From: Tom Hawtin <tackline@ta...> - 2000-08-29 01:37:02
> > > > 4. Scroll when the pointer gets near the end (RISC OS style).
> > >
> > > This worked really well - so long as it works smoothly and quickly
> > > (nothing worse than waiting for it to chunder along at one
> > OK. I was hoping to skip 4 though, as it's the hardest to implement ;-)
> > > Some sort of indication at either end of the panel showing that there
> > > stuff off the edge of the screen would also be useful (purhaps a
> > > arrow in the appropriate top corner of the panel?)
> > Hmm, but if we're adding an arrow without making the bar shorter then
> > might as well have click-to-scroll on the little arrow...
> Thats true, but remember thingie's (can't remember his name, sorry) law
> GUI useability. A corner is easier to hit than a space or a side of the
Fitz, IIRC. (I have just finished Jef Raskin's the Humane Interface which
is quite a good read - it suggests getting rid of fillers amongst other
things.) Time taken to position pointer is roughly proportional to the log
of the ratio of distance to and perceived distance across figure (distances
along the same line). Edge of screen is good because experienced users move
the mouse way over the required distance. The little resize triangles in
some schemes doesn't work well, because the distance across is tiny when
coming from the center of its window.
The time taken and speed of the RISC OS iconbar scrolling irritated me. I
think to do it well you need to detect the retardation of the pointer as it
is about to stop in the scroll zone. Sitting there and waiting because any
shorter time might indicate that the pointer is about to leave is not good
enough, IMHO. Perhaps some subtle indication of when the scroll is about to
come would be helpful.
From: Thomas Leonard <tal197@le...> - 2000-08-29 15:45:24
On Tue, 29 Aug 2000, Tom Hawtin wrote:
> The time taken and speed of the RISC OS iconbar scrolling irritated me. I
> think to do it well you need to detect the retardation of the pointer as it
> is about to stop in the scroll zone. Sitting there and waiting because any
> shorter time might indicate that the pointer is about to leave is not good
> enough, IMHO. Perhaps some subtle indication of when the scroll is about to
> come would be helpful.
If we have a visible arrow super-imposed on the bar, then I guess it will
only scroll when you hold the mouse button down (although clicking should
probably move it a noticable distance as well).
This might need to be made an option, though...
tal197 at users.sourceforge.net
The ROX desktop (free/GPL) : http://rox.sourceforge.net