Ideation — Large-Scale Fixes & Expansions

Drafted 2026-07-01. This folder is the third documentation layer for Queer the Stacks: extensive ideation on deep structural fixes and larger expansion bets that the existing planning documents do not already contain. It was produced from a fresh, full read of the source tree, tests, CI, and audit artifacts on main (HEAD a19142c), plus the research pass on the research-panel-and-roadmap branch.

How this relates to the existing documents

Document	Role	This folder's relationship
../ROADMAP.md	The original build spec (M0–M6), ADRs, gates	Referenced, never restated
../ROADMAP-FUTURE.md	The shipped expansion plan (N1–N6) + inline deferrals	Items cited by section (e.g. "ROADMAP-FUTURE B3") when an idea builds on them
../RESEARCH-ROADMAP.md + ../USER-RESEARCH.md (on branch research-panel-and-roadmap)	The 2026-06-30 synthetic-stakeholder pass (R1–R12, E1–E12)	Cited by ID (e.g. "R6", "E5"); this folder deliberately does not repeat any R/E item
../RESPONSIBLE-TECH-AUDITS.md + ../audits/	The responsible-tech framework instantiation	Every idea here is checked against the four hard guardrails

Rule followed throughout: NET-NEW only. Where an idea extends an existing roadmap item, the existing ID/section is cited and the idea states what it adds beyond it. Where an idea would collide with a guardrail (read-only sources; no egress of reading data; no Goodreads/gatekept catalogs; sourced descriptors, never author labels), the tension is named rather than papered over.

Contents

1. 01-deep-dive.md — current-state assessment from the code itself: architecture map, genuine strengths, structural debt actually observed (with file paths), and strategic position in the 21-repo portfolio.
2. 02-large-scale-fixes.md — FIX-01…FIX-14: deep structural fixes across correctness, data model, security, privacy, accessibility, performance, operability, maintainability.
3. 03-expansions.md — EXP-01…EXP-16 in three horizons: deepen the core, adjacent capabilities/audiences, transformative bets.
4. 04-impact-and-sequencing.md — impact × effort matrix over every FIX/EXP ID, dependency notes, a Now/Next/Later sequence that goes beyond the existing roadmaps, and the explicit list of items gated on humans, legal review, SMEs, or real data.

Honest framing

These are ideas for evaluation, not commitments. They were generated by one deep code-reading pass, without new user conversations, without running the system against a real library, and without consulting the data stewards whose terms several ideas depend on. Some observations (e.g. the demo-candidate wiring in app/view.py, the missing md5 join in ingest/unify.py) are verifiable in code today; impact estimates and horizon placements are judgment calls and say so. Nothing here should ship without the same gate discipline the repo already practices: auto-gates for what is checkable, dated committed artifacts for what requires human judgment, and honest deferral for everything else.