{"version":"1.0.0","sourceCommit":"5091f70608a9305cb7bbbf5bb292d60c4c9b361e","qualification":"Research map and mechanism-level bibliography. Citations contextualize ideas; they do not validate the GenOS implementation.","sections":[{"index":"01","title":"Runtime, state, and continuity","question":"How can agent work be resumed and inspected?","summary":"GenOS explores versioned state, snapshots, counterfactual branches, and recovery. Decisions retain their provenance so results can be compared, challenged, or resumed.","links":[["Agent runtime","01-concepts/runtime-agentique.md"],["Counterfactual workspaces","02-orchestration/workspaces-contrefactuel.md"],["Orchestration and branches","02-orchestration/orchestration.md"]],"externalBasis":"Convergent replicated state and versioned histories","externalAnchor":"Conflict-free replicated data types (Shapiro et al., 2011)","genosHypothesis":"Concurrent member writes can converge on shared versioned state under explicit merge operations and invariant checks.","status":"RUNNING","evidenceHref":"/runs","routes":{"en":"/en/research","fr":"/fr/research"}},{"index":"02","title":"Orchestration and topologies","question":"What organization fits a group of collaborating agents?","summary":"The work defines eight topologies, their contracts, capabilities, and limits: specialist teams, populations, communities, shared state, and execution across comparison worlds. Their presence in the runtime does not imply equal maturity.","links":[["Topology and capability contract","02-orchestration/topologies-et-capacites.md"],["Topology index","02-orchestration/topologies/README.md"],["Agent communication","02-orchestration/communication.md"]],"externalBasis":"Distributed scheduling and organizational coordination","genosHypothesis":"Each topology answers a falsifiable question — e.g. three heterogeneous trajectories beat one single trajectory net of cost — testable per topology.","status":"PROTOCOL","evidenceHref":"/benchmarks","routes":{"en":"/en/research","fr":"/fr/research"}},{"index":"03","title":"Memory, learning, and plasticity","question":"How can experience influence later decisions?","summary":"This research area covers episodic and semantic memory, retrieval, lessons, plasticity, and consolidation. Effects observed in controlled scenarios remain distinct from quality measured on real tasks.","links":[["Memory and learning","01-concepts/memoire-et-apprentissage.md"],["Neurobiology and plasticity","01-concepts/neurobiologie-et-plasticite.md"],["Autobiographical memory","02-orchestration/memoire-autobiographique.md"]],"externalBasis":"Human episodic and semantic memory systems","genosHypothesis":"Retrieval that combines similarity, recency, credibility, and success signals may improve long-conversation factual recall over a same-model raw-context baseline. The published LoCoMo result reports absolute F1, without that controlled comparison.","status":"INCONCLUSIVE","evidenceHref":"https://github.com/PISSARAW/GenOS/blob/5091f70608a9305cb7bbbf5bb292d60c4c9b361e/docs/06-qualite-preuves/benchmarks/locomo.md","routes":{"en":"/en/research","fr":"/fr/research"}},{"index":"04","title":"Epistemology and evidence","question":"What justifies a conclusion or promotion?","summary":"GenOS treats outputs as results to examine. The work covers uncertainty, abstention, provenance, independent evidence, and metric limitations, separating execution success from the validity of a claim.","links":[["Epistemology and evidence","01-concepts/epistemologie-et-evidence.md"],["Product completeness and status","03-reference/contrat-produit-et-completude.md"],["Quality and evidence","06-qualite-preuves/README.md"]],"externalBasis":"Probabilistic forecast scoring","externalAnchor":"Brier score (Brier, 1950)","genosHypothesis":"Deliberation with explicit ballots and Brier-scored calibration improves error detection instead of merely producing consensus.","status":"PROTOCOL","evidenceHref":"/benchmarks","routes":{"en":"/en/research","fr":"/fr/research"}},{"index":"05","title":"Evolution and verified development","question":"How can an agent evolve without confusing change with progress?","summary":"GVX and evolutionary research examine transformations, mutations, lineages, experimental nurseries, and promotion conditions. They emphasize external measurements, replication, and safeguards against circular self-evaluation.","links":[["Verified development (GVX)","01-concepts/gvx.md"],["GVX experimental