Re: Helix stuff...

[aj facade dhs org]

> > These are traditionally understood to confuse users.  I hope this doesn't 
> >come across to harshly, but you have just described the "vi" window manager.  
> 
> Some people like vi.  :O  (Not me, I confess.)

I'm writing in it now.  I was also brainwashed by a HCI course that modal
is bad for newbies.  Anyone planing gnome-elvis?  (Manic Laughter...)

> >If you like I can give you a play by play analogy.  Escape sequences come 
> >naturally  to programmers, because we deal with them all the time.  But it 
> >confuses the innocent.
> 
> I don't think keyboard shortcuts for window manipulation are something the
> "innocent" use very often. Most people use the mouse for manipulating windows;
> power users might want something quicker. If so, they can take the time to 
> master a modal interface, as they would do with vi. It's no more confusing 
> than the "insert" key, as long as you have some visual indication of which 
> mode you're in.

Thats a good point.  Providing shortcuts for the power user is good
design.  I would still prefer a non-modal solution over a modal one,
but a solution is also better than none.

The new user who accidentally presses the command key need needs help
getting out.  Perhaps dropping out of command mode whenever a non-command
key is pressed would make things a little more natural in general.
Anyway, I guess this is going a little far off the lists topic.

> >Change the spec?  Build gconf into E?  Heh...  I guess I don't have a better
> >solution, but I have serious reservations about a modal WM.
> 
> The spec is being changed anyway. The "window manager key" could be indicated 
> by a property on the root window - no need to use GConf.

I'm all for relaxing the spec.  For example one shouldn't need keyboard
gymnastics to switch between windows.  That just won't sell.

Is there a wm-spec list I should be sending these comments to instead?

Adrian Johnston