Project Forge
A living scientific instrument.
Project Forge is the maintained public record of a scientific instrument — not a website about one. It treats an instrument’s digital presence as a small ontology of objects, events, and measures, published as durable, timestamped documents: a living, citable, evolving record of science itself. Not another dashboard. Not another documentation site.
The guiding conviction is that scientific infrastructure deserves a permanent memory, and that the most radical feature such a platform can offer is honesty — every fact stored once, carrying its provenance and time; every surface derived from that one truth; every claim traceable to its source.
Architectural principles
Project Forge is defined by a founding constitutional corpus, indexed in
docs/forge/README.md:
| Document | Title | Establishes |
|---|---|---|
| FORGE-001 | The Vision | The instrument, not the website — one calm, living, traversable medium. |
| FORGE-002 | The Principles & Ontology | Truth Over Illustration; Calm Is a Feature; the triad of Objects, Events, and Measures. |
| FORGE-003 | The Experience Bible | Invisible design; motion only when the world moves; the four-hundredth visit. |
| FORGE-100 | The Architecture | The unidirectional flow (Reality → Truth → Publication → Knowledge → Views → Readers) and the governing invariant. |
| FORGE-101 | Composition | The runtime, domain, and instance planes. |
| FORGE-102 | Product Roadmap | The evidence gates and sequence of work. |
| FORGE-103 | Development Standards | The implementation discipline for code, review, testing, and maintenance. |
The governing invariant (FORGE-100 §1.5): A fact is stored exactly once, in canonical form, with its provenance and time; everything a reader ever sees is a pure function of stored facts and the reader’s identity.
This repository
This repository is the Forge engine — the architecture, the constitution, the runtime’s canonical source, and the domain packages that let organisms exist. It contains no organism: no instance declaration, no publication, no organism-branded frontend. See FORGE-202 for the complete statement of what the engine owns and what an organism owns.
index.html # the engine's own landing page (not an organism's)
css/forge.css # the visual language
js/forge.js # boot sequence + scroll reveals (both optional)
docs/forge/README.md # the constitutional corpus index
forge-core/ # the engine: runtime, contracts, core capabilities
forge-hpc/ # the HPC domain package
instances/README.md # the organism registry
The page is intentionally minimal and durable:
- Pure HTML, CSS, and vanilla JavaScript — no frameworks, no build step, no package manager, no dependencies.
- No external runtime resources — no CDNs, fonts, analytics, or trackers.
It works when opened directly from the filesystem (
file://) and over the web. - Accessible and responsive — keyboard-navigable, screen-reader legible, reduced-motion aware, and fully readable with JavaScript disabled.
- No fabricated data — nothing is running yet, so no “live” readings are shown (FORGE-002 Principle 1, Truth Over Illustration).
Future work
This landing page is the first artifact, not the destination. Future work will expand Project Forge into the complete platform described by the founding documents: the canonical ontology, the truth and publication layers, the six worlds, universal search, and living knowledge — all engine capability, reused by every organism the engine composes.
Mjolnir, at the University of Copenhagen, is the first Forge organism, and
is currently in active development. It no longer lives here: as of
FORGE-201, Mjolnir was split into
its own repository, forge-mjolnir,
running at mjolnir.projectforge.dk. This
repository keeps the runtime that composes it
(FORGE-200) and the registry of
organisms it has produced (instances/README.md) —
never the organism itself.
The architecture is becoming reality.
The future should inherit more than our papers. It should inherit our instruments.