It can be ugly as long as the issues stated above apply

For me, and my purposes, this is true.

“Better” means nothing until you state: Better for whom, and for what purpose?

I have to agree with the author’s explanation for this, though. Software developers aren’t the best UI designers. The sad thing is, the two groups should work hand-in-hand, where interface designers are clients of software devs who have their requirements and software developers and designers strive to suffice these requirements.

I belong more to the software developer camp and I can already see developers thinking (with their narrow view of things) how eye candy can’t contribute to productivity or ease-of-use. Worse, more features usually mean a hit on software performance. (This is akin to console-only users who think that everything is much better on console compared to GUI apps).

This isn’t true, though, as evidenced in designs from companies like Apple. The important thing is for information and opinions to flow freely and for people to have the openmindedness to at least consider the possibility that there are better ideas out there than their own.

Can’t wait to hear the results from Hackfest.

In over twenty-five years of computing, I have yet to see a user interface I would describe as “pretty”.

Somehow, I survive.

- Firstly, I like the looks of most GTK2 apps vs QT apps. This is one of the reasons my projects ( http://entropy.homelinux.org/axis/ ) use Gtk2 instead of QT. In particular, I like the ‘aurora’ theme.

- Secondly, any work done on theming should maintain backward compatibility with Gtk2. Yes, I’m in the camp that opposes a so-called ‘clean break’ and creation of Gtk3. I especially don’t want to see a Gtk3 created *purely* for theming. I have a lot of production code in Gtk2, and I’m not interested in rewriting it for some damned themers. And I don’t think the community is big enough to support Gtk2 and Gtk3 together, which leads to my next point ..

- Thirdly, there are more important things to do than improve theming. There are a number of bugs I’ve reported, including DATALOSS bugs, in Gtk2. I’ve even submitted patches to a number of these. Sadly, there doesn’t seem to be enough manpower to even review these patches. Again, we’re talking about DATALOSS bugs. I’d rather see this junk sorted out before the eye-candy stuff is started on. Seriously.

If all these issues are address, by all means, go crazy with theming improvements. But for people who really want to break with the old Gtk and get some real eye candy, why not use EWL, ETK or Elementary from the Enlightenment team. They know how to produce eye candy. And frankly, I’m starting to think the Enlightenment dev community is bigger than the Gtk dev community. They’ve got *three* widget toolkits based on Evas. Yeah maybe that’s too many. But no-one can accuse them of being inactive.

- Do you have concrete examples of custom widgets so we can understand that problem better?

- What is the deal about Eugenia’s theme mockup there in the article?

- How does QT theming, for instance, solve these problems?

And btw., I use Murrina theming engine and like the look of my GTK+ applications very much. It is pretty to me.

http://audidude.com/blog/?p=62
http://github.com/chergert/custom-gtk-widgets/tree/836ca5502bf766c00e055544c201015ebda8062d/gtkmodebutton