[linux-audio-dev] mouse wheel behavior and RFC: human interface guidelines

[John Check]

On Saturday 21 August 2004 04:41 pm, Lee Revell wrote:
> On Sat, 2004-08-21 at 14:45, Thorsten Wilms wrote:
> > On Sat, Aug 21, 2004 at 06:35:44PM +0200, Melanie wrote:
> > > Hi,
> > >
> > > it's backwards in a "numerical" sense, in that the numbers increase
> > > with one slider type, but decrease with another, using the same
> > > command.
> > >
> > > However, UI designers don't think in numbers, but associations.
> > >
> > > Left is generally associated with up, right with down, as we read left
> > > to right, top to bottom. Therefore, up MUST map to left, down MUST map
> > > to right, otherwise, non-mathematically minded people get uttely
> > > confused.
> >
> > I couldn't find anything on the web abou this.
> > But I asume the behaviour was thought out for
> > scrollbars and transfered to sliders.
> >
> > With scrollbars scrolling lets say a table, the vertical scrollbar
> > has top = start, bottom = end. Horizontal scrollbar left = start,
> > right = end. Wheeling up on vertical sliders means scrolling
> > toward the vertical start. Wheeling up on the horizontal slider
> > should therefor mean scrolling to the horizontal start -> scrolling
> > left.
>
> Very interesting.  You are probably correct.
>
> > But there's nothing to scroll with sliders. They're not about
> > a position in space.
> >
> > Today might well have been the first time I used the wheel
> > on common sliders, and it felt backwards!
>
> Agreed.  I can understand why Microsoft (and thus QT and GTK) chose to

The latter two, probably to fit in with the former. I think I may be missing 
something because my KDE sliders work like I expect.

> do it this way, but I bet audio users make MUCH heavier use of sliders
> than almost anyone else.  So, an informal survey of Linux audio users is
> actually pretty good data.
>
> > Fan-slider wheeling will stay as is, differing from QT and
> > GTK sliders. But I doubt the folks behind the toolkits would
> > listen and change wheeling direction.
>
> True, no reason to break it for people who are used to the old behavior,
> and MS does do extensive usability testing.  However this NEEDS to be
> made configurable system-wide.  This way CCRMA and AGNULA (for example)
> can ship with the non-default slider behavior if their users prefer it.
>
> Lee