Clipped from uni-versum.me on 2026-03-30. 54 pages consolidated: 8 narrative argument pages, 3 vocabulary index pages, 12 schema vocabulary terms, 29 domain vocabulary terms, plus 2 JSON-LD indexes.
“You know more than your tools can hold. The connections between your projects, your collaborations, your references, your ideas: you carry them in your head because no tool captures them.”
Your information is scattered across dozens of systems that do not talk to each other. The only thing connecting them is you.
Uni-Versum is an architecture for solving this. Not another app; a pattern. A personal system of interconnected workspaces, organized from your perspective, where relationships have meaning and boundaries protect focus without destroying connections. Sovereign, distributed, connected.
The name is Latin: one verse, one turn of the gaze. Your knowledge, your relationships, your work, seen from your point of view.
You understand that an email from last Tuesday connects to a conversation from three months ago, which relates to a project in a different application. You comprehend these connections. Your mind retains them. Your applications cannot.
The Flood: Early personal computing involved manageable digital volumes. Then multiplication occurred — social platforms, cloud storage, communication apps. By the 2020s, every application generates content exceeding human processing capacity.
Fragmentation: Volume is not the essential problem; fragmentation is. Information scatters across systems incapable of mutual communication, connected solely through human presence.
Data, Information, Knowledge: Almost every tool marketed as a “knowledge management system” manages information. Knowledge demands three elements absent from existing applications: (1) semantic relationships — typed connections bearing implications, (2) a point of view — understanding is perspectival, (3) boundaries preserving concentration without severing links.
The Imposed Perspective: Every system shapes how you contemplate your information. File systems enforce hierarchy. Gmail offers labels. Notion supplies databases. Each imposes frameworks. What if systems adjusted to your perspective rather than the inverse?
The Sovereignty Problem: Your information does not belong to you. Email resides on Google infrastructure. Notes occupy Notion servers. You function as tenant, not proprietor. True data sovereignty means documents reside on your hardware, in accessible formats. But sovereignty over meaning — connections, perspective, organization — is what must be constructed.
The file system: A hierarchy. One structure, one path per document. No concept of relationships, types, or meaning. A library with no catalog and no librarian.
The SaaS era: Gmail, Google Drive, Trello — each handles its domain well. The trade-off is sovereignty. Each service optimizes for its own domain with no incentive to connect with others. Integration (Zapier, IFTTT) is plumbing, not architecture.
The “second brain”: Obsidian, Logseq, Roam — local files, bidirectional links. Local files solve sovereignty. But bidirectional links do not solve the connection problem. A link says “these are connected” without saying how, why, or what follows. A graph of untyped links is a map with roads but no labels.
PKM tools inherit the oldest assumption in personal computing: the tool defines the structure.
The pattern of failure: Every failed solution follows the same pattern: (1) recognize a real problem, (2) solve it within the tool’s boundary, (3) ignore everything outside that boundary, (4) force the human to be the integrator.
What is missing is a layer that sits across tools and above information: a layer that captures relationships, respects perspective, and maintains boundaries.
What if the architecture started from the human instead? Not “here is a tool, adapt to it,” but “here is a person with work, relationships, and a perspective; build the system around that.”
The watershed: A watershed is not a canal. A canal is engineered; a watershed is emergent. The terrain shapes the flow, and the system adapts because the structure is the adaptation. Uni-Versum is organized like a watershed: the force that connects things is gravity — more shared context, more pull.
Four pillars:
Workspaces with boundaries (satelles): Autonomous but part of a system. Types: communal spaces (insula), focused workshops (taberna), warehouses (horreum). Each workspace orbits the person at center.
The map (census): Aggregates what exists and shows relationships. Two faces: human-readable and machine-readable. Captures gravitational force — more shared context, stronger pull.
Communications (tabellarium): Routes external channels to workspaces. The boundary between automatic and human-reviewed is a trust decision.
A point of view (nos): The person at center defines the organizing principle. Agents extend perspective without replacing it. Each agent operates within a specific workspace with that workspace’s context and no more.
“The shift from ‘the tool defines the structure’ to ‘the person defines the structure’ is the single most important design decision in this architecture.”
A cell’s defining feature is not what it contains but its membrane — a selective barrier, not a wall. Workspace boundaries determine what passes through, who makes decisions, and internal organization.
The four pillars replicate at workspace level (fractal pattern): sub-spaces within a workspace, internal entity mapping, workspace-specific channels, contextualized perspective.
The fractal architecture enables distributed design: each workspace maintains its own knowledge. Problems remain localized.
Each workspace is a specialized organ. A library processes published knowledge (digestive system). A collaboration manages external exchange (respiratory system). A journal integrates signals (nervous system). The human is the circulatory system.
Trust at the boundary: Explicit agreements about what each participant can do + which channels connect the workspace to outside. Security by architecture — the workspace is an airlock.
Crystallization: When an idea accumulates enough mass, it crystallizes into its own workspace. A new workspace at any scale follows the same pattern.
