NWB road network: is it worth wiring in? #184

Merged

viberfox-agent merged 1 commit from docs/180-nwb-not-worth-wiring into main

2026-08-29 18:54:30 +00:00

viberfox-agent commented

2026-08-29 18:54:27 +00:00

Collaborator

Closes #180

Merged by the autopilot (docs/direction.md) after every Actions job passed on the branch head.

Closes #180 Merged by the autopilot (docs/direction.md) after every Actions job passed on the branch head.

viberfox-agent added 1 commit

2026-08-29 18:54:27 +00:00

docs(direction): NWB is centrelines, and is not worth wiring in 4fd5cb2336

The frontier row promised "measured carriageways, and bridge and tunnel
geometry". Checked against the live PDOK service on 2026-08-29: both halves
are wrong. `wegvakken` is multilinestring, `hectopunten` is multipoint, and
there is no polygon anywhere in the service. Its 60-odd properties are
administrative — no width, no lane count, no clearance. The only
structure-adjacent field is `rel_hoogte`, an integer grade level, null in
969 of 1000 features over Groningen centre and 363 of 400 at Knooppunt
Oudenrijn.

The decision is taken under the goal at `docs/direction.md:10-16` — the
content comes from registries, and where none covers a place the world
degrades to OpenStreetMap. NWB *is* a registry, so the test is whether it
says anything OSM does not. On width it does not: the ribbon's width comes
from a constant table at `crates/geo/src/road_texture.rs:164`, and NWB
carries nothing to replace it. On structures it does not: `bridge-heights.md`
already gets the outline from Shortbread and a measured clearance from AHN.
On grade it does not: `grade-separation.md` records that the dangle rule
recovers the interchange case without an external `layer`, and that the
earlier expectation of needing one was wrong. `crates/geo/src/paving.rs:7-35`
is the precedent for the remainder — a more accurate outline is not on its
own an upgrade, and NWB's centrelines are strictly weaker than the BGT
polygons that lost that argument.

Volume, for the record: 1,392 features and 2.03 MB of raw GeoJSON over
~1.7 x 1.7 km of Groningen centre, in two requests. That is the order of the
heaviest thing the client streams today, spent on a second copy of
centrelines the MVT already delivers. Licence is CC0 1.0 off the landing
page's own rel="license" link, so terms were never the obstacle.

The frontier paragraph at `docs/direction.md:47-49` authorises recording this
here — "this one is not worth it" is a valid result — so the row moves out of
the candidates table into a new "Checked and not taken" table, which the
other seven rows will reuse. Nothing outside the frontier section is touched.

Refs #180

viberfox-agent merged commit 04fecc66e6 into main

2026-08-29 18:54:30 +00:00

viberfox-agent referenced this pull request

2026-08-29 18:54:31 +00:00

NWB road network: is it worth wiring in? #180

viberfox-agent referenced this pull request from a commit

2026-08-29 18:54:32 +00:00

Merge pull request 'NWB road network: is it worth wiring in?' (#184) from docs/180-nwb-not-worth-wiring into main