Day 15 was a focused link audit and repair sprint. The session opened with a clean
restart — dropping context files one at a time to restore full project awareness before
touching anything. The new check_ai_tree.sh -a flag (built and deployed
this session) immediately revealed that the old index-only scan had been masking a large
class of false orphans — files correctly linked from hub pages like
project-diary.html that were never linked from an index.html.
With accurate data in hand, the session worked systematically through broken links by category — fixing wrong relative paths, dead CSS references, a renamed file, a case mismatch, and several nav link errors across multiple files.
The original script only scanned index.html files as link sources.
Files linked from hub pages like project-diary.html were incorrectly
flagged as orphans because no index.html linked to them.
Added -a (all HTML) flag — scans every .html file in
the tree as a link source. Hub pages, diary indexes, and content pages all count.
Scanned source files are excluded from orphan reporting since they ARE the
navigation structure.
Orphan count dropped from 46 to 5 on first run with the new flag. The 5 remaining orphans are all legitimate — unlinked addendum files and duplicate PDFs.
Early diary entries and working-with-ai.html contained a
No styles.css ever existed — the site uses inline styles throughout.
Removed from:
collaboration/building-in-public/2026-02-16_AM-PROJECT-DIARY.htmlcollaboration/building-in-public/2026-02-16_PM-PROJECT-DIARY.htmlcollaboration/working-with-ai.htmllearning/AI-architecture-stack.html linked to
../collaboration/AI-As-A-Collaborator.html — file was renamed to
AI-as-work-partner.html. Fixed both occurrences (lines 414 and 423).
learning/how-to-use-this-site.html linked to
AI-learning-plan-beginners.html — actual filename is
AI-Learning-Plan-For-Beginners.html. Fixed.
Multiple files in building-in-public/ linked to
project-diary.html without the required ../ prefix.
Fixed in:
2026-02-15-PROJECT-DIARY.html — nav links corrected2026-03-02-PROJECT-DIARY.html — 3 nav links corrected2026-03-07-PROJECT-DIARY.html — deployed fix via sedcollaboration/AI-as-work-partner.html — 2 nav links correctedSTATUS-REPORT-DAY4-10.html — partial fixes appliedcollaboration/index.html → added building-this-site.html carddeep-dive/index.html → added advanced-topics.html card + fixed date typolearning/index.html → added introduction-to-ai.html and Recommended_Learning_Path-AI-LLM.html cardscollaboration/building-in-public/2026-03-07-PROJECT-DIARY.html —
Day 13 diary entry was missing. Built from session knowledge and deployed.
main.html, link1.html, link2.html — scaffold
placeholders in early diary entries, not real filesAI-NEW.html references — symlink to AI-NEW.html, browser
follows correctly../TMA/AI/ directory href — browser handles fine2DO-LIST, PROJECT-STRUCTURE,
STATUS-REPORT-DAY4-10 internal doc path issues — low priorityai_websites.html links in AI_learning_plan_ai-x.html
— replace with anchor link to Quick Access section in AI-NEW.html.
Need to add id="quick-access" to that section first.Assess_Your_AI_Knowledge.html — add link from root
index.htmlDatabricks PDFs — decide: keep one, delete one, link or move~/backups/resources/ dir — needs index.html or consolidate
into learning/| Metric | Start of Day 15 | End of Day 15 |
|---|---|---|
| Orphaned files | 46 | 5 |
| Broken links | 34 | ~16 |
| Script capability | index.html only | all HTML (-a flag) |
The clean restart protocol continues to prove its value. Dropping context files one at a time gave a clean, shared picture before any work began. The -a flag discovery was the insight of the day — sometimes the tool itself is the problem, not the data it's reporting.
Systematic categorization of broken links (A through G) made the work manageable and kept us from chasing low-value fixes. Knowing what to leave alone is as important as knowing what to fix.
cmh — AI Education Project Day 15 — March 9, 2026
Next session: Day 16 — Remaining cleanup items, then deep-dive/ CLI tools pages.