On Wednesday 25 June 2003 15:03, Monaghan, John wrote:

I think the problem with the replies is that they are all subjective (which is natural).

To a certain degree, yes. But I think most of those who replied have a somewhat educated opinion about the topic, more often than not based on personal experience. In my old job a few years ago (before I joined SuSE) I had looked around to find a suitable multi-platform toolkit for that exact purpose. I evaluated several of those things and came to the conclusion that the type of toolkit that just sits on top of other toolkits are insufficient for almost everything because they tend to give you the least common denominator for each target platform - which is very little. Don't expect a feature-rich environment if you choose one of those. I just saw you are writing from a .de domain, so also look at this: http://www.suse.de/~sh/qt/

Whilst I don't have any problems with QT, here is what I would say is the "other side" to give a balanced view.

1. I have to say that I emailed Julian Smart, the creator of wxwindows a few times (I too had lots of questions) and was very impressed with the replies; disregarding technical merit for the moment, I would recommend wxwindows politically.

Are you talking about technical support and response times or about the more "political" issues?

2. I know you said "apart from the licence", but I also found this to be an important factor.

No, not really. Because...

Basically, if you want to create free software (libre) and distribute it, then you can. However if you want to create commercial software, you can with wxwindows but can't with QT unless you buy the full product (expensive).

The Qt commercial development environment does have its price (although I consider it really reasonable). But for commercial products this really is no important factor at all: It's in the order of one or two "man-days" (the cost of a developer per day). This is negligible for a commerical company. It's runtime licences that would really hurt. And Qt does have none (execpt for Qt/Embedded which is a completely different issue). You can sell as many of your products as you like without paying any additional fee to TrollTech.

3. I personally think (subjective!) that the C++ implementation is cleaner in wxwindows and doesn't need things such as MOC. It also promises to be as platform independant as possible.

I find Qt's signal/slot concept simply ingenious. This means you can write independent UI components and "plug them together" at a later time. You don't have to subclass your own button from the generic button just to overwrite its "clicked" method to do your thing. You don't have to invent "user IDs" for button messages and clutter your .h files with them. This is what you get for the small price of having a (simple) additional preprocessor like "moc". Do yourself the favour and have a look at the Qt tutorial: http://doc.trolltech.com/3.1/tutorial.html I don't think UI programming can get any easier than this.

I think that a lot of people are enthusiastic about QT because it is used by the "darling" crowd for KDE.

It's the other way round. Qt was chosen as the base for KDE because of its features, its clarity and its ease of programming. They might as well have chosen some different toolkit - but they didn't, for good reasons.

I have no objections with KDE (I use it about 50% of the time) or QT, but like I said earlier we need to stay as objective as we can.

You do that. If you do, I am confident you won't abandon Qt so lightly. ;-) CU -- Stefan Hundhammer <sh@suse.de> Penguin by conviction. YaST2 Development SuSE Linux AG Nuernberg, Germany

1. I have to say that I emailed Julian Smart, the creator of wxwindows a few times (I too had lots of questions) and was very impressed with the replies; disregarding technical merit for the moment, I would recommend wxwindows politically.

The QT interest mailing list is also excellent.

2. I know you said "apart from the licence", but I also found this to be an important factor. Basically, if you want to create free software (libre) and distribute it, then you can. However if you want to create commercial software, you can with wxwindows but can't with QT unless you buy the full product (expensive).

Correct.

3. I personally think (subjective!) that the C++ implementation is cleaner in wxwindows and doesn't need things such as MOC. It also promises to be as platform independant as possible.

I don't know enough about wxwindows to give a judgement, but I find it hard to believe the implementation is cleaner than Qt. The comments I've heard suggest it's not really a class leading implementation, and I can tell you from first hand experience that Qt is. The moc thing is a red herring - it's just about invisible if you use Qt Designer (which nearly everyone does, AFAICT). Personally I prefer to use Qt from Python, and it shines brighter there than from C++. PyQt is the best scripting/GUI solution out there as far as I know.

I think that a lot of people are enthusiastic about QT because it is used by the "darling" crowd for KDE. I have no objections with KDE (I use it about 50% of the time) or QT, but like I said earlier we need to stay as objective as we can.

The reason a lot of people are enthusiastic about Qt is because it's a work of art. Really. Compared to every other GUI toolkit out there (that I have used or have heard first hand reports of), it's a superb piece of work. -- "...our desktop is falling behind stability-wise and feature wise to KDE ...when I went to Mexico in December to the facility where we launched gnome, they had all switched to KDE3." - Miguel de Icaza, March 2003