Re: gtkmm capabilities

[Alberto Paro <alberto paro homeunix org>]

Roel Vanhout wrote:
> Frank Naumann wrote:
> > I don't know the normal learn curve for such things but I'm sure that 
> Yes that's right, the 4 or 5 days is the amount of time it would take to 
> > That's I think it's much more (cost) relevant how expensive the 
> > development and maintenance of applications using gtkmm or qt is (at 
> > least in the long term).
> 
> indicated in another thread, I do not have enough experience with both 
> toolkits to give a valuable opinion on this.
> 
I develop multiplatform (Linux/Win32) and I have tried a lot of
toolkits. 
I everywhere use emacs or eclipse, gcc as compiler (mingw on win32) and
I had never got compatibility problems to port gtkmm/glibmm
applications.

Gtkmm, libglademm, and the other *mm stuff are the best way (in my
opinion) to learn or program in C++. They are implemented in the C++ way
(and UNIX way). It's easy to integrate a lot of other C++ packages as
boost and networking ones without complex change in code.

I don't like QT because is performance poor when you create big
application. The problem resides in MOC and in the look up table of
function call.

In this flame discussion, no mention to wxwidgets is made
(http://www.wxwidgets.org). It's a good library multi-platform, Open
Source, no cost. Learning curve is as QT. The approch of programming is
not as gtkmm but as QT. 
API and style programming of QT and wxwidgets is very similar to MFC.
The limit of old style programming are too much. Using library extension
is complex because they doesn't use std::string, namespace,....

I repeat this is my opinion. 

   Best regard,
            Alberto