A look at GNOME 2.14, comparison to KDE

[Luciano Montanaro]

On Wednesday 22 February 2006 20:57, Aaron J. Seigo wrote:
> On Wednesday 22 February 2006 02:05, Luciano Montanaro wrote:
> > It would be nice if applications were tested with a variety of styles.
> > Maybe a test with the default style and a weird (but useful) style like
> > "HighContrast" would be a useful routine.
>
> the challenge is that it's really not possible to make everything look
> good in every style. 

My point is not really about "good look"; it's quite hard to say 
HighContrast looks good, that's not its aim, and even if it may be 
improved, look is not the point of this style at all.

Anyway, there are widgets that do their own rendering for their bevel, which 
makes styles a bit less useful. For example, the color selectors use hard 
coded bevels (I have a patch about that, I think I sent it to this list 
fore review some time ago), metrics and colors are often hard coded.
I think these things should be handled by the style as much as possible; 
there may be exceptions, but if the style could handle most of these 
issues, I think it should.

> that's the power of having a HIG that covers these 
> kinds of issues: it allows us to say "these styles are all compliant with
> the HIG, and therefore things should look good with them" and "these
> style are no HIG compliant, so they may not look great with every app or
> situation". 
> this separations gives us a defined list of styles that we 
> can then check against, and people who use those styles can then
> rightfully file bugs against apps that may look good in HIG-compliant
> style A, but broken in HIG-compliant style B. we may only end up with a
> few HIG-compliant styles, and that's ok. at least we'll have a defined
> set of styles to work towards "looking good"

By the way, can you point to a draft of these HIG guidelines I can have a 
look at?

It would be good to know where we are heading. 

And I'd like to know which of the currently shipped styles would qualify for 
"HIG" compliant, and which do not, and why. 

Luciano 


-- 
Luciano Montanaro //
              \\ //
               \x/ www.cirulla.net