This is a good architectural/design rationale - it'd be worthwhile if you could add it to either Wiki or the doc/ directory?
Bryce
On Sat, 31 Jan 2004, MenTaLguY wrote:
On Thu, 2004-01-29 at 18:39, Alan Horkan wrote:
I'm not sure if it is this is a good keybinding for panning but it should not be used when there is a selection.
Panning must not depend on whether there is a selection.
I dont understand. Can you explain why?
Most of Bulia's user interface decisions serve to reduce the amount of "modality" in the interface.
[ ...an exception being the auxiliary toolbar, but that's an extra convenience rather than a requirement to get to most features, and the remaining features are specific to the particular tool. ]
In general usability studies have shown that "modal" interfaces are harmful for usability.
Part of the reason for this is that modality places an increased burden on the user -- they need to maintain a mental model of the internal "state" of the application before they can know what a given action/keypress will do.
[ Visual indications (such as the selection indicators) mitigate this burden a little, but they do not remove it entirely. ]
Modal interfaces also place an additional burden on the user: the user needs to maintain a mental map of the state transitions necessary to return to a state where the desired functionality is available.
At that point, they still need to perform the complex sequence of actions necessary to arrive there.
[ Specifically, unselecting objects in order to pan, only to have to re-select them afterwards, would make the interface inordinately difficult to use. ]
In practice, any sufficiently complex interface will need to have modal elements, but it's important to minimize the modality and avoid making it too fine-grained.
-mental