The central map becomes leaner: it registers workspaces and trusts each to manage its own contents. Each workspace is authoritative about its contents; the central map is authoritative about system-level relationships.
A link says “these two things are connected.” It does not say how, why, or what follows. The shift from links to relationships is the shift from information to knowledge.
Why words matter: “Workspace” means different things in different tools. The ambiguity is the reason integrations fail. When a tool does not define its terms, it is not being flexible; it is being opaque.
A shared dictionary: The vocabulary is a structural layer. Every entity has a type from the vocabulary. Every relationship has a named term. “A is connected to B” becomes “A funds B” or “A inhabits B” — each carrying different implications.
Why Latin: English words carry too much baggage. Latin terms arrive clean. When you encounter satelles, you have no preconceptions. Same principle as scientific nomenclature.
The machine layer: A vocabulary that machines can also read is an ontology. The system can answer: “Which collaborations are losing momentum?” or “What am I neglecting?” — questions requiring perspective, context, and relationship traversal.
Precision as freedom: “When the structural terms are well-defined, everything else can be loose.” The vocabulary governs architecture, not content. “It is the difference between a city’s zoning plan and the conversations people have in their houses.”
What changes: You work in focus. Connections survive without you. Collaborators have clear boundaries. Tools work together without integration. Questions get real answers. The system grows without collapsing.
One verse: Not “everything in one place” (the canal) and not “everything connected to everything” (the mess). One coherent system from one perspective, organized by gravity.
Personal means sovereign: Lives on your machine. Open formats. No company owns it. The architecture would survive if every tool disappeared.
A pattern, not a product: Current implementation uses Obsidian, Markdown, JSON-LD, automation scripts, and a graph as primary interface. But the pattern is independent of tools.
What comes next: Workspaces exist with boundaries and types. The registry maps entities. Two vocabularies published. Communications agents import and route. Still being built: full fractal implementation, formalized trust agreements, advanced routing, specialized vocabularies.
Infrastructure vocabulary for the schema layer. Defines the meta-types that all other vocabularies use. Maps to SKOS, RDFS. Serves as the @context foundation for JSON-LD serialization.
URI namespace: https://uni-versum.me/vocabularies/schema/v1/
class (rdfs:Class): A class defines a type of entity: a category of things that share a common nature. Examples: satelles, agens. Self-referential — class is itself a class. Maps to rdfs:Class.
concept (rdfs:Class): An abstract idea that belongs to a vocabulary but is neither a class nor a property. Concepts represent principles and abstractions: gravitas, autonomia. Maps to skos:Concept.
property (rdfs:Class): A relationship or attribute connecting entities or assigning values. Has domain (what it applies to) and range (what it targets). Examples: habitatores, collaborat. Maps to rdf:Property.
vocabulary (rdfs:Class): A named set of terms sharing a namespace and a domain. Each vocabulary is a folder under vocabularies/. Terms classified as class, property, or concept. Maps to skos:ConceptScheme.
broader (rdf:Property): Declares a term has a more general parent in a hierarchy. Insula, taberna, horreum all have satelles as their broader. Maps to skos:broader / rdfs:subClassOf.
domain (rdf:Property): Declares what class a property applies to. “What kind of entity can be the subject?” Habitatores has domain satelles. Maps to rdfs:domain.
range (rdf:Property): Declares what class a property targets. “What kind of entity can be the object?” Habitatores has range agens. Maps to rdfs:range.
instance-of (rdf:Property): The most fundamental property. Declares what kind of entity a note represents. Maps to rdf:type / @type in JSON-LD.
note-type (rdf:Property): Declares what document pattern a note follows, distinct from instance-of which declares semantic type. A note can be instance-of schema::class AND note-type uv::vocabulum simultaneously.
part-of (rdf:Property): Structural containment (child-to-parent). Distinct from broader (which is taxonomic). Maps to dcterms:isPartOf.
related (rdf:Property): Symmetric, non-hierarchical associative relationship. If A is related to B, then B is related to A. Maps to skos:related.
uri (rdf:Property): The actionable access point for an entity. Unlike @id (canonical identifier), uri is the practical address: Obsidian URIs, HTTP URLs, repository addresses. Maps to rdfs:seeAlso.
Domain vocabulary for the Uni-Versum model. Uses Latin terms. Source of truth: the design documents.
