Re: New 'GObject' as base for GtkObject?

["Shawn T . Amundson" <amundson eventloop com>]

On Tue, Dec 21, 1999 at 05:54:41PM -0800, Bill Huey wrote:
> 
> > You aren't going to convince anyone that the GTK core developers are
> > out-of-date idiots and you are some kind of programming genius (unless
> > you maybe point us to your cool code). So why not take it off the
> > list.
>  
> Well, part of *allowing* yourself to be convince by it and to actually
> look at other *modern* object systems. The current system is just plain
> barbaric. Anybody with any experience outside of X and Gtk would make
> the same comments.
> 
> Much of problems with GtkObject seems to be related to what I call a
> "least common denomiator" of software engineering understanding, where
> folks intentionally narrow their view of a problem just to get it done
> without proper design.

I've spent a lot of time working with this team, and I really don't 
think the "lowest common denominator" has anything to do with it.  Owen
and Tim, who have spent a massive amount of time coding (yes, writing
lines of C in an editor) the internals of GTK+, are more than a few 
notches above the average developer when it comes to understanding
software engineering.

After doing a good share of coding GUI stuff on BeOS (yes, after porting
GTK+ to BeOS I started coding in the native toolkit...), I can tell you
for certain that GTK+ is a hell of a lot nicer than some toolkits that
benefit from some more modern language features.

I think the only relevant discussion in this area can be how it will
affect language bindings and GUI builders.  Writing applications in C 
with GTK+ is currently top-notch.

-Shawn

--
Shawn T. Amundson                       amundson@eventloop.com	
Research and Development                http://www.eventloop.com/
EventLoop, Inc.                         http://www.snorfle.net/

"The assumption that the universe looks the same in every
 direction is clearly not true in reality." - Stephen Hawking