Tracker for the native-speaker review of the French edition by @Honaminto. It replaces #3, whose 26-box checklist could not be ticked by its reviewer (read access only) and which is closed in favour of this project — see the decision on project-translation#19.
Background
This repository is a machine-generated first draft of the QuantEcon Python Programming lectures in French (Claude Opus 4.8, one pass, 364-term glossary, no human edits). The precedent that matters: the Persian review found meaning-inverting mistranslations that read perfectly fluently. A native reading is the only thing that catches those. There is no deadline — one lecture a month is genuinely useful.
Where we stand (verified 2026-08-21)
| Fact |
State |
| Lectures in the ToC |
27 (polars.md was added after #3 was opened, which listed 26) |
| Reviewed so far |
2 — numpy.md (#24) and python_by_example.md (#25), Emile's edits merged 2026-08-04 |
| Reviewer's access |
write pending — added to translation-team-french (push) on 2026-08-21, awaiting his acceptance of the invitation; was read before, contrary to what #3 said |
| Roster |
.translate/config.yml names editors.primary: Honaminto |
| Glossary questions |
the six in action-translation#78 (merged 2026-07-22 "review needed") are unanswered; Cours vs chapitre deliberately deferred in action-translation#241 |
How the review works
One sub-issue per lecture, below. Each is complete when a PR editing that lecture by @Honaminto is merged, or when he approves the lecture as needing no change — nothing is ticked by hand; the PR is the record. Edit the French directly (a branch in this repo, one PR per lecture), or leave line comments / suggestion blocks on a PR if that is easier. Priorities, in order: meaning errors; wrong technical terms; inconsistent terminology; unnatural French (academic register); anything in code comments that looks wrong. Please ignore spacing before : ; ! ? (fixed automatically, action-translation#81), errors in the English (raise upstream), and the translation: header block. A term that is wrong is probably wrong in several lectures — say so on the terminology sub-issue and it goes into the glossary once, for every lecture.
Work plan
| Phase |
Issue |
Work item |
| 0 — access |
#41 |
@Honaminto accepts the team invitation (write) and acknowledges the editor agreement |
| 0 — terminology |
#42 |
Decision: the seven open glossary questions (six from QuantEcon/action-translation#78 + Cours vs chapitre) |
| 1 — lectures |
#43 |
Review intro.md |
| 1 — lectures |
#44 |
Review about_py.md |
| 1 — lectures |
#45 |
Review getting_started.md |
| 1 — lectures |
#46 |
Review python_by_example.md — done: fr#25 merged 2026-08-04 |
| 1 — lectures |
#47 |
Review functions.md |
| 1 — lectures |
#48 |
Review python_essentials.md |
| 1 — lectures |
#49 |
Review oop_intro.md |
| 1 — lectures |
#50 |
Review names.md |
| 1 — lectures |
#51 |
Review python_oop.md |
| 1 — lectures |
#52 |
Review need_for_speed.md |
| 1 — lectures |
#53 |
Review numpy.md — done: fr#24 merged 2026-08-04 |
| 1 — lectures |
#54 |
Review matplotlib.md |
| 1 — lectures |
#55 |
Review scipy.md |
| 1 — lectures |
#56 |
Review numba.md |
| 1 — lectures |
#57 |
Review jax_intro.md |
| 1 — lectures |
#58 |
Review numpy_vs_numba_vs_jax.md |
| 1 — lectures |
#59 |
Review autodiff.md |
| 1 — lectures |
#60 |
Review pandas.md |
| 1 — lectures |
#61 |
Review pandas_panel.md |
| 1 — lectures |
#62 |
Review polars.md |
| 1 — lectures |
#63 |
Review writing_good_code.md |
| 1 — lectures |
#64 |
Review workspace.md |
| 1 — lectures |
#65 |
Review python_advanced_features.md |
| 1 — lectures |
#66 |
Review debugging.md |
| 1 — lectures |
#67 |
Review sympy.md |
| 1 — lectures |
#68 |
Review troubleshooting.md |
| 1 — lectures |
#69 |
Review status.md |
Phase 0 gates nothing technically — Emile can already open PRs from a fork, as #24/#25 show — but accepting the invitation is what makes in-repo branches and one-click suggestions work, and the terminology decision improves every unreviewed lecture at once, so both come first. Phase 1 runs in any order; intro.md and about_py.md are the most-read after the two already done. The two reviewed lectures are filed as sub-issues and closed immediately so the progress bar reflects work that has actually happened.
