Glossary
The words used across r-io suite, grouped by theme: how work is scoped, what the collaborative model contains, how concepts relate, which statuses a node can carry, and the platform notions behind deduction, interpretation, simulation and execution. Definitions follow the metamodel resources and the code, not marketing.
Where a model lives
- Collaboration
- The top-level working context: a set of users and the knowledge spaces they share. Managed by the system governance (GovSystem) and selected in r-ioga; every REST call and WebSocket topic is scoped by it.
- Knowledge space
- One graph inside a collaboration: the nodes and edges of a situation, its projects, sensors and processes. A collaboration may hold several knowledge spaces (for instance one per use case); r-io4U and the web applications always work in one selected knowledge space.
- Project
- A
Projectnode that ties a domain, the model extensions it uses, a detection strategy, an operating mode (real or simulated), reference projects and shared resources. Use cases are ready-made projects. - Use case
- A packaged project delivered as a Maven module (
gind-usecases-*) with its models, images, game scenarios and datasets. Two are on the standalone runtime classpath; the others are test-scope demonstration projects. See use cases. - Shared resources project
- A project mounted into another one to reuse its context and partners (for example IMT Albi Context or Albi Partners).
- Reference project
- A reusable catalogue of common problems or roles for a domain (Crisis Commons Problems, Health Commons Roles…) declared in the project descriptor.
- Metamodel
- The grammar of a model: concepts, their properties, inheritance and connection rules, published as XML resources and JAXB classes. The core metamodel defines the technical concepts, the collaborative model the shared vocabulary. See Metamodel, domains and extensions.
- Domain
- A business specialization of the collaborative model: Crisis, Health and Production, plus the technical Core domain. A project selects one domain.
- Extension
- An installable vocabulary package adding concrete concepts to a domain (Urban, Road Network, Hospital, Skateboard…). Twelve extensions exist; each declares a privileged domain.
- Effective metamodel · effective domain
- The metamodel actually available in a knowledge space once core, collaborative, domain and selected extensions are merged and published by the meta-models governance.
- Concept · instance
- A concept is a type declared by a metamodel (
Person,Risk); an instance is a node of that type in a knowledge space. Abstract concepts (Asset,Resource,Potentiality) cannot be instantiated directly. - Node · edge · property
- The generic model is a graph: nodes carry a type, a status, labels and typed properties (including geometries); edges carry a type, a source and a target, and may hold properties such as 2L code or an activation condition.
- Label
- Free tag on a node used for scoping and queries:
instance, a solution identifier,EXPECTED_FREEZE… Detection and supervision rely on label combinations such asinstance:EXPECTED_FREEZE:<solutionId>. - brokenEdges
- Edges attached to a node whose other end is not yet in the graph (it lives in another file or batch). They are reconciled after all nodes are published. See model loading.
The vocabulary of the collaborative model
Abstract concepts are marked; domain and extension specializations are listed at the end. Full inventories with properties are on the Collaborative page.
- Asset abstract
- Physical or logical entity with operational value; parent of territories and resources.
- Territory
- Geographical perimeter on which assets can be positioned.
- Resource abstract
- Mobilizable asset used by collaborative work; parent of Person, Group, Good, Stock and Sensor. Its hidden presumed trajectory property holds the JSON trajectory computed when a spatial sensor moves it (expected trajectory is computed on each move task of a deduced process: the walk along the pedestrian networks, dated from the task’s expected start).
- Person
- Human resource that can carry roles and perform functions; linked to a system User through Assigned_To.
- Group
- Organizational grouping of people and assets (an organization, a team, a company).
- Role
- Concrete set of responsibilities provided by a resource; the pivot between actors and the functions they Provide.
- Good
- Material or logical good used by the collaboration (a vehicle, a building, a camera…). Extensions specialize it heavily. Its optional commissioning date property (date it entered service) gives its age to the strain indicator, like the birthdate of a Person.
- Stock
- Countable or consumable resource quantity, consumed or produced by functions and tasks.
- Sensor
- Core concept for a data source: spatial, quantitative, social, dedicated, media or other. An ACTIVE Sensor node becomes a running sensor and a CEP or ML configuration. See Sensors.
- Problem abstract
- Problematic situation shared by actualities and potentialities.
- Actuality
- Observed fact requiring treatment or representing a resolved state. Alerts (Crisis), Act_of_care (Health) and Order (Production) are actualities.
- Potentiality abstract
- Possible situation carrying a probability; specialized into Risk and Opportunity.
- Potential
- Generic potential situation before specialization into risk or opportunity.
- Risk
- Adverse potentiality that objectives may prevent; its
probabilityis written by CEP rule actions or 2L assignments. - Opportunity
- Beneficial potentiality that objectives may enable.
