Review Request 109792: Update 'dim display' algorithm.

[Thomas Lübking]

On Samstag, 13. April 2013 12:37:56 CEST, Danny Baumann wrote:

> If 'dim' does not mean 'alter brightness' (to be lower), what 
> else does it mean? ;)

Implies degradation, subset of "alter brightness" but not equal.

> But your expectation of it is interesting
I expect it to do something notable (otherwise it's a pointless feature) that is not the same as turning off the screen (a feature that already exists)

> So basically your use case is 'I don't use it'? ;)
I "use" it as a warning that i'm now stearing too long onto something w/o telling the system that i'm still active.

> To expand on the use case I have: I often let my laptop 'do 
> things' (compile code, download stuff, do a 'yum update', wait 
> for an answer of somebody via IM) where immediate attention 
> isn't needed, but it needs to be somewhat monitored nevertheless 
> (does the code still compile or did it abort because of a typo?, 

Actually I have the system ping me for that (and have PM inhibited) - iff i actually compile on battery (and in general if i'm not doing anything else)
But that's not relevant, what i claim is that this

> to 50% of last brightness' allows me to cover that use case.

is moot.

The ability to actually read onscreen content depends on the actual screen content (contrast) and the environment (sunshine, night, panel glas) why i claimed that for a particular brightness state (not too bright, not too dark - juts readable) user interaction is required (maybe a light sensor, but all i've used are rather nasty than anything else)
So i absolutely doubt that this usecase could be reasonably covered by a preset configuration - that kicks in timer based.

> That's what Sebastian is suggesting, though. With current code, 
> this static value is '0% of maximum', Sebastian proposes to 
> change it to '5% of maximum'.

So if i start at "4% of maximum" (what's not "off" on my system) and the function kicks in, i get the screen brightened, despite it says "dim"?

-> qMin(current, qMax(current*ratio, safeBase))?

Cheers,
Thomas