[Usability] Re: The state of keyboard navigation

[Soeren Sandmann <sandmann daimi au dk>]

Calum Benson <calum benson sun com> writes:

> > e->a->c->d->a->... is not good, because then the user can't get back
> > to the original focus only by using f6 and shift-f6.
> 
> You're right, they can't.  But they will get back to E if they press Tab
> often enough, or directly by using control E's access key.  (And every
> control should have an access key).  This, I believe, is how Java would
> handle this case.

Ok, I'll try to look into doing it like that this weekend.

The rule would then be: "when focus is cycled into a pane, focus is
given to [the first/previously focused] widget, unless the pane itself
contains panes, in which case focus is recursively transferred into
these panes". This will be just fine in almost all cases, because if a
pane contains subpanes, it normally doesn't contain any other widgets.

> So, I think the F6 sequence of a->c->d->a is sufficient. The user will
> soon learn that pressing F6 *only* cycles focus between the panes in a
> window, and not anything outside those panes.  Sometimes it's better
> just not to try and be too clever  :o)

I still think that my suggestion is the correct thing to do. It is not
like it depends on fragile heuristics or anything, I just can't find a
way to implement it. The tests I did with a an evil patch to
gtkcontainer.c all worked as expected with the common panes and
nothing else inside a pane as a special case.


Søren