it would be easier and cheaper than the alternative. But then the guy who rolled it for them left the company...


By Brandybuck at Fri, 2006/10/06 - 5:00am

"At least for us, it is more important to have as much toolkit independent code as possible and to be able to migrate to a different toolkit within short time."

Then you're going to be running around in circles trying to make everything 'independent' rather than actually doing anything.

"The advanced features which are usually missing in other toolkits (eg. HTML/XML browsing, container classes, etc.) are such portions of code"

Then you're wasting an awful lot of time and effort.

"Most of our developers' time goes into the algorithms and keeping the modularization uptodate, not into the GUI related stuff. Therefore, there won't be a saving of 10% time using Qt."

Qt isn't just a GUI toolkit.

"Even the new graphics features of Qt which we certainly require are not that appealing. Why? They are basic stuff, we would need to extend them anyways."

Wow. So you feel you can, and want to, replicate the effort that has gone into Qt because it's all just basic stuff? You don't do much programming, do you?

"If you produce a software where the major code portions may go into the GUI related stuff, then Qt would be a wise choice."

Again, you have no clue what Qt actually is, so your comment is meaningless. Qt is a whole programming framework, not just a GUI toolkit.


By segedunum at Sun, 2006/10/08 - 5:00am

And if trolltech goes down, QT will be released under a BSD-licence so it will be all for you to take: a wonderfull toolkit with no costs, independant to other companies.


By Matthias Logghe at Tue, 2006/10/10 - 5:00am

> if we can find better solutions

Well, that's the point, isn't it? If you find a better solution, you would be stupid not to use it.

However, for a lot of people, Qt is the best solution.


By Kevin Krammer at Fri, 2006/10/06 - 5:00am

> However, for a lot of people, Qt is the best solution.

at $3300/- per seat?


By Paul at Fri, 2006/10/06 - 5:00am

>> However, for a lot of people, Qt is the best solution.

>at $3300/- per seat?

Yes, don't you pay attention. That's a pittance in the proffesional SW development world.

The cost of a devloper is so much greater. In a non "low cost country" it's not uncommon that the cost of a developer will be in the range $1000-$5000 or more a week(That's cost, the sum of pay, taxes, rent, insurance etc).

With a 40week/year you get an cost increase about 2-8%. Given the productivity gain Qt may give you that's a nobrainer. And the $3300/- per seat is a onetime fee, so making it a 2 year investment things look even better.


By Morty at Fri, 2006/10/06 - 5:00am

It always depends what the next best solution is, if there is one at all.

Other options like Java/Swing or Mono/Gtk# might not have the same customer acceptance, or use "native" GUI like wxWidgets, which means it changes across platforms, or lack important non-GUI parts like networking, or lack commercial support...


By Kevin Krammer at Fri, 2006/10/06 - 5:00am

I think you make valid points. I've done some wxWidgets programming and found it a solid platform. I've done some Gtk programming and like it somehow too, bit complex though. Done some Fltk programming and was impressed by it simplicity.
Besides I would become a bit suspicious if non paid people are starting to calculate how much I could save buying Qt. But then again they probably drunk too much guinness and/or are still young and believe in the happy-ever-after.
No doubt Qt is the best toolkit of these, but of course it all depends what one needs and, as important, already developed over the years.
And exchanging wxWidgets for Qt is like buying a bigger house or faster car, it wont make you happy by itself.


By koos at Fri, 2006/10/06 - 5:00am

"Well, I am working for a software company creating engineering software. The problem: Specialized software mostly gets just a few, say maximum 30 customers....Why should we increase the cost of our products by 10% per licence if we can find better solutions."

Specialised software is where Qt excels, because specialised software requires specialised features and tools to build it with. If you have only 30 customers then I really hope that you're charging them enough, because if you're having to save on development tools (the stuff you use to build the stuff that pays the bills) then you're not going to sustain a business for very long. In fact, given your information you're not even going to provide a living for yourself and your employees if you're so cash strapped that you'll have to add 10% on to each license. Your arguments just don't add up.

