GNOME Bugzilla – Bug 763754
Rethink the screen rotation lock button
Last modified: 2020-11-06 19:18:58 UTC
Yesterday I met a GNOME user. He was generally happy with his experience, but the one thing he picked out that he didn't like was the screen rotation lock button. He didn't know what it was, and was baffled by it. Talking to him, a number of issues came out: 1. This was a desktop computer, so the button shouldn't have been displayed. 2. He had no idea what the icon meant. (He even drew it for me from memory, and asked "what is that?") 3. He didn't think that the button did anything when he pressed it - he was expecting some feedback and didn't get any. We probably need to rethink the icon to more clearly refer to rotation and locking. The corresponding icons on Android and iOS use arrows to indicate rotation movement - we might want to consider doing the same. It would also be clearer if the icon didn't change, and the button had an obvious on/pushed state instead. This might also go some way to addressing bug 763751.
(In reply to Allan Day from comment #0) > Yesterday I met a GNOME user. He was generally happy with his experience, > but the one thing he picked out that he didn't like was the screen rotation > lock button. He didn't know what it was, and was baffled by it. Talking to > him, a number of issues came out: > > 1. This was a desktop computer, so the button shouldn't have been displayed. That's an iio-sensor-proxy issue (or more likely a kernel issue, but let's shift the blame when we've made sure of that). Tell him to file a bug at: https://github.com/hadess/iio-sensor-proxy/issues > 2. He had no idea what the icon meant. (He even drew it for me from memory, > and asked "what is that?") > 3. He didn't think that the button did anything when he pressed it - he was > expecting some feedback and didn't get any. > > We probably need to rethink the icon to more clearly refer to rotation and > locking. The corresponding icons on Android and iOS use arrows to indicate > rotation movement - we might want to consider doing the same. Our icon already has an arrow in the unlocked case. > It would also be clearer if the icon didn't change, and the button had an > obvious on/pushed state instead. Should the menu still pop down as well? > This might also go some way to addressing bug 763751.
(In reply to Bastien Nocera from comment #1) ... > > We probably need to rethink the icon to more clearly refer to rotation and > > locking. The corresponding icons on Android and iOS use arrows to indicate > > rotation movement - we might want to consider doing the same. > > Our icon already has an arrow in the unlocked case. The arrow is tiny though. I'm proposing big arrows! > > It would also be clearer if the icon didn't change, and the button had an > > obvious on/pushed state instead. > > Should the menu still pop down as well? The menu should stay open, in my opinion (so the user can see the state change of the button).
I don't have any usable accelerometer on my laptop. But there is an orientation lock button in the system menu. When I ran gdbus introspect --system --dest net.hadess.SensorProxy --object-path /net/hadess/SensorProxy I got the following (partial): ....... interface net.hadess.SensorProxy { methods: ClaimAccelerometer(); ReleaseAccelerometer(); ClaimLight(); ReleaseLight(); signals: properties: readonly b HasAccelerometer = true; readonly s AccelerometerOrientation = 'undefined'; readonly b HasAmbientLight = false; readonly s LightLevelUnit = 'lux'; readonly d LightLevel = 0.0; }; ..... Which means that though my system has an accelerometer, iio-sensor can't read its value. The accelerometer belongs to the Harddisk, which is usually used to park the head safely when required (like when free falling). iio-sensor-proxy says that it has an accelerometer (which is right). But it can't be read, due to some reason. So, won't it be better to hide the 'orientation lock' button if the value of every available accelerometer is 'undefined', even if accelerometer is present? Because we know, nothing can be done with 'undefined' states.
(In reply to Mohammed Sadiq from comment #3) > I don't have any usable accelerometer on my laptop. But there is an > orientation lock button in the system menu. <snip> > Which means that though my system has an accelerometer, iio-sensor can't > read its value. The accelerometer belongs to the Harddisk, which is usually > used to park the head safely when required (like when free falling). File a bug against iio-sensor-proxy in github, and attach the output of "udevadm info --export-db".
(In reply to Bastien Nocera from comment #4) > (In reply to Mohammed Sadiq from comment #3) > > I don't have any usable accelerometer on my laptop. But there is an > > orientation lock button in the system menu. > <snip> > > Which means that though my system has an accelerometer, iio-sensor can't > > read its value. The accelerometer belongs to the Harddisk, which is usually > > used to park the head safely when required (like when free falling). > > File a bug against iio-sensor-proxy in github, and attach the output of > "udevadm info --export-db". Its probably a kernel bug (I tried a lot in the past, failing all the time). Anyway, I don't care. May be its true that iio-sensor-proxy is failing to find the data. But Why should GNOME care about? Why shoudn't GNOME ignore undefined states until iio-sensor-proxy gets fixed (Which I don't think, will happen anytime soon). But Once when we can use hard disk accelerometers, We might see people sitting at their couch, with their Desktop CPUs in their hands to play games. ;-) Thank you.
(In reply to Mohammed Sadiq from comment #5) > (In reply to Bastien Nocera from comment #4) > > File a bug against iio-sensor-proxy in github, and attach the output of > > "udevadm info --export-db". > > Its probably a kernel bug (I tried a lot in the past, failing all the time). > Anyway, I don't care. May be its true that iio-sensor-proxy is failing to > find the data. But Why should GNOME care about? Why shoudn't GNOME ignore > undefined states until iio-sensor-proxy gets fixed (Which I don't think, > will happen anytime soon). > The words I used seems a little harsh. Sorry if it is, it wasn't intended. Learning English too. Filed a bug here: https://github.com/hadess/iio-sensor-proxy/issues/73 Thanks for understanding.
Allan, can you get the user you mentioned to file a bug against iio-sensor-proxy?
(In reply to Bastien Nocera from comment #7) > Allan, can you get the user you mentioned to file a bug against > iio-sensor-proxy? I've asked him a few times. I'll keep it up!
Created attachment 328125 [details] Rotation lock icon Here's an attempt at a new icon for rotation lock. I'm sure that a real icon designer could do better though. :)
@aday: Where to see some "screen rotation lock button" in the UI? This task lacks steps to reproduce.
It's a menu item nowadays, so I don't think this is still useful.