Re: Feature freeze imminent

On 27 Feb 2007, at 06:45, Tuomo Valkonen wrote:

> On 2007-02-26, Samer Abdallah <[email protected]> wrote:
>> Speaking of ideas for Ion4, if such a thing should ever come to pass,
>> has anyone come across/have opinions about acme (from plan9)?
>
> Sucks on the tabletop computer, being heavily pointing-device oriented
> in an application requiring textual input. Whereas combined use of the
> mouse and the keyboard can work great in applications like Blender and
> FPS games, it simply won't work in applications where you type a lot.
Agreed, the worst thing is *forcing* the user to switch rapidly between
mouse and keyboard, but not all applications involve a lot of text  
input.
I think it's good to have the option of doing as much as possible one
way or the other, because I find it helps combat fatigue and muscle  
strain
to be able to switch from pure keyboard use to, say, leaning back,
changing position, and doing things with the mouse for a while. If you
give people the option, they can choose whichever mode is most  
comfortable
for them *at that time*.

As far as this relates to ion, the keyboard-orientation aspect
is orthogonal to the tiled/tabbed window management aspect and there is
no reason that a window manager like ion can't extract the maximum  
utility
from the mouse. It's actually not too bad as it is, and the cfg_mouse  
script does
a bit more. The main problem is the overall lack of creative ideas  
(everywhere
that is, not specifically in ion) about how to do useful and powerful  
things
with the mouse without it being too fiddly and taking up a lot of  
screen space
with pointless and ugly widgets.

> Not with the standard two-handed keyboards, and I've yet to see a  
> better
> text input method.
>
> Such a "type/store commands somewhere" interface could, however, work
> nicely on a pocket book sized touch screen tablet (a real lightweight
> one, not a bloated pseudo-tablet of the hinged laptops) with hand- 
> writing
> recognition, where the keyboard would necessarily be too crappy to  
> have
> at all.
I tried morse code on my iPod linux once, which was ..um..  
interesting. At least
it doesn't use any screen space.

Some other ideas:
	http://www.strout.net/info/ideas/hexinput.html
	http://www.speedscript.biz/default.aspx
	http://www.inference.phy.cam.ac.uk/dasher/
A common problem with these is that they take up a lot of screen  
space, but I
think the gestural aspects of the first two along with the predictive
aspect of dasher could be quite promising, especially when the  
predictive
model could be specialised for different applications, eg English  
language text
vs coding vs terminal session etc.

Samer



>
> -- 
> Tuomo
>