Appendix — feedback to the style guide & action-style-guide#
This audit doesn’t just score lectures — it also feeds back into the tooling and conventions it scores against. This page summarises that feedback. The source material (issue bodies, ready-to-merge rule entries) lives in the repo under contributions/.
How the pieces fit#
The QuantEcon manual style guide is the human-readable source of conventions.
action-style-guideis the tool that enforces them — a registry ofqe-*rules checked at PR time. This audit scores lectures against that registry (see the scoring spec).Where the audit found a convention in the manual that isn’t yet a coded rule, or a way to make the tool catch more, faster, it became feedback. Four issues capture it.
Four issues opened#
Issue |
What it proposes |
|---|---|
Encode 7 conventions the manual documents but the registry is missing (see table below). |
|
Extend the planned deterministic-checker scope from ~13 to 22 rules, prioritised by audit-measured frequency, with this corpus as regression-test data. |
|
A design discussion: should cross-series scoring / synthesis (what produced this report) live inside |
|
Offers this 299-lecture labelled corpus as test / evaluation fixtures for the tool. No action required. |
Proposed new rules (issue #18)#
Seven conventions documented in the manual but not yet in the registry. They appear throughout this report tagged (proposed). The five strong-evidence rules are recommended for adoption; the last two are weaker-evidence and may be deferred.
Proposed ID |
Convention |
Corpus evidence |
|---|---|---|
|
Write “IID”, not “i.i.d.” / “iid” |
~30 lectures |
|
|
~60 lectures |
|
Plain letters for distribution names ( |
~30 lectures |
|
|
~5 lectures |
|
Reference equations via |
moderate |
|
Braces |
limited |
|
Lowercase for densities/PMFs, uppercase for CDFs |
limited |
Ready-to-merge rule entries for each are in contributions/rule-drafts/.
Direct feedback to a lecture repo#
Two findings are build-breaking and were flagged for lecture-python.myst regardless of the tooling discussion (also on the front page):
divergence_measures.md:134—\begin{align}inside$$(breaks the PDF build).cross_product_trick.md:133— broken{eq}reference.
Status#
All four issues are open. Since they were filed, the broader style-guide program (coordinated in QuantEcon/project-style-guide, a private planning hub) has set a new direction: rules are being consolidated into a single QuantEcon/style-guide rule database, and action-style-guide is slated to be split and retired. Under that direction the rule-drafts/ here serve as transcription inputs for the consolidated database rather than a PR against action-style-guide. See contributions/README.md for the current status.