Japanese beginner quickstart — 2026-08-12
Decision: LOCAL_JAPANESE_ENTRY_PASS / PUBLICATION_PENDING
Candidate base:
36f8cad98397d570b0aef1968d42e1b434e74f41Remote mutations performed: none
Why this increment
The product candidate already had one no-argument terminal home, a read-only system doctor, one-command Docker and source demos, and a guided own-bag path. The README and maintained documentation exposed those routes only in English. A Japanese user still had to translate the choice, failure, and distribution boundaries before entering the same product workflow.
A read-only live review on 2026-08-12 found 837 Stars and an unchanged Draft PR
#427 at public head 3f4dd70cdc58ad421192559213cdee0bdc41eba8.
All nine reported PR checks pass. The two ndt_omp_ros2 rosdistro PRs remain
open with the package-lineage question unanswered. The local follow-up was not
substituted for that public state, and no GitHub comment, label, push, release,
or review reply was made.
Product change
docs/getting-started-ja.md is a deliberately short decision page rather than
a translation fork of the full manual. It exposes only:
- read-only installation and bag diagnosis;
- the fixed Humble/Jazzy Docker demo;
- the six-package source quickstart and dry-run;
- the canonical
lidarslam-map startown-bag path; - Docker own-bag execution from a checkout; and
- verified success, stable failure actions, and safe demo post-processing resume.
The README caption and Docs home link directly to the page. MkDocs navigation and the deterministic release bundle include it. The page states the 517 MB, 8 GiB, and approximately 30-minute first-demo expectations and explicitly says that GLIM-style PPA/package-manager installation is not yet available. Detailed English contracts remain authoritative, so maintenance does not require two complete manuals.
Synchronization contract
The docs entrypoint regression requires the Japanese page to retain the exact
canonical commands, read-only doctor boundary, read-only Docker bag mount,
mapping-free demo --resume, map_verify: PASS, resource expectations,
package-manager limitation, English canonical links, README/Docs discovery,
and release-bundle inclusion. This turns translation drift into a test failure.
Verification
| Check | Result |
|---|---|
| docs, release bundle, published-release, and install/upgrade regressions | 48 passed |
| complete maintained Python gate | graph: 1,429 passed / 13 skipped / 11 existing warnings; lidar_slam: 694 passed; 2,123 total |
| Python syntax/undefined-name safety lint | PASS |
git diff --check |
PASS |
mkdocs build --strict |
PASS with existing Material and unlisted-page notices |
| rendered Japanese page | title, doctor commands, and map_verify: PASS present |
| remote writes | none |
Honest boundary and next measurement
This change reduces Japanese command-discovery cost; it does not prove a lower completion time, create a binary package, answer the rosdistro review, publish the larger local candidate, or close any onboarding matrix row. After an exact candidate is public, one first-time Japanese operator should be timed from README arrival to a selected safe command, then through the same neutral onboarding contract. Findings must repair the canonical workflow, not create a Japanese-only execution path.