I also hope that you're keeping those 30 customers happy, because that's all you have. I hope that your software actually works well and completely cross-platform, otherwise those customers are going to see right through the lack of quality, up sticks and leave you high and dry.

"(3) it is completely cross platform"

No, it isn't. The number of bugs it has on various platforms, and how bad it is on a Mac, tells you it isn't. Cross platform software is hard, and you're going to have to invest in it.

"Missing features are developed by us or are provided by other toolkits."

So rather than use developer time, and money, on creating the software that pays your bills, you'd rather spend it reinventing the wheel and creating.........frameworks? When you have thirty customers?

Sorry, but I don't believe you work for a software company at all. This load of twaddle is just not viable. It's just another person trying to tell us why paying for Qt is bad I suspect.


By segedunum at Sun, 2006/10/08 - 5:00am

3300$ is for the full license for 3 platforms. If you only want to do KDE stuff, you only need the X11 platform which is 1000$ and something


By carewolf at Fri, 2006/10/06 - 5:00am

You can work with "KDE stuff" on three platforms.


By Jaroslaw Staniek at Fri, 2006/10/06 - 5:00am

I'm in this business, and I find that commercial developers don't have a problem with this price. Even in small companies. Heck, I even know some Open Source developers who have a commercial license. If you can't make enough money with your app to afford this price, maybe your app should be Open Source to begin with.


By Brandybuck at Fri, 2006/10/06 - 5:00am

I find this argument is either a question from a largely ignorant postion or a cheap shot from a troll who knows better. So assuming it's a legitimate question the answer begins with saying you have no idea what the world of business and/or programming is like. Granting that there may be reasons you might not choose Qt for your project, price isn't one of them. There is one exception, where you want to take contributions from free software people and turn around and sell the code. Personally I don't see how that stands up with the community as there are retail solutions.

Looking at the cost of licensing, if licensing Qt is substantially more than 1% you should look at your ability to create a business model and should consider doing something else. Take a look at Programmers Paradise (pparadise.com) for prices of professional software tools. Like how IBM supports Linux? How come I don't see a lot of complaints that a fully outfitted Websphere software package costs over $50K? BTW don't expect it to produce W3C compliant HTML. Borland's Kylix Professional on Linux was $2K and in spite of all it's nifty stuff Qt is easier for me to use. I could go on, but why...

When the original developers of Quanta and I split they went to produce a commercial version. The paradox is that they had tens of thousands of users and yet they have done half a dozen releases to our several dozen while having a head start on our KDE 3 release and they have nowhere near the functionality. The point is that with all these advantages their commercial package is a failure with little activity today. A smarter choice would have been support and supplemental packages on a free package or consulting.

Commercial software is a very difficult place to be with a very high failure rate. To do it you need a fat wallet to get started. A few million dollars is nice. Succeeding in commercial software is worth noting and it takes savvy. I note that Trolltech has gone from a handful of people to over 100 in around a decade and they are obviously doing something right and selling a lot of software. To me it seems obvious that the people using their software are not lacking the faculties to make reasonable decisions and they are happy with the products and service they are getting. Therefore the question if the price is too high is really really difficult to ask if you ask even the most obvious questions. It may seem like a lot of money for your personal budget, but then if companies had those restrictions we'd all be raising sheep, riding horses and living in grass huts.

BTW the biggest expense any company has is employees... Software tools companies make money because anything that gets a slight performance improvement in developer performance is worth thousands of dollars. Qt is in a particularly difficult market to have any degree of success in and is thriving. I wish I had a company a fraction as successful but as I do have a successful company I know very well you get there by courting the customer and giving them more than they expect if you want to emerge a leader. If my company were developing commercial software I'd be using Qt, but if my company produced software it would be GPL and service model.


