Re: GNOME 3.2: mapping extra mouse buttons

[Federico Mena Quintero <federico ximian com>]

On Thu, 2011-03-24 at 08:46 +1000, Peter Hutterer wrote:

> just because zoom is mentioned here: zoom, rotate and a few other actions
> are actions that are ideally continuous. Rotate isn't just "rotate" but
> "rotate by X degrees". as such, they are difficult to map to pure
> actions unless you also provide the angle, direction, velocity, etc.
> 
> an example where we already did it wrong are scroll wheels that are mapped
> to button clicks (though we're working on exporting those as valuators). so
> when considering actions, please don't forget that they may be more complex
> than just a "zoom in" action.

This is not a super-important problem, I think... 

Consider a scroll wheel for scrolling.  An app can get smart and take
"clicks" from the wheel, as well as their timestamps, to figure out how
fast the user is spinning the wheel - and to do some kind of nonlinear
scrolling in that case.

For zooming, you can do the same thing (although I'm convinced that the
best zooming behavior is already what is described in
http://live.gnome.org/UsabilityProject/HIG/ThreeZero/ImageZoomExamplePattern - you zoom by a constant *factor*, not by a predefined list like "100%, 150%, 200%".

If I remember correctly, Qt even does the scroll-wheel speed computation
for you; its scroll event has an "amount" field, computed from the
interval between clicks.

> other than that, I thnk the concept of "actions" is a good one, especially
> when looking a bit further ahead with touch gestures and their potential
> mapping to actions.

Yeah, gestures are definitely interesting.  They map pretty well to the
mechanism for actions.

  Federico