Re: Notification Area guidelines

[Rodney Dawes <dobey free fr>]

On Tue, 2003-03-11 at 20:55, Evan Martin wrote:
> On Wed, Mar 12, 2003 at 01:59:20PM +1300, Mark McLoughlin wrote:
> > > > 	I think the gnomeicu approach is much, much more sane. If you think
> > > > about it from the "software should act like a butler" analogy, the gaim
> > > > approach is like telling your butler you have no need of his services to
> > > > which he responds by running behind the sofa and staying there until you
> > > > roar "will you ever F**K off and leave me alone!".
> > > > 
> > > > 	We don't need "minimise", "close" and "really close" buttons imho.
> > > 
> > > WRT user expectation, my Windows-using (non-geek) friend came over the
> > > other day and hit the close button on my GAIM window.  He expected it to
> > > minimize, but since I wasn't using the GNOME version, it closed GAIM,
> > > and he got confused and a bit frustrated.
> >
> > 	And therein lies the problem. Such behaviour confuses the meaning of
> > the "close" button. I don't think we want such confusion in GNOME thank
> > you very much :-)
> 
> Well, there is a rationalization, sorta: you really are closing the window,
> in that it's not visible anymore and it's not in the standard list of
> minimized windows.  The difference here is that you intend to get rid of
> the window and not the application.
> 
> So I could imagine someone becoming equally confused by your proposed
> behavior: they'll click the minimize button and the window will suddenly
> disappear from the window list (which any casual user likely thinks of
> as a task list).
> 
> I'm not suggesting that close-hides behavior is better; I'm just pointing
> out that it's not as cut-and-dry as you might think.

Iconify shouldn't hide the window, it should iconify it. Unfortunately I
don't see any way to tell if the window is iconified or not in gtk+. If
there is an easy way to tell that the window was iconified in code, it
would be trivial to implement iconify-hides as a pref. Close-hides is
already trivial, since you can just do it in the delete_event callback.

-- dobey