- Objective
- Expected result used to drive deduction and planning; treats actualities, prevents risks or enables opportunities.
- Objectives_Manager
- Container and coordination point for the objectives of a project; the root a deduction starts from (its Plans edges).
- Function
- Capability or activity that a resource can perform; satisfies objectives, needs, implies or supports other functions.
- Solution
- Candidate or selected response satisfying the objectives: the deduced process with its assignments and indicators.
- Process
- Executable organization of collaborative work: tasks, events, gateways and flows compiled and deployed by the PRIO engine.
- Task · Personal_Task · SubProcess_Task
- Executable unit of work; a personal task targets one participant, a subprocess task delegates to a nested process.
- Start_Event · End_Event · Timer_Event
- Process events controlling the flow: entry point, terminal event, time-based trigger.
- Exclusive · Inclusive · Parallel_Gateway
- Branching and synchronization points selecting one, several compatible, or all branches.
- Remove_Tag
- Process action used to remove a designated task.
- User system
- System account with login and rights; a Person or Group is Assigned_To its User.
- Alert crisis
- Actuality specialization of the Crisis domain; thirteen typed alerts (fire, flood, toxic gas…) inherit the generic Alert.
- Act_of_care health
- Observed care situation that can consume a Patient or another resource.
- Order production
- Observed production or logistics demand (a Damage-typed actuality) that objectives treat.
- Building · Parking · Road · Road_Network · Pedestrian_Path · Pedestrian_Network · Water_Network · Electrical_Network · Urban_rail_network
- Goods brought by the Urban, Road Network, Pedestrian Network, Water Network, Electrical Network and Urban Rail Network extensions.
- Camera · Photo_Camera · Phone · Watch
- Goods of the Electronic Tools extension; cameras are the visual sources served by r-iopiot.
- Patient · Operating_room · Wake_up_room · Transfert_room · Company · Skateboard · Dog
- Concepts of the Hospital, Industry, Skateboard and Animal extensions.
How concepts connect
Relation names are the edge types of the collaborative model (plus Near from the core). Five relations (Generates, Proposes, Concerns, Treated_By, Applies_On) are declared without a connection rule and used programmatically only.
- Plans
- Places objectives, processes or tasks in an operational plan; the Plans edges of an Objectives_Manager are the entry of a deduction.
- Satisfies
- Links a Function to the Objective it can satisfy.
- Provides
- Links a Role, or a resource capability, to the Functions it can perform; used to find candidate actors.
- Consumes · Produces
- Declare the roles, stocks or resources a Function or Task requires or yields.
- Needs · Implies · Supports
- Functional dependencies between functions: a prerequisite, an implied follow-up, a supporting function. They shape the recursive function graph rebuilt during deduction. Implies, Supports and Needs can carry a 2L Condition Evaluation that guards the task of the dependent function at execution (see conditioned links); this condition is read once. To hold a task until a condition is true, a “Wait Until” Mediation function with a Wait Condition is placed between the two functions.
- Treats · Prevents · Enables
- Link an Objective to the Actuality it treats, the Risk it prevents or the Opportunity it enables; created or cloned by the Actuality/Potentiality interceptor.
- Generated_By
- Causal chain between actualities and potentialities; carries an activation condition evaluated by the 2L interpreter. See Generated_By interceptor.
- Impacts
- Concrete effect of a problem on an asset, role or resource.
- Is_Susceptible_To
- Declares that an asset or role can be affected by a problem or potentiality; the Has interceptor uses it to clone potentiality structures onto resources.
- Near
- Core proximity relation between instances (a node and its FREEZE or REFERENCE twin, a sensor and its reference instance).
- Has
- Links a Resource to its Roles, or a Group to the Assets it holds. See Has interceptor.
- Pilots
- A Sensor controls a Good, for example a radar controlling a camera; used by reflex events.
- Fixed_Onto · Acts_On
- Fix a resource onto another resource or territory; state that a resource acts on a territory or asset.
- Assigned_To
- Links a Person or Group to its system User, and assigns a Resource to an Objective, an Objectives_Manager or a Task.
- Belongs_To
- Membership in a group, process or structural container; computer tasks find their Good through it.
- Solves
- Connects a Solution to the objectives manager it resolves.
- Induces · Created_By
- Derived consequence between concepts; creator or generating concept of an element.
- Sequence_Flow
- Normal ordering between tasks, events and gateways of a process. It may carry a 2L condition copied from Implies/Needs, paired with an “otherwise” flow to the End_Event.
- Exception_Flow
- Alternate process path triggered by an exception; its condition is evaluated in 2L by the PRIO engine.
- Event_Flow
- Connects an event occurrence to the process path it activates; its condition can qualify the actuality that generated the event.
