Building a Content Graph from Plain Slug References

Every reference field in Barkpark is just a slug string pointing at another document — and that single design choice is what turns a flat content store into a navigable content graph. In this post we walk through how the v1 reference model resolves a post's author, a media asset's collection, and a Paper's related documents into real graph edges, with no join tables and no GraphQL stitching. We cover why we chose plain slugs over opaque IDs, how the graph view became the dual of our Indx retriever seam, and what it means to query content by its relationships instead of its rows. If you've read the schema and content-graph docs, this is the engineering story behind them.