diff --git a/docs/superpowers/plans/2026-09-24-v1.6-understand-implementation.md b/docs/superpowers/plans/2026-09-24-v1.6-understand-implementation.md
index b00c378..e2368c4 100644
--- a/docs/superpowers/plans/2026-09-24-v1.6-understand-implementation.md
+++ b/docs/superpowers/plans/2026-09-24-v1.6-understand-implementation.md
@@ -526,7 +526,7 @@ results out, the drawing never modified.
 """Label parsers for foreign drawings: diameters, equipment tags, panels, electrical data.
 
 Pure functions over one string each - no ezdxf, no I/O - so every rule here is
-tested on the exact spellings the field drawings of 2026-09-24 used, and the
+tested on the spellings real drawings use, and the
 readers above it (the network, the takeoffs, the scale check) share one
 reading of a label instead of three.
 
@@ -830,7 +830,7 @@ def wiring_target(text: str | None) -> str | None:
 - [ ] **Step 4: Run it to verify it passes**
 
 Run: `uv run --frozen pytest tests/test_understand_labels.py -q`
-Expected: `41 passed` (the look-alike test is parametrized over the eleven letters). A failing spelling is fixed in `labels.py`, never by deleting its test: each one is a spelling a field drawing used.
+Expected: `41 passed` (the look-alike test is parametrized over the eleven letters). A failing spelling is fixed in `labels.py`, never by deleting its test: each one is a spelling real drawings use.
 
 - [ ] **Step 5: Write the failing test for the vocabulary**
 
@@ -1086,9 +1086,9 @@ The tables are vocabulary a drafter writes, each row with its language — not a
 """Multilingual keyword tables for reading foreign drawings: EN / TR / RU / NL / DE.
 
 A foreign drawing says what a layer carries in its name and in the texts on it,
-in whatever language the office that drew it speaks: the field P&ID of
-2026-09-24 named its piping layers in English and Dutch (``P_cipsupplyline``,
-``KLEPPEN``) and labelled its lines in Russian. These tables are what lets the
+in whatever language the office that drew it speaks: a P&ID may name
+its piping layers in English and Dutch (``P_cipsupplyline``,
+``KLEPPEN``) and label its lines in Russian. These tables are what lets the
 reader tell a CIP supply line from a product line without a human.
 
 They are **vocabulary, not a standard**: words a drafter writes, each with the
@@ -2665,7 +2665,7 @@ MSG
   plus additive keys that no pinned key changes: per run `"id"`, `"layer"`, `"diameters": {size: schematic length on the P&ID}`, `"pid_length_mm"`; per cable row `"phases"`, `"voltage"`, `"vfd"`; top level `"pid_insunits": 4`, `"copy_offset": [100000.0, 0.0]`, `"allowance": 0.20`, `"label_offset_mm": 300.0`, `"tag_positions": {"layout": {tag: (x, y)}, "pid": {tag: (x, y)}}`, `"scale": {"overall", "room_1", "room_2": {"pairs", "median", "within_10pct"}, "room_1_factor": 1.3}`. Also module constants `RUNS`, `LAYOUT_TAGS`, `ROOM1_TAGS`, `PID_ROOM2`, `LAYER_SERVICES`, `ALLOWANCE`, `LABEL_OFFSET`, and the helpers `manhattan(a, b)`, `rmst_length(points)`, `scale_statistics(pid, layout, tags)`, `pid_tags()`.
   Conventions later tasks rely on: `tags` is **sorted**; a layout tag's position is its TEXT insertion, which is also the centre of the equipment circle it labels (so "tag position" and "equipment position" are one point by construction); `layout_length_mm` is `None` for the status-C run (`PR4`, ending on `T105`, which only the P&ID has); `PR3` crosses a reducer, so it has two sizes — `"diameter"` is the one at its first tag (`Ø51`) and `"diameters"` gives both; the layout draws the plan twice, so each shared tag appears twice on it and is unique only within a copy (Review Focus 3).
 
-The truth, computed by the generator and worked by hand in the test comments: nine runs — `PR1` product Ø51 A 5500 mm; `PR2` product SMS51 B (a T-junction, an inline valve, an arc bend, one segment drawn twice) 13200; `PR3` product Ø51→Ø38 A (a reducer) 9000; `PR4` product DN20 C (to T105) —; `CS1` CIP supply Ø38 A 20700; `CS2` CIP supply 20x27 B (an inline valve) 8000; `CR1` CIP return Ø38 A 25700; `CR2` CIP return Ø38 A **drawn on the supply layer** 10100; `UT1` ice water DN25 A 7100 (a utility, outside the default takeoff scope). Cables: M11 → CP-1 6900 mm, 9 m; T102 → CP-1 14100, 17; M12 → CP-2 15800, 19; T202 → CP-2 8000, 10 (no power stated). Every name is invented; no value comes from a client drawing.
+The truth, computed by the generator and worked by hand in the test comments: nine runs — `PR1` product Ø51 A 5500 mm; `PR2` product SMS51 B (a T-junction, an inline valve, an arc bend, one segment drawn twice) 13200; `PR3` product Ø51→Ø38 A (a reducer) 9000; `PR4` product DN20 C (to T105) —; `CS1` CIP supply Ø38 A 20700; `CS2` CIP supply 20x27 B (an inline valve) 8000; `CR1` CIP return Ø38 A 25700; `CR2` CIP return Ø38 A **drawn on the supply layer** 10100; `UT1` ice water DN25 A 7100 (a utility, outside the default takeoff scope). Cables: M11 → CP-1 6900 mm, 9 m; T102 → CP-1 14100, 17; M12 → CP-2 15800, 19; T202 → CP-2 8000, 10 (no power stated). Every tag and every number is invented and no value comes from a client drawing; the layer and block names are ordinary drafting words in the languages the readers must handle.
 
 - [ ] **Step 1: Write the failing test**