Goal: by the end of this lesson you can narrow the Explorer to one kind of entity with the kind chips, and read everything about a selected fn — bindings, runs, versions — in the right-hand Inspector without opening a single popover.
Concepts introduced: the kind filters (the chips under the filter box), kind markers, and the Inspector panel with its four tabs.
Under the Explorer’s filter box sits a row of kind chips:
◍ all λ fn T types 🔒 secrets ⚙ services 2 ▣ apps 1 ✓ tests 3
✕ failed 1 ⚠ type errors 2 ⚐ lint 1
+ filter
Click a chip and the tree narrows to rows of that kind — services under services, type definitions under types, and so on. A chip is a filter, not a search: the tree keeps its namespace shape, it just hides rows that don’t match. A second chip adds to the selection (services or apps); clicking a pressed chip removes it; ◍ all clears every filter at once — the one gesture that always brings the whole tree back.
Details worth knowing:
fx marks chip is a detail toggle, not a filter:
it marks every fn whose execution carries effects with a small
fx (hover it for the exact footprint — io, time, …)
without hiding anything. Off by default; flip it on to see the
footprints, off again and the tree returns to normal.· N✗); the apps chip
appears only when the deployment has app routing at all
(Lesson 30).2 filters) — Lesson 22 makes it
a named view.The second row — ✕ failed, ⚠ type errors, ⚐ lint — filters on what needs attention rather than on what a row is; it composes with the kind chips (a failing fn is still a fn). The counts, the marks on rows and cards, and where each mark leads in the Inspector are lesson 19’s subject.
Kinds are the fixed half of the row. + filter opens the rest — only these namespaces, only what is built on this fn, only fns with this effect, unused — each of which appears as its own chip with an ×. Those, and saving a set of chips as a view, are lesson 22.
The < in the Explorer’s header (or Space e) collapses the whole
panel to a slim tab at the left edge — it still names the branch you
are on; click the tab (or Space e again) to bring the panel back. A
tour step that needs the Explorer expands it for you.
⋯ on a card → the ns badge shows the fn’s namespace path and
offers Reveal in Explorer — the tree unfolds to the row and
selects it. The ↗ beside it opens the fn in a new tab.
Click any fn — in the tree or any node on the graph canvas — and the right-hand Inspector panel shows it. Four tabs:
The Inspector’s head (name, namespace, description) renders instantly from what the editor already knows; all four tabs, Overview included, are fetched from the server when you open them, so they’re always current.
Two more reading affordances live outside the Inspector:
The cards themselves have one setting: Settings → Appearance → Graph cards. Compact hides the type / effect strips on every card until you reveal them — the same facts stay one click away in the Inspector, and a large graph gets noticeably calmer.
The Inspector is the “read” side of the editor: popovers are for acting (edit a binding, run, publish), the Inspector is for understanding what’s in front of you.
Prefer to be shown? This lesson exists as a guided in-editor tour: open the demo with the tour running (no sign-up), or pick “Interactive tutorial” in the editor’s account menu.
const (core.logic) and open Overview: the Used-by
section lists the crowd of fns that pin constants through it,
“Extended by” first. Click a row — the editor jumps there.tutorial-page-attrs /
tutorial-row-attrs) and click ⚐ lint: the tree collapses to
the two rows, each with ⚐1, their namespace row with ⚐ 2, the
chip with 1 — one finding, two members. Delete one of the pair
and the chip reads nothing.Lesson 21 — Working without the mouse: everything this lesson did with clicks — find, read, walk the graph — from the keyboard.