[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Jonas Thiem]

https://bugs.kde.org/show_bug.cgi?id=368902

--- Comment #8 from Jonas Thiem  ---
In that case I propose simply disabling it per default, and allow advanced
users to activate it.

This way, you can make a default configuration that focuses on making it easy
to use when enabled, and inexperienced users won't accidentally trigger it and
panic in the default state since it's turned off.

I would rather need to turn it on manually rather than being tricked into
thinking it is implemented poorly, and as a result being frustrated with the
desktop.

(of course this might just be my opinion, so feel free to ignore. The KDE
desktop in general has sometimes a few rough edges tho, so I would rather
recommend a config that is simple and allows simple changes, rather than trying
to cater both for the super user and panicing default user with the same
defaults and getting an IMHO weirdly half-usable in-between result that might
appear to be broke to the usern)

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Martin Gräßlin]

https://bugs.kde.org/show_bug.cgi?id=368902

--- Comment #7 from Martin Gräßlin  ---
I have seen users getting under real stress and panic when they activated the
top left edge by accident. That's why I say any change in that area requires
user testing.

What we also saw in the past is a strong difference depending on the input
device. On X11 we don't know what kind of device triggered the mouse input. On
Wayland we could change depending on input device and it's configuration. But
also that will require some user testing

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Jonas Thiem]

https://bugs.kde.org/show_bug.cgi?id=368902

--- Comment #6 from Jonas Thiem  ---
Clarification: a usability disaster for the given feature, simply because it is
not obvious this needs to be configured to work reliably so it strongly
suggests to be simply broken and buggy until the user manages to figure this
out - in my opinion a pretty unfortunate way to default configure a feature,
and not really reasonable way to choose a default configuration for it.

If you're that worried that some users don't want anything to happen if they
accidentally use that gesture because they don't like it, you should probably
simply not enable it per default in the first place instead of having a default
activation level that makes it semi-useless unless altered. At least then you'd
have a clear matched expectation that nothing happens, and in case a user
enables it without immediately noticing the need for detail configuration, a
clear matched expectations that it works reliably if enabled with no further
tweaking - and people can still configure the detailed thresholds to their
preferences if for some reason they liked the old
works-but-hard-and-effortful-to-trigger behavior that is the current default.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Jonas Thiem]

https://bugs.kde.org/show_bug.cgi?id=368902

Jonas Thiem  changed:

   What|Removed |Added

 CC||jonas...@web.de

--- Comment #5 from Jonas Thiem  ---
What do you mean, activation by accident?

The entire point is to have something happen if the cursor touches the corner,
not to have something happen if the cursor is slammed for a whole second with
notable effort (who wants to do that regularly for a mouse gesture anyway?
Hello Repetitive Strain Injury around the corner). This is what the UI suggests
the feature does, so this is what the default behavior should be geared
towards.

It is all about user expectations with the default values, and right now your
default values are chosen in a way that makes the feature seem brittle, almost
broken with a "detection" that works 5% of the time.

There is nothing wrong with making this configurable, and nothing wrong with
allowing users to configure it to require significant effort to do the gesture
to make it not happen that easily IF THEY LIKE THAT, but the current default
values are simply super unexpected and IMHO a usability disaster.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Martin Gräßlin]

https://bugs.kde.org/show_bug.cgi?id=368902

Martin Gräßlin  changed:

   What|Removed |Added

 Status|UNCONFIRMED |RESOLVED
 Resolution|--- |WONTFIX

--- Comment #4 from Martin Gräßlin  ---
The idea is to prevent activation by accident and due to that we decided for a
strong push into the corner. Any change in that area requires proper user
testing with a large sample base on different hardware (mouse, trackball,
touchpad).

The comparison to how it is in GNOME doesn't hold as the corner has a much
higher relevance in GNOME. In Plasma it's an addition but not a primary way to
interact with the system.

Anyway KWin passes complete control to the user through settings which mean
that every user can adjust to personal preferences and personal hardware
responsiveness.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Jonas Thiem via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368902

--- Comment #3 from Jonas Thiem  ---
Ok I tested some more, and I believe the problem is the default Activation
Delay.

It is way too high, and as a result you need to push with a lot of effort or it
will simply appear to be sporadically not working. If you compare it to GNOME 3
where this feature is much more central, a slight tip with the mouse is enough.
However, in KDE you need to consciously make either a very heavy push (which
you don't do if this gesture is part of your trained workflow where it becomes
a sort of non-attentive thing) or get frustrated with your gesture not being
recognized all the time.

Therefore, I highly recommend lowering the Activation Delay to 10ms.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Jonas Thiem via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368902

Jonas Thiem  changed:

   What|Removed |Added

 Status|RESOLVED|UNCONFIRMED
 Resolution|WORKSFORME  |---

--- Comment #2 from Jonas Thiem  ---
Reopening. It is *very often* pushed back even when it wasn't there for a long
time. It simply doesn't work reliably. Are you sure that is intended??

-- 
You are receiving this mail because:
You are watching all bug changes.

[kwin] [Bug 368902] Top-left cot corner doesn't work reliably when window extends to top-left corner

[Martin Gräßlin via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=368902

Martin Gräßlin  changed:

   What|Removed |Added

 Resolution|--- |WORKSFORME
 Status|UNCONFIRMED |RESOLVED

--- Comment #1 from Martin Gräßlin  ---
The cursor is pushed back and there is a cooldown delay before you can trigger
it again. Please go to systemsettings -> desktop behavior -> Screen edges and
adjust the settings to your personal preferences.

-- 
You are receiving this mail because:
You are watching all bug changes.