nursery","02-orchestration/nursery-experimentale-gvx.md"],["Evolutionary research on selfhood","06-benchmarks/recherche-evolutionnaire-soi.md"]],"externalBasis":"Metapopulation ecology: local extinction and recolonization","externalAnchor":"Levins metapopulation model (Levins, 1969)","genosHypothesis":"Semi-independent groups under extinction pressure maintain capabilities through migration and recolonization.","status":"INCONCLUSIVE","evidenceHref":"/benchmarks","routes":{"en":"/en/research","fr":"/fr/research"}},{"index":"06","title":"Biomimetic systems and cognition","question":"Which biological ideas can become testable computational mechanisms?","summary":"The program examines memory, homeostasis, communication, perception, immunity, and collective coordination as engineering models. Analogies organize mechanisms; they do not prove biological properties or consciousness.","links":[["Computational biology","01-concepts/biologie-computationnelle.md"],["Epistemic immune system","01-concepts/adaptive-epistemic-immune-system.md"],["Vital agent organs","01-concepts/organes-vitaux-agents.md"]],"externalBasis":"Physiology, immunology, and collective behavior as design analogies","genosHypothesis":"Biological organization patterns translate into bounded runtime operations with explicit limits — inspiration first, validation per mechanism.","status":"HYPOTHESIS","evidenceHref":"/concepts","routes":{"en":"/en/research","fr":"/fr/research"}}],"primaryReferences":[{"id":"episodic-memory","authors":"Endel Tulving","year":1972,"title":"Episodic and semantic memory","venue":"Organization of Memory, pp. 381–403","url":"https://books.google.com/books?id=jGpKAAAAMAAJ","mechanism":"Episodic vs. semantic memory","relevance":"Introduces the distinction used as a vocabulary for contextual episodes and generalized facts; it does not establish that software records have human recollective experience."},{"id":"semantic-memory","authors":"John R. Anderson","year":1991,"title":"The Adaptive Nature of Human Categorization","venue":"Psychological Review 98(3), 409–429","url":"https://doi.org/10.1037/0033-295X.98.3.409","mechanism":"Associative retrieval and categorization","relevance":"A probabilistic account of categorization and retrieval from prior instances; GenOS retrieval scores are software rankings, not a cognitive model validated by this paper."},{"id":"memory-retrieval","authors":"John J. Hopfield","year":1982,"title":"Neural networks and physical systems with emergent collective computational abilities","venue":"Proceedings of the National Academy of Sciences 79(8), 2554–2558","url":"https://doi.org/10.1073/pnas.79.8.2554","mechanism":"Content-addressable associative memory","relevance":"Shows a dynamical network that retrieves stable patterns; it is a mechanism reference, not evidence that GenOS vector search reproduces neural memory."},{"id":"synaptic-plasticity","authors":"Guo-qiang Bi and Mu-ming Poo","year":1998,"title":"Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type","venue":"The Journal of Neuroscience 18(24), 10464–10472","url":"https://doi.org/10.1523/JNEUROSCI.18-24-10464.1998","mechanism":"Timing-dependent synaptic modification","relevance":"Reports timing-dependent changes in cultured neurons. A runtime weight update is an engineered rule and does not inherit the biological mechanism or its causal claims."},{"id":"stdp","authors":"Guo-qiang Bi and Mu-ming Poo","year":1998,"title":"Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type","venue":"The Journal of Neuroscience 18(24), 10464–10472","url":"https://doi.org/10.1523/JNEUROSCI.18-24-10464.1998","mechanism":"Spike-timing-dependent plasticity","relevance":"Primary experimental characterization of timing-sensitive synaptic change; it motivates the analogy but does not validate software STDP implementations."},{"id":"brier-calibration","authors":"Glenn W. Brier","year":1950,"title":"Verification of forecasts expressed in terms of probability","venue":"Monthly Weather Review 78, 1–3","url":"https://doi.org/10.1175/1520-0493(1950)078%3C0001%3AVOFEIT%3E2.0.CO%3B2","mechanism":"Quadratic scoring of probabilistic forecasts","relevance":"Defines the probability score now bearing Brier’s name. Calibration still requires a suitable set of resolved forecasts; a score alone does not establish general forecasting skill."