No "remaining time" option in battery monitor?

[Sebastian Kügler]

Hey,

On Tuesday, June 19, 2012 18:07:48 Viranch Mehta wrote:
> I read through the thread provided by Marco and I'd like to give in
> my bit of suggestion:
> 
> We clearly do not want to mislead the user by giving him a misleading
> number (the remaining time). I agree with Aaron that this will make
> the user trust his computer less. What I suggest is let the user know
> down right that the number might be inaccurate before he chooses to
> use the feature. So in the config ui, put a warning in the brackets
> beside the config check box: (Warning: the reported remaining time of
> the battery will be a predicted value, and hence may be inaccurate).
> Now when the user is aware that the remaining time might be misleading
> he will keep that in mind whenever he sees the information and whenever
> it is wrong, he won't blame his computer.

Config options don't work "conditionally", I'm against having the option in 
the UI. To make a small group of users happy, I'm OK with a config-file only 
option (as we had in the past releases).

This is also what we agreed on previously, and I don't see that input data has 
changed in a way that we should now change the outcome of that discussion.

It's not just about "inaccurate", it maybe entirely wrong. Unscientific tests 
on my notebook say that the battery might say 4 hours at one point, I start a 
compile job (or really any other program that actually uses the CPU), and I've 
got only one hour left. Understanding what influences the power consumption is 
nothing that we can expect from a user, most developers don't even have a firm 
grasp on it, and besides that, it depends on the hardware as well. There's no 
way we can even come close to a sensible prediction, hence the decision is the 
above.

> I'd generally not give my suggestion in a burning discussion, but since
> we all agree this is a good feature to have 

I don't think we ever agreed on that. It's impossible to get remotely 
accurate, so we agreed on not wanting this feature.

> and knowing that the only
> roadblock in shipping this feature is the inaccuracy of it, I thought the
> above suggestion might work out to "kind of" dissolve the road block,
> while relevant people can continue working on making the numbers
> more accurate.

I'm really against solving social problems (which this one is an example of) 
with config options. That will quickly lead us to KDE3-style UIs, and we've 
paid so much effort to get away from that...

By the way, the /Plasma Way/ of solving this: Write a battery monitor for 
Geeks, that includes this and other features, tada. You know yourself how easy 
that would be with the current QML implementation. (I even have some code, 
somewhere that puts the power consumption into a graph, could be some really 
cool geek / debugging toy, but really nothing for a default component. If the 
energy poured into this discussion would have gone there, everybody would be 
happy.)

Cheers,
-- 
sebas

http://www.kde.org | http://vizZzion.org | GPG Key ID: 9119 0EF9