Review Request 122505: Warn about brightness getting too low

Thomas Pfeiffer colomar at autistici.org
Tue Feb 10 23:02:40 UTC 2015



> On Feb. 10, 2015, 9:01 a.m., Mark Gaiser wrote:
> > I'm not quite sure if a user wants to see a warning message at all.
> > When i use my notebook in a dark environment i usually put the brightness all the way down (depending on the notebook).
> 
> Kai Uwe Broulik wrote:
>     As I stated above, "all the way down" can mean "completely off", which I wouldn't expect as a user. I've never seen any other device that does that, apart from some black and white seven-segment display calculators.
> 
> Mark Gaiser wrote:
>     Then it's device specific even!
>     - My notebook: all the way down = still visible
>     - My macbook: all the way down = off
>     
>     Can you detect that?
>     
>     You message does make sense if the lowest step = off. It doesn't if the lowest step is still on.
> 
> Martin Klapetek wrote:
>     Note that this is only for when you are dragging it by mouse - if you drag it all the way down and your screen goes black, there's no way to recover if your keys don't work. If your keys work and stuff, you probably never use the slider, so this for a minority of users. It's the same reason we ask confirmation when deleting a file.
> 
> Thomas Pfeiffer wrote:
>     Wait wait wait wait... Could it be that we've come up with an overly complex solution to a rather simple problem?
>     Actually, those devices which turn the backlight off at 0% brightness are the only ones doing it _right_. I always found it very weird when my screen brightness OSD said "0%" but I could still see things. 0% brightness means zero brightness means _dark_.
>     So why should the user even be able to set the brightness to 0% anyway?
>     Since turning off the backlight without turning off the screen doesn't make sense practically, there just should be no way for the user to set the brightness to 0%, period.
>     So let the slider start at 1% and don't allow the brightness to go zero neither via power management nor via brightness keys.
>     That solves two problems: Accidentally setting to zero _and_ that semantic bullshit of "0% brightness but I can still see stuff".
> 
> Martin Klapetek wrote:
>     > So why should the user even be able to set the brightness to 0% anyway?
>     
>     To save battery time when the screen is not needed *right now*, perhaps? I quite often compile things on my laptop when on battery, this can take up to 5 minutes and it's already quite a battery drainer, why the screen backlight should help it when it's not needed? I listen to music while cooking, screen backlight not needed. Etc etc.
>     
>     > So let the slider start at 1% and don't allow the brightness to go zero neither via power management nor via brightness keys.
>     
>     I disagree there. It's a hardware design after all, there's no reason the software couldn't/shouldn't take advantage of that. Also setting it to 1% does not really make a difference (not on my laptop at least), it's so dark it's useless, so I'd have to be higher, like 5% or 10%, which is...weird, I think.
> 
> Mark Gaiser wrote:
>     Quote: "So let the slider start at 1% and don't allow the brightness to go zero neither via power management nor via brightness keys."
>     
>     Please, no! I don't really care if 0% is a hardware flaw or design. We apparently are stuck with the fact that we have hardware behaving differently. The software should not limit prevent me to use my hardware at it's full capacity. If 0% in my case is still visible then so be it and that should be allowed just fine. If you forbid this then i expect quite some bug reports for that will flow in.
>     
>     If you want 0% to be off then you should buy hardware that obeys that.
> 
> Thomas Pfeiffer wrote:
>     > To save battery time when the screen is not needed right now, perhaps?
>     
>     Turning the screen off when it's not in use is a perfectly useful thing to do, but that is _not_ what a brightness slider is for. The brightness slider is there to allow users to set the optimum brightness for their current surroundings.
>     Check out your mobile phone or tablet. Regardless of which OS it runs, I am very confident that pushing the slider all the way to the left will _not_ turn off the backlight.
>     
>     With a sensible power setting, the screen will turn off after some idle period when on battery anyway. If we want to allow the user to turn it off manually, there should be a keyboard shortcut for it. It _must not_ be a button in the GUI, because then there would be now way to turn it on again because you could not see it.
>     
>     Turning off the screen via brightness slider doesn't only have the problem this patch is supposed to solve, but also the disadvantage that you have to find your optimal brightness setting again afterwards. If the screen is turned off by a shortcut or via power management, it should return to the previous brightness setting again after it's turned on again.
>     
>     > Also setting it to 1% does not really make a difference (not on my laptop at least), it's so dark it's useless, so I'd have to be higher, like 5% or 10%, which is...weird, I think.
>     
>     That's why other OSes display percentages but only graphical representations of relative britghness. Which makes sense given that the same percentage of brightness means very different things with different hardware.
> 
> Martin Klapetek wrote:
>     Btw. sometimes it's (just) a driver issue, if you're on Intel, try adding
>     
>     ```
>     Section "Device"
>             Driver      "intel"
>             Option      "Backlight"    "intel_backlight"
>     EndSection
>     ```
>     
>     to your xorg.conf (or equivalent), then you should get 0% == backlight off. Similar is with nvidia binary, that requires
>     
>     ```
>     Option  "RegistryDwords"  "EnableBrightnessControl=1"
>     ```
> 
> Martin Klapetek wrote:
>     > I am very confident that pushing the slider all the way to the left will not turn off the backlight.
>     
>     But you also don't have dedicated keys to control brightness like you do on laptop. How else would you turn in on if you've turned it off? On laptop, you have dedicated keys. If you try to turn it off using the slider, it warns you. Seems pretty sensible to me. Plus, on mobile phone you have a dedicated button to turn off your screen right now. I don't think many laptops have that same dedicated key...?
>     
>     >  but also the disadvantage that you have to find your optimal brightness setting again afterwards.
>     
>     If I'm skilled enough to turn my brightness off, I'm also skilled enough to find my optimal brigthness. This honestly is a non-issue, it could as well be used ad-absurdum to disallow the users any brightness changes at all.
>     
>     > there should be a keyboard shortcut for it
>     
>     We already do have rather simple way for it, 0% brightness means no brightness (as far as drivers allow). It's that simple. Why should the user learn another shortcut? (which would be just as undiscoverable as say krunner).
> 
> Thomas Pfeiffer wrote:
>     > But you also don't have dedicated keys to control brightness like you do on laptop. How else would you turn in on if you've turned it off? On laptop, you have dedicated keys. If you try to turn it off using the slider, it warns you. Seems pretty sensible to me.
>     
>     See, the problem is that we allow users to switch off the screen via GUI but they don't have a way to turn it on again via GUI (because, you know, they can't see squat). That makes zero sense.
>     If we allow users only to turn it off via keyboard, they can turn it on again via keyboard. Everything is fine.
>     
>     If a user doesn't have keys to turn the screen on again, it can only turn off via power management, which means it will turn on again on any input. Everything is fine.
>     
>     > We already do have rather simple way for it, 0% brightness means no brightness (as far as drivers allow). It's that simple. Why should the user learn another shortcut? (which would be just as undiscoverable as say krunner).
>     
>     Maybe I see what you're getting at: How about allowing setting brightness to 0 (= off) via keyboard brightness controls (because that way it can be turned on again as well) but not via the slider?
>     
>     There is a rule in design for safety:
>     1. Prevent users from performing a dangerous action
>     2. If 1. is not possible, use a warning
>     
>     This rule exists because warnings are never 100% safe, so they can only ever be "better than nothing".

