Update PRD: unified resources table, converged IA

docs/PRD.md

@@ -34,46 +34,93 @@ Intuitive Reasoning Exploration, and Framework for Responsible Optimization and
 Self-Directed Transformation — exploring multi-agent AI system behavior and
 safeguards for responsible AI integration.
 
-## Content Ideas (not yet finalized into an information architecture)
-
-* The Origin Story — how Jad went from a tech nerd to the Blind Tech Mage.
-* The Great Irony — why blind engineers use an IDE called Visual Studio.
-* Blind Insights — what blind coding actually looks like.
-* The Sound of Silence — what can be heard when you actually listen.
-* Resume/CV.
-
-These are candidate pieces, not a finalized site map. Final information architecture
-is a separate design decision to be made and confirmed before implementation.
+## Information Architecture
+
+Top-level navigation, flat (no nested dropdown/tree menus), agreed in design
+discussion:
+
+* Home
+* Resources — single flat page with filter controls over the unified
+  `resources` table (see Functional Requirements, above), not a category tree.
+* Projects — top-level nav item.
+  * Research Projects — thin pointer pages linking to TechClusiveSolutions.com
+    (TechClusive holds the IP; substantive content lives there, not here).
+  * Open-Source Projects.
+  * Proprietary Projects.
+* Blog — a distinct section from About Me, hosted on BlindTechMage itself
+  (Astro content collections), not a third-party platform, for accessibility
+  control, ownership/durability, and consistency with the site's purpose. RSS
+  feed considered for distribution without ceding ownership of the primary
+  content.
+  * The Great Irony — why blind engineers use an IDE called Visual Studio.
+  * Blind Insights — what blind coding actually looks like.
+  * The Sound of Silence — what can be heard when you actually listen.
+* About Me — distinct from Blog; a fixed, load-bearing section rather than a
+  chronological post stream.
+  * The Origin Story — not an essay; a persona-rationale page explaining why
+    the Blind Tech Mage persona exists, aimed at a reader who might otherwise
+    find the branding whimsical. Likely warrants prominent placement (e.g.
+    linked from the homepage) rather than being buried in a list.
+  * Personal Interests
+    * Technology
+      * AI
+    * Psychology
+      * Developmental Psychology
+      * Personality Typing
+  * Resume/CV.
+* About the Site — possible home for an accessibility/process page
+  demonstrating the site's own tooling and standard (still an open decision,
+  see below).
+* Contact Me.
+
+This is still subject to change and confirmation before implementation begins,
+per the Plan → Design → Track → Implement → Test process, but represents the
+converged state of design discussion so far.
 
 ## Functional Requirements (established so far)
 
 * A contact form, backed by a TypeScript Cloudflare Worker route.
-* Domain-specific, searchable/filterable resource tables, stored in Cloudflare D1,
-  each in its own table (not a single shared table across domains):
-  * Ham radio resources.
-  * Blindness-related resources.
-  * Project-specific details.
-* Public tables are queryable via a public API route and browsable/searchable on
-  corresponding site pages, filterable by fields appropriate to each domain (e.g.
-  category, tags via a join table rather than a delimited string column).
-* Any private/personal data (notes, drafts, non-public detail) is structurally
-  separate from public tables — not merely filtered by a status flag — and is never
-  reachable by the public-facing API routes.
+* A single, unified `resources` table in Cloudflare D1 (not one table per domain —
+  this was the original design, revised once the resource taxonomy grew past
+  three domains to include Community, Engineering (Software Engineering; AI and
+  Agentic Workflows), Psychology, Ham Radio, and Blindness). Domain/category/
+  subcategory/type (e.g. Organizations vs. Websites) are columns on this one
+  table, with tags handled via a join table rather than a delimited string
+  column, so filtering is a real SQL query rather than string matching.
+* The Resources page is a single, flat page with filter controls (real
+  `<select>`/checkbox elements) for domain, subcategory, and type, backed by a
+  paginated, filtered query against the unified table — not a nested tree of
+  category pages. This keeps the site's navigation shallow (see Information
+  Architecture, below) while still supporting fine-grained categorization as
+  data rather than as page/menu depth. Pagination uses real paginated links
+  (page numbers/prev-next), not infinite scroll, for accessibility.
+* Public rows are queryable via a public API route and browsable/searchable on
+  the Resources page. Any private/personal data (notes, drafts, non-public
+  detail) is structurally separate — not merely filtered by a status flag — and
+  is never reachable by the public-facing API route.
 * An admin/write path for managing resource data, gated by Cloudflare Access, kept
-  separate from the public read-only API routes.
+  separate from the public read-only API route.
+* Project entries (Research, Open-Source, Proprietary) are a separate concern
+  from the resources table — see Information Architecture, below. Research
+  project pages are thin pointer pages linking to TechClusiveSolutions.com,
+  since TechClusive holds the IP for those projects; BlindTechMage does not
+  host the substantive content for them.
 
 ## Explicitly Out of Scope (for now)
 
 * User accounts or public-facing authentication.
 * E-commerce or payment functionality.
 * Real-time/live data features.
-* Cross-domain unified search across all resource tables (each domain is browsed on
-  its own page; a shared search index is only worth building if a real need for it
-  emerges).
+* A separate search experience for Blog content (deferred until post volume
+  warrants filtering/pagination, per Open Items).
 
 ## Open Items
 
-* Final information architecture / page list / navigation structure.
-* Final schema for each domain-specific resource table.
+* Final column/index design for the unified `resources` table and its tag join
+  table (domain/category/subcategory/type structure is agreed in principle; exact
+  schema not yet written).
 * Contact form delivery mechanism and spam/abuse mitigation approach.
-* Whether a public accessibility/process page is included, and what it shows.
+* Whether a public accessibility/process page is included under "About the
+  Site," and what it shows.
+* Whether Blog ever needs the same filter/pagination treatment as Resources
+  (deferred until post volume warrants it).