[Pm-utils] low power mode

[Matthew Garrett]

On Mon, May 01, 2006 at 02:29:31PM -0400, Peter Jones wrote:
> On Mon, 2006-05-01 at 19:17 +0100, Matthew Garrett wrote:
> >  rather than just having static configurations 
> > that mostly just serve to give the user more buttons to press, and 
> > "powersave" against "performance" just encourages the same sort of 
> > binary thinking that led us to the dire speedstep situation on Windows 
> > of "Oh look, my CPU goes really slowly now". Providing it doesn't cost 
> > us usability, we should *always* be saving power.
> 
> Well, that's the tricky bit.  Things like "should I have bluetooth on
> right now?" are really hard to do without a button somewhere.

Right, there's certain cases where specific functionality does have to 
be binary - both because of theoretical (can the hardware listen for 
incoming connections while in a low power state?) and practical (the 
button disconnects the bluetooth whether the user likes it or not) 
considerations. But that should always be seen as the worse-case 
scenario - if the user is on wireless, the wired chipset should be 
powered down to the point where it can only detect link beat, for 
instance.

> Except they haven't; I've got to use a menu to turn things on and off on
> my phone.  Likewise, it has a preferences menu for how long to wait
> before setting dimming (and later disabling) the screen.

And sometimes there are things that are going to be user preferences, 
like how long the screen takes to blank. But compare to the Nokia 770, 
which intelligently disables wireless, stops the CPU, turns off the 
screen and so on - it lasts /ages/ on a charge without ever actually 
entering a traditional suspend state. That's what we should be aiming 
for on the desktop, to the extent it's achievable with current hardware.

-- 
Matthew Garrett | mjg59 at srcf.ucam.org