},{"id":"neuron-model","authors":"Warren S. McCulloch and Walter Pitts","year":1943,"title":"A logical calculus of the ideas immanent in nervous activity","venue":"Bulletin of Mathematical Biophysics 5, 115–133","url":"https://doi.org/10.1007/BF02478259","mechanism":"Threshold logic networks","relevance":"Formalizes idealized all-or-none units as logical networks; the abstraction is not a full biological neuron model."},{"id":"foraging","authors":"Eric L. Charnov","year":1976,"title":"Optimal foraging, the marginal value theorem","venue":"Theoretical Population Biology 9(2), 129–136","url":"https://doi.org/10.1016/0040-5809(76)90040-X","mechanism":"Patch departure under diminishing returns","relevance":"Derives an idealized patch-leaving rule under stated assumptions. A software resource-selection heuristic is not empirically validated by the ecological model."},{"id":"ecosystem","authors":"Arthur G. Tansley","year":1935,"title":"The Use and Abuse of Vegetational Concepts and Terms","venue":"Ecology 16(3), 284–307","url":"https://doi.org/10.2307/1930070","mechanism":"Ecosystem as organisms and physical factors in interaction","relevance":"Introduces the ecosystem as an ecological system. The GenOS biome analogy is a software allocation model, not an ecological measurement."},{"id":"ecological-niche","authors":"G. Evelyn Hutchinson","year":1957,"title":"Concluding Remarks","venue":"Cold Spring Harbor Symposia on Quantitative Biology 22, 415–427","url":"https://doi.org/10.1101/SQB.1957.022.01.039","mechanism":"Multidimensional ecological niche","relevance":"Formalizes the niche as conditions under which a population can persist. A worker's task niche is a constrained software role, not a species habitat."},{"id":"metapopulation","authors":"Richard Levins","year":1969,"title":"Some Demographic and Genetic Consequences of Environmental Heterogeneity for Biological Control","venue":"Bulletin of the Entomological Society of America 15(3), 237–240","url":"https://doi.org/10.1093/besa/15.3.237","mechanism":"Colonization and extinction among habitat patches","relevance":"Original metapopulation model. GenOS regional groups borrow the population-of-populations structure without reproducing ecological rates."},{"id":"stigmergy","authors":"Pierre-Paul Grassé","year":1959,"title":"La reconstruction du nid et les coordinations interindividuelles chez Bellicositermes natalensis et Cubitermes sp. La théorie de la stigmergie : essai d'interprétation du comportement des termites constructeurs","venue":"Insectes Sociaux 6, 41–80","url":"https://doi.org/10.1007/BF02223791","mechanism":"Stigmergic coordination through environmental traces","relevance":"Introduces stigmergy in the study of termite nest reconstruction; shared software traces are a design analogy, not evidence of biological equivalence."},{"id":"swarm","authors":"Craig W. Reynolds","year":1987,"title":"Flocks, herds, and schools: A distributed behavioral model","venue":"Computer Graphics 21(4), 25–34","url":"https://doi.org/10.1145/37401.37406","mechanism":"Local rules yielding group motion","relevance":"Presents a computational flocking model. It supports discussion of local coordination rules, not a general claim that a multi-agent system is intelligent or robust."},{"id":"selection","authors":"John Maynard Smith and George R. Price","year":1973,"title":"The logic of animal conflict","venue":"Nature 246, 15–18","url":"https://doi.org/10.1038/246015a0","mechanism":"Evolutionarily stable strategies","relevance":"Formalizes strategic stability in evolutionary conflict. It does not prescribe a fair or safe software selection policy."},{"id":"embodied-control","authors":"Rodney A. Brooks","year":1991,"title":"Intelligence without representation","venue":"Artificial Intelligence 47(1–3), 139–159","url":"https://doi.org/10.1016/0004-3702(91)90053-M","mechanism":"Layered perception–action control","relevance":"Describes a robot control architecture grounded in direct perception and action. It is a robotics design precedent, not evidence for agent cognition or consciousness."},{"id":"epigenetics","authors":"David L. Nanney","year":1958,"title":"Epigenetic Control Systems","venue":"Proceedings of the National Academy of Sciences 44(7), 712–717","url":"https://doi.org/10.1073/pnas.44.7.712","mechanism":"Cellular epigenetic regulation","relevance":"Frames epigenetic control as regulation distinct from genetic sequence. GenOS contextual configuration is only an analogy and does not implement molecular regulation."