troubleshooting#
Series: lecture-python-intro
File:
lectures/troubleshooting.mdAudit date: 2026-08-26
Corpus snapshot:
a12d17c0efCategories audited: writing, figures, links (JAX out of scope)
Overall score: 8.8 / 10
Priority: NONE
Score breakdown#
Category |
Score |
One-line note |
|---|---|---|
Writing |
8.5/10 |
|
Math |
N/A |
no mathematical content. |
Code |
N/A |
no executable code cells. |
JAX |
out of scope |
JAX rules target |
Figures |
9/10 |
|
References |
N/A |
no citations in this lecture. |
Links |
9/10 |
|
Admonitions |
N/A |
no admonitions, exercises or solutions. |
Issues#
Critical#
None found.
High severity#
None found.
Medium severity#
[qe-link-002] — Use doc links for cross-series references. Count: 1. Lines: 32. Example: raw link to python-programming.quantecon.org.
[qe-writing-001] — Use one sentence per paragraph. Count: 1. Lines: 65. Example: 2 sentences in one paragraph.
Low severity#
[qe-fig-005] — Descriptive figure names for cross-referencing. Count: 1. Lines: 52. Example: {image} without :name:.
[qe-writing-008] — Remove excessive whitespace between words. Count: 1. Lines: 65. Example: 2 spaces.
Strengths#
The page states its own scope in one sentence and never exceeds it: line 23, “This page is for readers experiencing errors when running the code from the lectures”, and the two sections that follow are exactly the two things such a reader can do - fix the environment (25-59) or report the problem (61-68).
27-30 makes the lectures’ maintained assumption explicit as a numbered pair - code should execute in a Jupyter notebook on a machine with the latest Anaconda - which is what turns a support page into a diagnostic: a reader who fails either condition knows immediately which one.
The remedies are ordered by cost to the reader: update Anaconda (34-37), reinstall it (39), update the external libraries (41-46), fall back to a remote machine (50-54), and only then report an issue (56). Nothing is asked of the reader before the cheaper fix has been offered.
50-54 pairs the instruction with the screenshot it refers to: the sentence names the “Launch Notebook icon available for each lecture” and the
{image}immediately below shows what that icon looks like, which is the one place on this page where a figure is doing real work.58-59 and 68 give two different routes for feedback - the issue tracker for reproducible problems, a mailto address for everything else - and say that feedback is wanted, which is the right note to end a support page on.
There is no maths, no code cell, no admonition and no citation on the page, and it does not manufacture any: the drafted report correctly scores four categories
N/A, and there is nothing here for the eight judgment rules to bite on.
Recommended actions#
Point the issue tracker at the right repository: line 63 links to
https://github.com/QuantEcon/lecture-python/issues, which is the intermediate series, while this page ships withlecture-python-intro- so a reader following the instruction files an intro-series bug in the wrong tracker. It should behttps://github.com/QuantEcon/lecture-python-intro/issues.Replace the raw cross-series URL at 32 with a
{doc}reference -{doc}`programming:getting_started`- which is the one mechanical Links finding on the page (qe-link-002).Put the two install commands in inline code at 45-46:
conda install -y quanteconand!conda install -y quanteconcurrently sit in plain prose, so the leading!that distinguishes the notebook form from the shell form reads as punctuation.Split the two-sentence paragraph at 65-66 into two blocks (qe-writing-001), which also removes the double space after “possible.” (qe-writing-008), and add the missing full stop at the end of 68.
Convert the
{image}at 52-54 to a{figure}with a:name:and a short caption such as “the Launch Notebook icon” (qe-fig-005), and give the path a leading slash -/_static/lecture_specific/troubleshooting/launch.png- to match the fifteen other image and figure directives in this series; only this file andlong_run_growth.md:34use the relative form.Leave the rest as it is. Two comparable support pages in other series returned no judgment findings, and this one should too: the page is 69 lines, one sentence per paragraph almost throughout, with no term to define and no result to visualise.