ekklesiaandClaude Opus 5
6050144e5d
Records what was measured this pass and, as importantly, what was not. KGlobalAccel will accept a bare Escape and its component carries both a pressed and a released signal, so hold-and-cancel is expressible. What could not be measured here is whether a global shortcut takes the key away from applications, because key injection cannot reach the nested session. The mechanism says it does — KWin matches shortcuts in an input filter ahead of the focused surface — which would cost the classroom every ordinary use of Escape. That is inference, and it is written down as inference, with the thirty-second bench test that settles it. The recommendation is to change the gesture rather than grab the key an appliance uses most: hold-to-confirm on an on-screen control, which is ordinary Qt inside our own surface, or a modifier chord like the one the maintenance terminal just proved. Either keeps the deliberate ceremony without an input daemon. Co-Authored-By: Claude Opus 5 <[email protected]>
3.8 KiB
3.8 KiB
Screenshots from the classroom v2 experiment
All 1920×1080, captured with grim from a real running session in the
container lab. See ../classroom-v2-experiment.md.
The desktop
v2-idle-desktop |
the idle desktop: the wallpaper, one bar, nothing else |
v2-from-package |
the same desktop, produced by installing the .deb on a clean machine |
v2-dark-bar, v2-light-bar |
the bar in both themes |
v2-theme-compare |
the two, one above the other |
v2-testpage |
a maximised browser, edge to edge above the bar |
v2-wonderink |
Wonder Ink as an application, in its own profile |
Application state
state-1-closed |
both launchers closed: no indicator |
state-3-behind |
one active (teal), one running behind it (grey) |
state-4-minimized |
minimised — measured identical to behind, which is why minimise was removed as a route |
state-5-restored |
back |
v2-both-apps |
both applications, one button each, from the packaged install |
labelled-tasks |
the labelled task manager, evaluated and rejected: it shows changing page titles rather than application names |
Parent Call
pc-icon-vs-notifications |
the Parent Call mark beside Plasma's notification bell — the comparison that decided it |
pc-mark |
the mark at 32, 48 and 64px, beside the two marks it replaced |
pc-panel |
the whole bar: Parent Call left, applications centre, status right |
pc-ready |
the card, ready, in approval mode |
pc-waiting |
waiting for a volunteer to approve |
pc-sent |
Sent to sanctuary — never "displayed" |
pc-rejected |
not approved, offering Try again |
pc-off |
the coordinator's mode is off |
pc-nobroker |
the broker is unreachable |
pc-typed |
a name in the field |
icon-candidates-2 |
sixteen Breeze person icons at real panel size, none of which meant Parent Call |
DVD
dvd-panel |
DVD in the row with the applications, out of the status area |
dvd-cover |
the DVD launcher's tooltip on the packaged install |
v2-dvd-overlay, v2-layershell |
the loading cover: a gtk-layer-shell overlay under KWin |
Windows
v2-secondary-window |
a window a page opened, floating over an untouched browser, with its own close button |
v2-print-dialog |
the system print dialog, over the page that raised it, offering only Elmer |
v2-fullscreen |
fullscreen video: the whole display, panel covered |
v2-layershell |
a gtk-layer-shell overlay under KWin — what the DVD loading cover is made of |
v2-notification-crop |
a notification card, top right |
Administration
admin-terminal |
the maintenance terminal, opened by its chord's action |
admin-terminal-bar |
the bar with the terminal open, and after closing it: a task button while it runs, no launcher ever |
Reliability and lockdown
recovery-after-8-kills |
the panel after eight consecutive SIGKILLs |
recovery-after-kwin-kill |
the session after KWin was killed twice; the application window survived |
recover-panel |
the bar rebuilt from the layout script after its configuration was deleted |
lock-desktop-rmb |
right-clicking the wallpaper: no menu |
lock-panel-rmb |
right-clicking the bar: no menu |
lock-krunner |
KRunner asked for, and not there |
Failures, kept
| — | keyboard injection never reached the nested session, so every interaction in this experiment was driven by the pointer. It is why Parent Call has never had a real key pressed into it, and why that is the first bench test. |
m-closed, m-minimized |
the pair whose pixels were sampled to show that Plasma draws minimised and closed differently, but minimised and behind the same |