},{"id":"genetic-structure","authors":"James D. Watson and Francis H. C. Crick","year":1953,"title":"Molecular Structure of Nucleic Acids: A Structure for Deoxyribose Nucleic Acid","venue":"Nature 171, 737–738","url":"https://doi.org/10.1038/171737a0","mechanism":"DNA structure and complementary pairing","relevance":"Reports a structural model for DNA and suggests a copying mechanism. GenOS genome files are software artifacts, not molecular sequences."},{"id":"distributed-ordering","authors":"Leslie Lamport","year":1978,"title":"Time, Clocks, and the Ordering of Events in a Distributed System","venue":"Communications of the ACM 21(7), 558–565","url":"https://doi.org/10.1145/359545.359563","mechanism":"Causal ordering in distributed systems","relevance":"Defines a partial order of events and logical clocks. A published run without event timestamps cannot be reconstructed into a causal timeline from this result alone."},{"id":"byzantine-agreement","authors":"Leslie Lamport, Robert Shostak, and Marshall Pease","year":1982,"title":"The Byzantine Generals Problem","venue":"ACM Transactions on Programming Languages and Systems 4(3), 382–401","url":"https://www.microsoft.com/en-us/research/publication/byzantine-generals-problem/","mechanism":"Agreement despite faulty participants","relevance":"Formalizes agreement under specified faulty-participant assumptions. A voting topology does not inherit Byzantine fault tolerance unless its protocol meets those assumptions."},{"id":"program-correctness","authors":"C. A. R. Hoare","year":1969,"title":"An axiomatic basis for computer programming","venue":"Communications of the ACM 12(10), 576–580","url":"https://doi.org/10.1145/363235.363259","mechanism":"Hoare logic and program assertions","relevance":"Sets out axiomatic reasoning about program correctness. A checked property is only as relevant as its specification and assumptions."},{"id":"security","authors":"Jerome H. Saltzer and Michael D. Schroeder","year":1975,"title":"The protection of information in computer systems","venue":"Proceedings of the IEEE 63(9), 1278–1308","url":"https://doi.org/10.1109/PROC.1975.9939","mechanism":"Protection principles and access control","relevance":"Analyzes design principles for protecting information in computer systems. It is foundational guidance, not an audit of this site's runtime boundary."},{"id":"fault-tolerance","authors":"Algirdas Avižienis","year":1967,"title":"Design of fault-tolerant computers","venue":"AFIPS Fall Joint Computer Conference, 733–743","url":"https://doi.org/10.1145/1465611.1465708","mechanism":"Detection and containment of computer faults","relevance":"Original fault-tolerance design work. It motivates software failure analysis but does not validate a GenOS pathology label."},{"id":"information-channel","authors":"Claude E. Shannon","year":1948,"title":"A mathematical theory of communication","venue":"Bell System Technical Journal 27, 379–423 and 623–656","url":"https://doi.org/10.1002/j.1538-7305.1948.tb00917.x","mechanism":"Messages, channels and noise","relevance":"Foundational communication model; GenOS messages require their own delivery and authorization evidence."},{"id":"distributed-snapshot","authors":"K. Mani Chandy and Leslie Lamport","year":1985,"title":"Distributed snapshots: determining global states of distributed systems","venue":"ACM Transactions on Computer Systems 3(1), 63–75","url":"https://doi.org/10.1145/214451.214456","mechanism":"Consistent distributed state capture","relevance":"Describes a snapshot algorithm under explicit assumptions; it does not prove a GenOS snapshot covers external effects."},{"id":"biological-morphogenesis","authors":"Alan M. Turing","year":1952,"title":"The chemical basis of morphogenesis","venue":"Philosophical Transactions of the Royal Society B 237, 37–72","url":"https://doi.org/10.1098/rstb.1952.0012","mechanism":"Reaction and diffusion as a pattern-forming model","relevance":"A primary mathematical biology model. GenOS graph composition borrows the concept of constrained form generation, not reaction–diffusion dynamics."},{"id":"biological-homeostasis","authors":"Walter B. Cannon","year":1932,"title":"The wisdom of the body","venue":"Kegan Paul, Trench, Trubner (London edition)","url":"https://wellcomecollection.org/works/avvwjxg6","mechanism":"Physiological regulation and homeostasis","relevance":"Original synthesis of homeostatic regulation; software feedback guards are an analogy, not physiology."