Review Request 122229: Smooth the battery remaining time with moving average

[Martin Klapetek]

> On Jan. 23, 2015, 9:08 p.m., Kai Uwe Broulik wrote:
> > File Attachment: Graph 1 - snapshot_J27264.png
> > <https://git.reviewboard.kde.org/r/122229/#fcomment277>
> >
> >     The values consistently seem to be too optimistic

I wouldn't say too optimistic. You can see the actual average of those is moreless constant, the moving average is slowly converging to it, then the actual value spikes up, down again and up again...I think it's quite ok. That's also why I'd like to collect more data from more batteries than just mine ;)


> On Jan. 23, 2015, 9:08 p.m., Kai Uwe Broulik wrote:
> > File Attachment: Graph 2 - chart_2.png
> > <https://git.reviewboard.kde.org/r/122229/#fcomment278>
> >
> >     What's that?

That's powerdevil :) I unplugged the power and that happened (and actually the spike was 3x higher but I cut it so it fits into the graph, this did not change any other values)

Maybe something to look into in powerdevil.


> On Jan. 23, 2015, 9:08 p.m., Kai Uwe Broulik wrote:
> > applets/batterymonitor/package/contents/ui/batterymonitor.qml, lines 57-59
> > <https://git.reviewboard.kde.org/r/122229/diff/1/?file=344393#file344393line57>
> >
> >     Seeing this makes me really wonder why this isn't on the C++ side in PowerDevil.

The actual presentation of remaining time is purely a UI thing, I'd like to keep powerdevil, the backend (for this), with the raw data and only "correct" them in the UI. But if you think otherwise, it can sure be moved there.


- Martin


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/122229/#review74628
-----------------------------------------------------------


On Jan. 23, 2015, 8:11 p.m., Martin Klapetek wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/122229/
> -----------------------------------------------------------
> 
> (Updated Jan. 23, 2015, 8:11 p.m.)
> 
> 
> Review request for Plasma.
> 
> 
> Repository: plasma-workspace
> 
> 
> Description
> -------
> 
> This makes the remaining time in the battery applet much more smoother and does not jump up and down by hours when running some short-time extensive tasks. Whenever the battery gets below 20%, the smoothing is relaxed and the averaged value is much more closer to the powerdevil value.
> 
> Couple graphs showing the difference are attached. As you can see, the blue is what powerdevil reports, the orange is the smoothed out value. The bigger graph then also shows battery percentage for better idea.
> 
> Obviously a 5.3 material.
> 
> At this point I'd like to get some more data to see if we need to tweak the alpha more. So please try out this patch, run with "plasmashell 2>&1 | grep movingAverage" (or use your other favorite filtering way), use your laptops casually or try also some heavy load. The output will then have all the values and it'd be awesome if you mailed them to me directly.
> 
> Thanks
> 
> 
> Diffs
> -----
> 
>   applets/batterymonitor/package/contents/ui/batterymonitor.qml 57801a9 
> 
> Diff: https://git.reviewboard.kde.org/r/122229/diff/
> 
> 
> Testing
> -------
> 
> Tried running for couple days and collected quite some samples, definitely looks better than the time jumping all over.
> 
> 
> File Attachments
> ----------------
> 
> Graph 1
>   https://git.reviewboard.kde.org/media/uploaded/files/2015/01/23/b4533731-40f3-4037-b87e-02a0888650d6__snapshot_J27264.png
> Graph 2
>   https://git.reviewboard.kde.org/media/uploaded/files/2015/01/23/0634bdd1-e717-4c6b-8130-b7fabd31e068__chart_2.png
> 
> 
> Thanks,
> 
> Martin Klapetek
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20150123/a01dc6a7/attachment.html>