URI namespace: https://uni-versum.me/vocabularies/uni-versum/v1/
agens (rdfs:Class): Anyone (human or non-human) who operates within the Uni-Versum. Agents are the connective tissue between workspaces. Every agent operates under a nos. A census distinguishes local from external agents based on operational authority.
agens-humanus (rdfs:Class): A human person who inhabits, visits, or interacts with workspaces. Distinguished by being biological, not software. Includes the sovereign human establishing the nos perspective and other human collaborators.
agens-scriptum (rdfs:Class): An automated script or process. Executes predefined operations (generating files, fixing typography, producing visualizations) but does not use natural language as its interface. Tools that act, not interlocutors.
agens-scriptum-loquens (rdfs:Class): An AI agent that uses natural language as its primary interface. Distinguished from plain agens-scriptum by the capacity to converse, reason through language, and collaborate through dialogue. Large language models fall into this category.
uni-versum (rdfs:Class): A personal system of interconnected workspaces for knowledge and creative work. Sovereign, distributed, connected, replicable. Four pillars: nos, satelles, census, tabellarium.
nos (rdfs:Class): Latin “we.” The human agent at the center of a personal Uni-Versum, and the perspective from which everything else is organized. One of four pillars. A workspace whose purpose is “making sense of everything else, prioritizing, discarding, connecting.” Holds perspective, not data. Provides continuity as projects end and tools are replaced.
census (rdfs:Class): The unified registry: a specialized horreum that records all entities and relationships across the system. Three constituencies: humans (browsable), AI agents (queryable graph), automated systems (machine-readable substrate). Does not contain project content, manage communications, or track nos’s internal structure. Each Uni-Versum has exactly one census.
tabellarium (rdfs:Class): The communications layer. From tabellarius (Roman messenger). Receives, classifies, routes, and distributes. Counterpart to census — where census identifies entities, tabellarium determines how to reach them. Does not store long-term knowledge or decide relationships.
satelles (rdfs:Class): Any workspace registered in the Uni-Versum via census. The base class for all workspaces.
insula (rdfs:Class): A communal workspace where multiple activities coexist. A community, an organization, a group. Defining characteristic: multiplicity, not size. May contain tabernae and horrea.
taberna (rdfs:Class): A focused workspace for one project or practice. Private (single agent) or shared (collaboration). Defining characteristic: bounded scope, evolving.
horreum (rdfs:Class): A workspace for accumulation and organization. A reference library, a registry, an archive. Fundamentally static: grows by addition, not transformation. Material meant to be consulted, not evolved.
tabella (rdfs:Class): An individual entry in the census registry: the card recording a single entity with its relationships. The registry’s basic unit. Named after Roman wax tablets. Multiple tabellae form a tabula.
tabula (rdfs:Class): A collection of tabellae grouped by class: the table organizing cards of the same kind. Named after large Roman wooden boards for public records. Groups tabellae by their instance-of value.
instrumentum (rdfs:Class): A tool or utility that agents use but that is not itself an agent. The autonomy distinction: agents act with judgment; instrumenta are acted upon. Examples: Git, Obsidian Publish.
lectio (rdfs:Class): A perspectival reading note: the record of an encounter between a source and a specific context of work. Not a review or summary — a relationship between source, reader, and context. Must declare fontem (source) and agens (reader). Resides in the workspace where the reading applies.
collaboratio (rdfs:Class): A documented collaboration between agentes within a specific context. Records not just that agents work together, but where (which satelles) and under what terms (what autonomia). A structured registry record richer than simple dyadic properties. Standard semantic web terms (FOAF’s knows, Schema.org’s colleague) lack the workspace context and trust dimensions.
vocabulum (rdfs:Class): An entry in a vocabulary that names and defines something. Every class, property, concept, and vocabulary is a vocabulum. The unit of meaning in the vocabulary system. Maps broadly to skos:Concept.
habitatores (rdf:Property): Which agents inhabit a satelles as residents. The fundamental agent-workspace relationship. Inhabitants contribute, maintain, and shape spaces. Concentration of agents determines character.
visitatores (rdf:Property): Which agents visit a satelles without being residents. Access for specific purposes (consultation, collaboration, delivery). Managed by diploma credential.
conditor (rdf:Property): Which nos founded a satelles. The perspective that initiated the workspace and set its foundational conditions. Only one nos per workspace.
fidelis-est (rdf:Property): “Is loyal to.” Declares which nos an agens operates under. From Roman fides — the bond of trust between patron and operators. Not obedience; allegiance. Distinguishes local agents (loyal to this nos) from external ones (loyal elsewhere — diplomats, not citizens).
collaborat (rdf:Property): Declares agent-to-agent collaboration. The fundamental inter-agent relationship. For full context (workspace, autonomia), see collaboratio.
concordat (rdf:Property): Declares a satelles synchronizes its state through an instrumentum. Value is a wikilink to a registered tool.
fontem (rdf:Property): Declares the source of a lectio. Range intentionally unconstrained. From Latin fons (source, spring, origin).
gravitas (skos:Concept): The founding metaphor. Relationships as force. Shared context creates pull proportional to collaboration frequency and multiplicity. Calculated from relationship records, not stored as data.
autonomia (skos:Concept): The trust contract between human agent and automated agent: the explicit definition of what the automated agent can and cannot do. The boundary is dynamic, evolving as trust develops.
diploma (skos:Concept): Trust credential for an agens from another nos operating in a Uni-Versum. Specifies permissions, restrictions, workspace access. Named after Roman documents certifying privileges. Scope: autonomia governs intra-system trust; diploma governs inter-system trust.