Re: [Re: [2.4 feature freeze- bring it on]]

Depends on how it is integrated with the desktop preferences. Probably the
biggest single issue with the current gswitchit is the separation between
keyboard layout configuration and switching keyboard layouts (its not obvious
from the applet how to configure the ... available options). In general,
point-of-use "configuration" is better than centralized configuration (of
course, this can vary based on other factors, but I would say that's the
general rule of thumb). Is the user going to interact with keyboard layouts
anywhere but the switcher? If not... I think there are two basic routes to go
with the keyboard layout stuff:

1) Put configuration in the control center. Make the switcher not a panel
object (applet) but a status docklet thingee.
   Advantage:    Fairly clear how to change keyboard layouts, needs no right
   Disadvantage: Adds more stuff to the centralized desktop prefs.

2) Make it an applet, disconnect it completely from the desktop prefs
   Advantage:    Point-of-use configuration
   Disadvantage: Its currently tricky to add panel applets (this needs to be
fixed anyway, but in a longer time frame), and it won't be obvious that this
is the way to "change keyboard layout" (in fact, it probably isn't).

I think (1) is better for now, at least given the current state of
adding/removing panel applets. It would be nice if (2) could be the preferred
option in the future, because it would have a very nice effect of reducing
load on the desktop prefs. ah well.

Anyway, I'll try to get in touch with Sergey.... My sense is that it would be
a bad idea to try to fix gswitchit starting with the existing interface and
that the "feedback" should come in the form of a glade file started from
scratch and a series of mockups. Mimicking gkb's interface might also be a
reasonable place to start (but just to start ;-).


Jeff Waugh <jdub perkypants org> wrote:
> <quote who="Seth Nickell">
> > >         *control center and gswitchit are excepted until we get the
> > >         inclusion of gswitchit/libxklavier straightened out.
> > 
> > gswitchit looks like it has a lot of great features. Its also possibly
> > hardest to use piece of software considered for GNOME inclusion that I
> > recall, which is ironic because the basic role filled by the software is
> > very simple. gkb is not exactly an exemplary piece of work itself, but
> > (particularly after a lot of work by Kevin & Dennis) its to the point
> > its reasonable. It looks like the underlying architecture for gswitchit
> > could provide much better usability than gkb, but right now it'd be a
> > *huge* regression. The interface needs to be completely redone from
> > scratch, label fudging and pixel tweaking isn't going to be enough.
> The intention is that it would be integrated with the control-center, which
> would most likely fix up a lot of the usability issues (Sergey is very open
> and helpful when it comes to suggestions, and he would love feedback). The
> question is whether that work can be done in time for the initial UI
> freeze...
> - Jeff
> -- 
> GU4DEC: June 16th-18th in Dublin, Ireland   
>    "Warning: we're not going to use the testing sessions as vehicles for
>                     personal UI vendettas." - Anna Dirks


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]