Daniel John Murray · Science from within.
Play Inside — a constructed observer world →
Explore the six gates · Evidence atlas · Research library · Laboratories · Source
An observer with finite access cannot see everything about the world it measures. What must it remember to predict what happens next? Which actions remain useful before a boundary is reached? The Bounded Observer follows those questions through six gates: access, state, action, chart and boundary, geometry, and prediction.
The public site makes the models explorable through twelve interactive demonstrations. The repository holds the assumptions, capstone manuscript, executable checks, IDA Live prototype and claim-by-claim atlas. Each claim carries a status and a next test, so a mathematical consequence, a simulation, an empirical result and an open hypothesis remain distinct.
The central composition result is conditional: continuous, associative, strictly monotone composition with a neutral state admits an additive coordinate with a bound at an unreachable horizon. Boundedness by itself does not establish that law for every system, a particular geometry, or a general biological prediction.
| Start here | What you can inspect |
|---|---|
| Inside experience | Move, look and zoom in a constructed world; test what a single reading loses, chase a chart horizon and carry a compass around a loop. |
| Interactive site | Six gates and twelve demonstrations with their mathematical or empirical status shown. |
| Evidence atlas | Claim status, scope, sources, limitations and next tests. |
| Research library | The earlier 41-record manuscript inventory and prediction ledger, mapped into the six gates with dated source receipts. |
| Laboratories | Earlier experiments, negative results, instruments and the next tests they require. |
| Capstone paper | Bounded Composition and Its Horizons, with figures and the assumptions behind the result. |
| Code and checks | Reproducible implementations and tests of the stated mathematics. |
| IDA Live | An experimental instrument; physiological validity and added predictive value remain open. |
Earlier programme, research records and contributions
The earlier work remains available as a source record. The Boundedness Atlas contains a manuscript library, eight connected contributions, predictions and failure criteria, and a biology exhibit. Its 41 manuscript records include historical versions; they are not 41 independent discoveries. The programme direction and observer lens preserve the wider synthesis.
- Empirical Architecture preserves manuscripts, mathematical foundations, a theory atlas, proofs and executable certificates, counterexamples and corrections.
- Empirical Observatory preserves laboratory software and research records. The hosted interactive Observatory has restricted access; public source and a working hosted exhibit have separate release states.
- Recovery Lab compared 214 older female mice; follow-up coverage and the added-history improvement missed the declared gates.
- Virtual Cell Flight 01 and Flight 02 retain their failed held-out success gates.
- Cancer Lab contains an experimental recovery model and proposed tests. It does not establish a cancer treatment or clinical efficacy.
- Active Context is a runnable diagnostic. Its first comparison did not establish superiority over conventional dependency checking.
- IDA / StateAtlas preserves measurement and intervention prototypes. Hardware validity and useful added physiological prediction remain open.
Published model-specific results: Hormesis, Dose-Response and Glutathione homeostasis, Redox Biochemistry and Chemistry. The redox paper distinguishes kinetic calibration from a prospective cancer prediction.
Engineering contribution records: Ray Serve, NeuroGym, Pertpy, Graphiti history and Graphiti timestamps. Each pull request has its own review status; submission is not acceptance or validation of the wider framework.
The programme does not yet establish a complete theory of nature or a general predictive advantage across biology. Published results, conditional proofs, passing software checks, failed empirical gates and prospective hypotheses should be read at their own evidence levels. Research records retain their individual sources, versions and reuse terms.