Reviving Praderas (Day 3) — Phase 2: making the site tell you where you are
Phase 1 stabilized core behavior: listing, search, pagination, and a useful sidebar. Phase 2 moved to discovery: finding content without menu overload, and understanding where to go next from each post.
This is product work, not only CSS. The workflow stayed the same: AI agent proposes and implements, human validates and prioritizes.
What Phase 2 meant in practice
- Four-item main menu — Home, Blog, Categories, About. No automatic "every
.mdgoes to navbar" behavior. - Categories page — each blog tag with a short Spanish description, approximate post count, and filtered-list link (
/tags/?tag=...). - Breadcrumbs across blog, search, tags, and posts to avoid navigation guesswork.
- Discovery at post end — related entries by shared tag, plus previous/next navigation by publication time.
The last point required server-side logic for maintainability.
Technical decision for maintainability
We added a focused plugin, plugins/50-BlogNeighbors.php, that calculates:
- previous and next post neighbors,
- related posts,
- category counts for category-index rendering.
Small plugin, no database, easy to evolve.
Scope honesty
Today focused on structure and comprehension, not every Priority 2 visual/polish task. Language unification and media upgrades stay in backlog and are explicitly documented in proposed-improvements.md.
Day 3 closing
The blog feels less like a folder of files and more like a place: you know where you are, what topic you are reading, and what you may want to open next.
Update (same day)
We also revised the homepage (content/index.md) to remove old messaging centered on ChatGPT as a direct reader "partner", because it no longer matched the project narrative. The current focus is the blog and repository workflow, with agents and tooling where appropriate and a human guiding direction.