Re: [g-a-devel]Nautilus menu labels

[David Bolter <david bolter utoronto ca>]

Mario Lang wrote:
> Bill Haneman <bill haneman sun com> writes:
>
>
> > Long long ago Peter Korn wrote:
> > ...
> > ...
> >
> > > Do you know what the AccessibleStateSet is for this menu?  (you can check it
> > > with at-poke)  Objects that are aren't "visible" and "showing" should not be
> > > rendered to the user via an AT.  If an AT is doing that, it is probably a
> > > bug.
> >
> > Actually there are some explicit exceptions, particularly for GOK. 
> > Since GOK is about providing a high-efficiency interface for users who
> > can't readily enter keyboard commands, GOK exposes certain kinds of
> > 'hidden' UI elements rather than forcing the user to simulate the
> > keyboard shortcuts for posting them.
>
>
> I think the same will be true at some point for blind related
> ATs.  We do not see the GUI anyway, so we really do not
> care about visibility.  If some functionality is easily accessible
> by activating a "hidden" object, it is IMHO preferable to
> having to navigate the desktop and manually do things which are
> visually oriented.
>
> I can not provide a concrete example for this yet, but
> I do have the strong feeling that this is true.

I agree Mario. Someone once said "the best interface is the kind you 
don't notice". The term "navigating", to me implies that one does notice 
it on some level. Sorry it seems I digress...  I think with interfaces 
there is always a trade off between learning curve and navigation. So 
your point is right on, that we should be able to bypass the navigation 
as we develop expertise.  For example in the case of menu items 
(essential program functions), it is difficult to map shortcuts to the 
regular static keyboard exhaustively, although it is done in some highly 
used, or common cases (e.g. CTRL-C), but we all know this so I'll be 
quiet now.

cheers,

David