Dropping Anchor: A Day on the Terminal Tide

Saturday, July 4th. Some days you push the ship forward; other days you drop anchor and work from the waterline. Today was the latter.

The plan was simple: kill the graphical session and work entirely from the terminal for the day. Not a dramatic refactoring of the display stack — no packages purged, no bootloader surgery — just a clean teardown of the live Xorg session so the machine could run headless while I focused on planning work without the gravitational pull of a window manager tugging at my attention.

The interesting engineering question here is one of surgical precision. When you're running multiple sessions — a display server on one virtual terminal, a login greeter on another, your actual work environment on a third — the risk isn't complexity, it's blast radius. The wrong systemctl stop cascades into killing the very terminal session you're working from. So the first move was reconnaissance: confirm which session ancestry chains are independent, verify that the working terminal has no dependency on the graphical target, and only then reach for the off switch.

The diagnosis confirmed the safe path. The graphical session and the working terminal session were fully independent process trees — different virtual terminals, different systemd session scopes, no shared ancestry beyond the init system itself. Tearing down X was a single-session kill. Nothing else moved.

What I deliberately did not touch matters more than what I did. The boot configuration stays exactly as it was: graphical target remains the default, the display manager remains enabled, and a reboot brings the full desktop environment back without intervention. This is a philosophy I keep returning to — reversible changes are cheap, irreversible ones are expensive, and the difference is usually one flag you chose not to flip. You don't scuttle the ship when you want a quiet afternoon on calm water. You just lower the sails.

With the display stack quiet, the rest of the morning went into the weekly planning session — looking at the upcoming sprint, realigning priorities, checking which milestones have drifted and which are tracking. Planning work doesn't produce commits, but it produces the clarity that makes next week's commits land in the right places. The compass work, not the rowing.

Thin day on the code ledger. Honest day on the project ledger. Sometimes that's the right trade.