},{"id":"biological-dormancy","authors":"David Keilin","year":1959,"title":"The problem of anabiosis or latent life: history and current concept","venue":"Proceedings of the Royal Society B","url":"https://doi.org/10.1098/rspb.1959.0013","mechanism":"Dormancy and latent life","relevance":"Biological basis for the cryptobiosis analogy; a persisted software process is not latent life."},{"id":"mobile-genetic-elements","authors":"Barbara McClintock","year":1950,"title":"The origin and behavior of mutable loci in maize","venue":"Proceedings of the National Academy of Sciences 36(6), 344–355","url":"https://doi.org/10.1073/pnas.36.6.344","mechanism":"Mobile genetic elements and mutable loci","relevance":"Primary biological observations for mobile hereditary elements; transferable software packages are a design analogy."},{"id":"cancer-clonal-evolution","authors":"Peter C. Nowell","year":1976,"title":"The clonal evolution of tumor cell populations","venue":"Science 194(4260), 23–28","url":"https://doi.org/10.1126/science.959840","mechanism":"Clonal variation and selection in tumors","relevance":"Biological model of tumor progression; uncontrolled software growth is only a metaphor."},{"id":"source-control","authors":"Marc J. Rochkind","year":1975,"title":"The source code control system","venue":"IEEE Transactions on Software Engineering SE-1(4), 364–370","url":"https://doi.org/10.1109/TSE.1975.6312866","mechanism":"Versioned source history","relevance":"Original source-control design, used here to contextualize versioned agent state rather than certify replay fidelity."},{"id":"transaction-model","authors":"Jim Gray","year":1981,"title":"The transaction concept: virtues and limitations","venue":"VLDB 1981, 144–154","url":"https://www.cs.utexas.edu/~dahlin/Classes/GradOS/papers/Gray81.pdf","mechanism":"Atomic transactions and their limits","relevance":"Primary database research; a GenOS SQLite integration must still state its own transaction and recovery boundaries."},{"id":"sqlite-architecture","authors":"SQLite project","year":2026,"title":"Architecture of SQLite","venue":"SQLite primary technical documentation","url":"https://www.sqlite.org/arch.html","mechanism":"SQLite compiler, virtual machine, B-tree and pager","relevance":"Primary implementation documentation for the specific engine; it is a technical source rather than a scientific experiment.","kind":"technical"},{"id":"experimental-design","authors":"Ronald A. Fisher","year":1935,"title":"The design of experiments","venue":"Oliver and Boyd","url":"https://books.google.com/books?id=-EsNAQAAIAAJ","mechanism":"Controlled comparison, randomization and replication","relevance":"Original experimental-design treatment; GenOS comparisons need matched inputs and a declared measurement protocol."},{"id":"gettier-problem","authors":"Edmund L. Gettier","year":1963,"title":"Is justified true belief knowledge?","venue":"Analysis 23(6), 121–123","url":"https://doi.org/10.1093/analys/23.6.121","mechanism":"Counterexamples to justified true belief","relevance":"Original philosophical counterexamples. They motivate keeping truth, support and provenance separate; they are not a software test of knowledge."},{"id":"procedural-skill","authors":"John R. Anderson","year":1982,"title":"Acquisition of cognitive skill","venue":"Psychological Review 89(4), 369–406","url":"https://doi.org/10.1037/0033-295X.89.4.369","mechanism":"Declarative-to-procedural skill model","relevance":"Primary cognitive model of skill acquisition. A stored GenOS procedure is not evidence of human-like skill learning."},{"id":"conscious-workspace","authors":"Stanislas Dehaene and Lionel Naccache","year":2001,"title":"Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework","venue":"Cognition 79(1–2), 1–37","url":"https://doi.org/10.1016/S0010-0277(00)00123-2","mechanism":"Neuronal workspace hypothesis and functional indicators","relevance":"A primary theoretical and empirical synthesis. It does not establish consciousness in a software workspace."},{"id":"clonal-selection","authors":"Frank Macfarlane Burnet","year":1957,"title":"A modification of Jerne's theory of antibody production using the concept of clonal selection","venue":"Australian Journal of Science 20, 67–69; original paper reprinted in Nature Immunology","url":"https://www.nature.com/articles/ni1007-1019","mechanism":"Clonal selection in adaptive immunity","relevance":"Biological basis for adaptive recognition and memory; GenOS threat policies do not implement lymphocyte selection."}]}