1.4 KiB
Research initiation — 2026-10-12
User brief
Create content-explorer, a general-purpose browser for interconnected knowledge, influenced by onco.cc's excellent information presentation and MIT-licensed repository. Connect to arbitrary dataspaces by projecting varied inputs into an abstract model. Reuse and refactor OnCo's UI where possible. Deliver INTENT and research-backed implementation workplans as ZIP.
Initial synthesis
OnCo supplies valuable generic scaffolding: graph validation/backlinks, index/detail routes, a faceted EntityBrowser, and static data publication. Those elements still embed cancer-specific types, icons and branching. The project should pin an upstream version, separate rights-controlled datasets from reusable code, and extract a generic UI behind a manifest and snapshot boundary.
Selected design assumptions
Static-first MVP; Next.js/React/TypeScript continuity; canonical qualified identifiers, typed relation and evidence model; two small independent domains as a non-coupling test; initial local JSON and Markdown connectors; GitHub/Wikidata later. Claims, conflicts and license gates must not be erased by the normalized read projection.
Known limitations
This research accessed a number of original source files and the live site but could not clone and execute the complete upstream repository. Baseline reproduction is the mandatory first workplan, not an already completed task.