By Eric Laffoon at Sun, 2006/10/08 - 5:00am

Personally, I think you hit the nail on the head. The criticism of Qt's commercial license price often comes from people equating price to value. And fortunately for Trolltech, business people don't think this way. Value proposition is a whole different thing, and armchair quarterbacks frequently overlook that.

I'm not a developer, so can only rely on what I read and what I hear. And even with all the flaming Qt sometimes gets because of it's commercial "cost", I've rarely heard people complain about it's capability. Instead I generally see people comment on the advantages of coding in Qt versus (free) alternatives. So in a world where time is money, I would think that if Qt is an effective tool that can reduce development time, then clearly it pays for itself. The actual ROI would depend on the nature of the project and the cost of developers etc., but I hardly think commercial developers or organizations would dismiss it simply because it carries a cost.

Besides that, the criticism of Qt's commercial cost is mostly theoretical. "Developers won't pay when they can develop for free using another platform!" Yet the empirical evidence would suggest otherwise. Trolltech continues to grow, and we even see examples of commercial software produced using Qt specifically for Windows, not even ported to Linux (Adobe Photo Album comes to mind.) So if actual commercial developers see value in Qt even in the face of "free" alternatives, where does that leave the critics claiming otherwise?


By elsewhere at Sun, 2006/10/08 - 5:00am

Sorry...You're far from reality. How much do you think does an average company in the industrial sector pay for a professional CAD license?
That's > 15.000,00 Eur / seat (~ > 19,000.00 US$ !)

Now come back and tell me Trolltech has excessive pricing... Probably Qt as a toolkit is just reasonably priced. it's not that you get nothing for this price. Actually you get a hell of a nice and feature complete toolkit (with complete sources).
If you want be efficient (cause you want to make some money, do you?) try to make some wise decisions regarding the tools you work with.


By thomas at Sun, 2006/10/08 - 5:00am

You kidding? $3300 is peanuts for any non-garage company.


By Robert at Mon, 2006/10/16 - 5:00am

I've sent it to the spanish Digg (Meneame). Could you please digg it? You don't need to register to digg it:
http://meneame.net/story/publicada-qt-4.2


By Víctor Fernández at Fri, 2006/10/06 - 5:00am

Yes I know, there are Klearlooks and other ones, but nearly every theme does have problems in display error and/or performance. So I use Plastik theme.

An simple "integrated" Clearlook theme as "default" would be nice in Qt3.5 too. Is a backport possible?

Thanks!


By Anonymous at Fri, 2006/10/06 - 5:00am

No.

Qt 3.x is in very deep freeze. Only major bugfixes go into it.


By Thiago Macieira at Sun, 2006/10/08 - 5:00am

Is an additional theme something what is very deep? Clearlooks looks not that difficult ;)


By Anonymous at Sun, 2006/10/08 - 5:00am

You can install additional themes seperate from KDE, so no need to break the freeze for it.


By AC at Sun, 2006/10/08 - 5:00am

But there is no functional, fast & stable (as Plastik) Clearlooks Theme for Qt 3.x

So I asked for it, but no one knows anything. Thx


By Anonymous at Tue, 2006/10/10 - 5:00am

Does KDE4 plan to run with qt4.2 or a later version (4.3)?


By Vlad at Sun, 2006/10/08 - 5:00am

Right now I believe the plan is to require Qt4.2. It probably depends on how quickly Qt4.3 and KDE4 are released, and if theres any features that KDE devs want (like QGraphicsView in 4.2 is used extensively in Plasma). I think most likely only Qt4.2 will be required (though you may get a more accurate guess by flipping a coin or asking /dev/random).


By Corbin at Sun, 2006/10/08 - 5:00am

Later, I hope. In 4.2 the dock widgets, while really much, much better than before, are not quite good enough yet.


By Boudewijn Rempt at Sun, 2006/10/08 - 5:00am

Pages