What does not need to change
The translations are not re-generated during the review. Glossary changes from Phase 0 flow into future syncs automatically; they are not back-applied to reviewed lectures by machine.
Sources
#3 (superseded), project-translation#19 (decision, 2026-08-21), plans/human-reviews.md WS5 (nominated audits, derived status). Repo state verified at main on 2026-08-21.
Merci beaucoup, Emile — questions and disagreement are very welcome.
Tracker for the native-speaker review of the French edition by @Honaminto. It replaces #3, whose 26-box checklist could not be ticked by its reviewer (read access only) and which is closed in favour of this project — see the decision on project-translation#19.
Background
This repository is a machine-generated first draft of the QuantEcon Python Programming lectures in French (Claude Opus 4.8, one pass, 364-term glossary, no human edits). The precedent that matters: the Persian review found meaning-inverting mistranslations that read perfectly fluently. A native reading is the only thing that catches those. There is no deadline — one lecture a month is genuinely useful.
Where we stand (verified 2026-08-21)
polars.mdwas added after #3 was opened, which listed 26)numpy.md(#24) andpython_by_example.md(#25), Emile's edits merged 2026-08-04translation-team-french(push) on 2026-08-21, awaiting his acceptance of the invitation; wasreadbefore, contrary to what #3 said.translate/config.ymlnameseditors.primary: HonamintoCoursvschapitredeliberately deferred in action-translation#241How the review works
One sub-issue per lecture, below. Each is complete when a PR editing that lecture by @Honaminto is merged, or when he approves the lecture as needing no change — nothing is ticked by hand; the PR is the record. Edit the French directly (a branch in this repo, one PR per lecture), or leave line comments / suggestion blocks on a PR if that is easier. Priorities, in order: meaning errors; wrong technical terms; inconsistent terminology; unnatural French (academic register); anything in code comments that looks wrong. Please ignore spacing before
: ; ! ?(fixed automatically, action-translation#81), errors in the English (raise upstream), and thetranslation:header block. A term that is wrong is probably wrong in several lectures — say so on the terminology sub-issue and it goes into the glossary once, for every lecture.Work plan
Coursvschapitre)intro.mdabout_py.mdgetting_started.mdpython_by_example.md— done: fr#25 merged 2026-08-04functions.mdpython_essentials.mdoop_intro.mdnames.mdpython_oop.mdneed_for_speed.mdnumpy.md— done: fr#24 merged 2026-08-04matplotlib.mdscipy.mdnumba.mdjax_intro.mdnumpy_vs_numba_vs_jax.mdautodiff.mdpandas.mdpandas_panel.mdpolars.mdwriting_good_code.mdworkspace.mdpython_advanced_features.mddebugging.mdsympy.mdtroubleshooting.mdstatus.mdPhase 0 gates nothing technically — Emile can already open PRs from a fork, as #24/#25 show — but accepting the invitation is what makes in-repo branches and one-click suggestions work, and the terminology decision improves every unreviewed lecture at once, so both come first. Phase 1 runs in any order;
intro.mdandabout_py.mdare the most-read after the two already done. The two reviewed lectures are filed as sub-issues and closed immediately so the progress bar reflects work that has actually happened.What does not need to change
The translations are not re-generated during the review. Glossary changes from Phase 0 flow into future syncs automatically; they are not back-applied to reviewed lectures by machine.
Sources
#3 (superseded), project-translation#19 (decision, 2026-08-21), plans/human-reviews.md WS5 (nominated audits, derived status). Repo state verified at
mainon 2026-08-21.Merci beaucoup, Emile — questions and disagreement are very welcome.