FLF epistack competition · submission type: Other (protocol / interop schema) · layer: Structure
Research ships as a sealed document. MIRAfying it makes a graph of signed, citable records — anchored on the actual data artifacts.
“There must be some medium for scientific exchange larger than raw data and smaller than a narrative.” — Matt Akamatsu
This submission is a protocol: an interoperability schema that lets different tools, labs and analysts write research into one shared graph — so their analyses compound instead of forking, without anyone being forced to agree on the interpretation. We’ve demonstrated on real scientific use cases and applied to examples for the contest (COVID-19 origins and “are eggs healthy?”), with a human-and-AI workflow described end to end. MIRA is an open architecture and a shared schema, co-developed at mira.science workshops for scientists and open science toolbuilders.
| Layer | MIRA’s stance | Where it lives here |
|---|---|---|
| Ingestion | touches — we MIRAfy real evidence with MIRA-extraction (verbatim source excerpts, dangling-edge reporting), but we don’t focus on ingestion. | Cases, Try it |
| Structure | focus — the main submission. The grammar, the artifact anchor, reified attributed relations, one format that links diverse subtopics without flattening them. | Everything, esp. the FAQ |
| Assessment | enables — we make the material assessable: convergence patterns, cruxes-as-Requests, a self-audit. | FAQ, Cases |
Current status: Shipped: the schema, two live public graphs on unrelated fields, the extraction tool (JSON-LD out), the KOI cross-org transport, and two round trips — PDF→JSON-LD→KOI and MyST→JSON-LD→viewer. Designed but not yet built at scale: cross-schema federation, the public ATProto layer, MIRA deployed across many organizations.
Start here — a ten-minute path
or jump to whatever you came forThe five numbered cards: learn the grammar, watch one result become a graph, see it do real work, get the protocol questions answered by name, then open the prototypes.