AI Agent
An AI agent is a software system in which a model can select actions, use tools, observe results and adapt its next steps toward an objective within defined boundaries.
KNOWLEDGE PLATFORM
Canonical definitions maintained by Electronic Artefacts and connected to implementations and evidence.
GRAPH INDEX
Every entry keeps its type, status, context and route into the wider knowledge system.
An AI agent is a software system in which a model can select actions, use tools, observe results and adapt its next steps toward an objective within defined boundaries.
Algorithmic composition is the use of rules, procedures, probability, computation or formal systems to generate or organize musical material.
Augmented intelligence is the design of tools, interfaces and systems that expand human perception, reasoning, memory and creative capability.
An autonomous system can perceive relevant state, select actions and adapt behavior toward objectives with limited direct human control during operation.
Browser software is application software delivered through web standards while relying on the browser runtime for interface, storage, permissions, media and distribution behavior.
Contextual execution is the practice of making identity, state, permissions, relationships and operating context active constraints on what a system can do.
Creative coding is the use of programming as an artistic, design or exploratory medium rather than only as an engineering implementation tool.
Cybernetic feedback is the circulation of information about a system's state back into the system so that behavior can be regulated, adapted or transformed.
Digital preservation is the long-term work of keeping digital objects understandable, authentic, usable and accessible despite changes in formats, software, hardware and context.
A domain operating system is a shared runtime and information layer that coordinates identities, entities, permissions, workflows, events and interfaces for a specific field of activity.
Entity identity is the durable naming of a person, concept, project, program, publication or artefact so that references remain stable as descriptions evolve.
Event-driven architecture is a system design approach in which components publish, transport and react to records of state change rather than depending only on direct synchronous calls.
Generative AI refers to machine-learning systems that produce or transform media such as text, image, audio, video, code or structured data from learned patterns.
A generative system is a rule-based, procedural or computational arrangement that produces variable outputs from defined constraints, materials and processes.
Graph modeling represents a domain through addressable nodes, typed relationships and properties chosen to answer specific questions and support specific operations.
A graph runtime is an execution environment in which addressable entities and typed relationships participate directly in contextual computation.
Human computer interaction is the study and design of relationships between people, computational systems, interfaces, tasks, environments and cultures of use.
Hypertext is a system of non-linear text and media connected by links, allowing readers and machines to traverse knowledge through associations.
Internet culture is the set of social forms, aesthetics, protocols, communities, memes, archives and practices that emerge through networked communication.
A knowledge graph is a structured network of entities, identifiers and typed relationships that makes knowledge navigable, queryable and reusable across contexts.
A large language model is a parameterized machine-learning model trained to predict and generate sequences of tokens from statistical patterns learned across large text or multimodal corpora.
Linked Data is a web publishing approach that identifies things with stable URIs and connects them through machine-readable statements.
Machine learning workflows are the ordered practices that move from source material and metadata through validation, preparation, training, evaluation and deployment decisions.
Metadata is structured information about a resource, such as title, creator, date, format, source, rights, subject, relation and preservation context.
Microservice architecture structures a software system as independently deployable services aligned with bounded capabilities and communicating through explicit contracts.
Motion design is the use of time, movement, transition, rhythm and sequencing to structure visual communication and interface behavior.
Multimodal AI describes systems that learn from, align, transform or generate more than one modality such as text, image, audio, video, sensor data or structured records.
In knowledge engineering, an ontology is an explicit specification of the kinds of entities, properties, relationships and constraints used to represent a domain.
Open source is a software and cultural practice in which source code is made available under licenses that allow use, study, modification and redistribution.
An open-weight model is a machine-learning model whose learned parameters are distributed for reuse under stated terms, without necessarily providing the training data, training code or freedoms required by open-source definitions.
A personal knowledge system is a practice and tool environment for capturing, linking, revising and retrieving notes, sources, ideas and decisions over time.
Procedural graphics are images or visual systems generated from rules, algorithms, parameters, simulation or code rather than only manually drawn assets.
Provenance is information about the people, activities, sources, transformations and decisions involved in producing a thing or piece of data.
Retrieval-augmented generation is an architecture that retrieves external information at query time and supplies selected evidence to a generative model before or during response production.
Signal archaeology is the interpretation of technical, sonic, visual or archival traces as evidence of hidden processes, transformations and histories.
In VASTE, an Actor is an extension-owned graph identity whose effective observation and action are resolved inside a specific System, Environment, topology and policy context.
Speech datasets are structured collections of recorded speech, transcripts, speaker context and quality metadata prepared for analysis, preservation or downstream machine-learning workflows.
Speech recording is the directed capture of spoken language with enough technical, contextual and consent metadata to support review, reuse and preservation.
Systems thinking is an approach that studies elements through their relationships, feedback, boundaries, delays, stocks, flows and patterns of behavior over time.
Typography is the design and arrangement of written language through typefaces, spacing, hierarchy, rhythm, scale and reading conditions.
VAB is VASTE’s Assisted Boot orchestration path for understanding an initial intent, composing a sealed plan, admitting a Founding Actor and materializing a governed starting graph.
A .vast file is an experimental `vast/1` sealed JSON package carrying a System-relative whole-System or subtree closure, its manifest and integrity hash—not a live Environment.
A VASTE extension is an optional, System-scoped capability unit that declares domain actions, bindings and surfaces while runtime authority remains in the core execution path.
VASTE has a closed primitive roster—Vertex, Tie, Action, Surface and Environment—from which graph identity, topology, admitted change, observation and execution-local context are composed.
Voice technology is the design and engineering of systems that capture, analyze, transform, preserve or present spoken voice as technical and cultural material.