> If I'm skilled enough to turn my brightness off, I'm also skilled enough to find my optimal brigthness. This honestly is a non-issue, it could as well be used ad-absurdum to disallow the users any brightness changes at all.

I meant that more as an argument for the convenience of having a way to directly turn off the screen instead of setting the brightness all the way to zero. It's exactly the same thing as with "mute" functions on volume controls. They do the same as setting the volume to zero, but
a) It's quicker
b) You get your previous volume setting back after un-muting


- Thomas


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/122505/#review75756
-----------------------------------------------------------


On Feb. 9, 2015, 10:25 p.m., Kai Uwe Broulik wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/122505/
> -----------------------------------------------------------
> 
> (Updated Feb. 9, 2015, 10:25 p.m.)
> 
> 
> Review request for Plasma and KDE Usability.
> 
> 
> Repository: plasma-workspace
> 
> 
> Description
> -------
> 
> Some graphics drivers, notably Intel, turn off the backlight completely when brightness reached zero, which is also in the spec (0 = off, 1 = very dim) but imho that's unexpected. To prevent the user from accidentally turnign the screen off, especially when keyboard brightness controls don't work, which sadly still happens quite often, the slider breaks free from the user's drag (by becoming disable for two (perhaps 1 is enough?) seconds, so we also catch the mouse wheel case) and displays a warning (which stays there until screen brightness is dialed up again).
> 
> 
> Diffs
> -----
> 
>   applets/batterymonitor/package/contents/ui/BrightnessItem.qml 546ab58 
>   applets/batterymonitor/package/contents/ui/PopupDialog.qml a2acf31 
> 
> Diff: https://git.reviewboard.kde.org/r/122505/diff/
> 
> 
> Testing
> -------
> 
> Works pretty well, I just realized I forgot the mousewheel-on-trayicon case. Also, I'm open to wording suggestions since it sounds more like "we suck, sorry about that". (Note in the screenshot I used the mouse wheel, hence the displayed 4% rather than 5)
> 
> 
> File Attachments
> ----------------
> 
> Screenshot
>   https://git.reviewboard.kde.org/media/uploaded/files/2015/02/09/8b585088-e33e-4862-9c46-207d06f566f1__dimwarning.png
> 
> 
> Thanks,
> 
> Kai Uwe Broulik
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20150210/d279f203/attachment-0001.html>


More information about the Plasma-devel mailing list