filebrowser, target resolution reduction

[Thomas Lübking]

Am Saturday 20 February 2010 schrieb Lydia Pintscher:
> On Sat, Feb 20, 2010 at 13:13, Thomas Lübking <thomas.luebking at web.de> 
wrote:
> > Am Saturday 20 February 2010 schrieb Jeff Mitchell:
> >> Are you suggesting something similar to what Kopete does with
> >> auto-contact-list scrolling, but horizontal?
> > 
> > eeehhh... <very tiny>i must confess that i've never really used kopete
> > nor IM at all, i'm a jerk</very tiny>
> > 
> > i guess the window has no scrollbar but implicitly scrolls against
> > mousemovement, so probably yes - bu i've no idea :)
> 
> It doesn't have a scroll bar but doesn't scroll automatically here,
> only on mouse-wheel scrolling.
Then NO.
The wheel operates on the wrong axis (so aside undetectable it'd feel a little 
weird).
The movement happens automatically.

The concept behind this proposal is seeing pointer moves as relative events.

Usually the user shares the widget context and the pointer moves against both, 
but if you need to "fake" higher resolutions, you can set the widget to a 
negative context.
(if you'd put the user into the pointer context the pointer would keep a 
static position and the environment changes - ego shooters work this way)

In addition:
As this will reduce the input resolution (for {pointer context}/{widget 
context} == 1.0, a 1px mouse movement will adjust the mouse by 2px on the 
widget) you can compensate this by increasing the pointer resolution (i.e. for 
a 1.0 vector scale warp the pointer back on every odd event)
This however get input devices like pens out of sync with the pointer and 
cannot be used with such devices.

I hope that was not too confusing ;-)

Greetings
Thomas