Orca on laptops.

[Benjamin Hawkes-Lewis]

There's nothing wrong with having sensible defaults, but I struggle to
believe there is a set of magic bindings which work on all hardware.
How about running a configuration program at Orca's first startup that
would take information about the user's keyboard as input (the Ubuntu
installer already includes a program for guessing keyboard layouts)
and generate a set of sane defaults based on that?

We would need two basic Orca sets: one for keyboards with easy number
pad access, and one for keyboard layouts without. The remaining
problem is what would be the Orca modifier key. How about a list of
potential Orca keys in order of preference, with the configuration
program presenting the first available key as default. How about
something like the following list:

Right Windows key
Left Windows key
Insert
AltGr/Right alt
Scroll lock
Caps Lock
"Access IBM" and similar keys
Special navigation keys (my Thinkpad has back and forwards keys for
web navigation)

It seems to me crucial to change Orca (and LSR) keys from the same
control panel that configures other Gnome key bindings in order to
minimize conflicts with core Gnome bindings for application
functionality and special character entry.

The question of the Orca keys is but a symptom of a much larger
problem. Key bindings, with their inherent tension between
supplementing and replacing the mouse, between interfaces accessible
and making them faster, are a naturally complex subject. In the case
of Linux, the horrible problem of key bindings is exacerbated both by
the sheer range of hardware that can run Linux and by the glorious
chaos of its software. Running Vim in a Gnome Terminal or a web
application in Mozilla  on Gnome with Orca enabled creates a serious
potential for conflicts between key bindings. And when you're
dependent on the keyboard, such conflicts threaten basic usability.

Despite its intrinsic difficulties, the problem of key bindings is
extremely unsexy. So if I were to say, for example, that we need a
control panel that, when the user changes a key binding, can
introspect the configuration files of key end-user applications for
conflicts, I am not optimistic I would attract many enthusiasts for
building it. Yet that is precisely what a humane interface for
configuring key bindings demands. If one baulks even at the idea of
coding a program that can juggle key bindings, is it fair to expect
human beings to do so? Surely a control panel that could resolve
conflicts between web access keys and Orca, or Vim and Orca is not
intrinsically impossible? The biggest obstacle would seem to be
gathering information about the modes and contexts in which console
application key combinations apply, but that's not exactly an
insurmountable problem.

--
Benjamin Hawkes-Lewis