- Code2L
- Not a relation but the property of an edge holding 2L code; the Code2L interceptor registers it with the low-level manager. See Code2L interceptor.
Values of the node status
Enumerated by statusType in genericModel.xsd.
- ACTIVE
- The node is part of the current situation: it is queried by rules, shown by the applications and counted by detection as the field model.
- FREEZE
- A frozen template or twin: not active, kept as a model to clone from (for instance potentiality structures cloned onto resources).
- EXPECTED_FREEZE
- The expected situation attached to a solution (its twin), compared with ACTIVE instances by detection and highlighted in blue or orange on r-iota maps.
- PROJECTED_ACTIVE · PROJECTED_EXPECTED_FREEZE
- Projected variants: a situation computed from a Forecast at a future date (a projection knowledge space), never executed as such.
- REFERENCE
- Reference instance of a concept, used as a template for potentiality clones and for ML alert validation (Near edge to the reference).
- NOT_VALIDATED
- Produced by an ML prediction and awaiting human validation in r-iota; excluded from sure alerts.
- SHARE
- Shared element made available across knowledge spaces.
Words from the services and tools
- Governance
- The container that validates, enriches and persists models: core (graph), system (users, collaborations), meta-core and meta-models (metamodels, effective domains), models (publication), deduction and AI assistants. See Core governance.
- Interceptor
- Hook executed before and after a node or edge is written (Sensor, Function, Actuality/Potentiality, Code2L, Has, Generated_By). See model loading.
- Publication (sync · async · bulk)
- Writing a model into a knowledge space through the models governance; the bulk path (UNWIND by lots) is used for initial imports and replays interceptors and history.
- 2L
- The small language written on the edges of the collaborative model (JavaCC grammar,
LLInterpretor), with$sourceand$targetbound to the edge: activation, susceptibility and impact rules that set statuses, probabilities and zones, and the Condition Evaluation of links and flows evaluated by the workflow engine. See the 2L language reference and Models & DSL. - KPIL · indicator · metric
- The indicator language: declarations
Concept:name[aggregation, objective, "unit", "description", {icon, color, thresholds}], optionally followed by a function body, evaluated on functions, tasks and processes to value solutions (duration, cost, distance, strain…). The optional display block gives the Material icon (plain or per concept), the colour and the ascending level thresholds that r-iosepe uses to draw the indicator. A body sees the task asthisand calls whitelisted methods (getAssignedTo,getResources,getAgeInYears,getSharersCount,Indicator.sum/Indicator.max…). When flows of the process carry conditions, an indicator is an interval from the scenario where it is lowest to the one where it is highest (one scenario per combination of the values the conditions read, up to 512; a process that loops is not measured). Task, function and resource counts follow the same rule, and moves are not counted as tasks. Each outcome of the branches is stored with the solution (branchOutcome: its decisions, executed tasks and indicators). A metric is a computed value in a strict canonical format. See Models & DSL. - Strain
- Default indicator of the r-ioga script (
Resource:strain, in %): fatigue of a person or wear of a good on a task, 10 % per hour shared between the engaged resources, plus 2 % per year of age (birthdate of a Person, commissioning date of a Good);Function:strainkeeps the most strained resource. Its thresholds 50 and 75 colour it green, orange or red. Two levels: the strain a task adds to each of its resources, and the current strain of the resource (real value of itsResource:strain), increased by the share of each task once done, up to 100 %; the deduction does not read the current level, the supervision fires Failing when it crosses 75 %. - Move · Wait
- Tasks added to the deduced process once it is planned, when the indicators script of the project allows them: a “Move to <task>” (move improver,
Function:distanceindicator) when the actor has to travel to its task, a “Wait before <activity>” (wait improver,Function:waitindicator) when it stands idle at least 60 s, held back by a task of another actor. Neither counts as a task or a function of the solution;Resource:waitgives the waiting time of each actor. - Interpretation config
- A deployable configuration carrying either a CEP rule (Siddhi) or an ML config (prediction module); created for each active sensor. See Interpretation.
- CEP rule
- Siddhi application correlating sensor events over windows; its actions publish typed events that concept interpreters turn into ACTIVE facts (Danger, Effect, Emerging_Risk…).
- ML config · predict module
- Interpretation config delegating to an
SPIAbstractPredictModule(ChatGPT today); predictions come back as a generic model whose nodes are NOT_VALIDATED. - Reflex event
- Event emitted by the high-level interpretation linking a produced fact to the sensor and the Good it pilots; consumed by r-iota for immediate reactions.
- Supervision · supervision strategy
- Continuous comparison, every 10 s by default, of the field (ACTIVE instances) with the supervised solution (its EXPECTED_FREEZE twins) and a baseline taken after their clone. Each strategy is a plugin on the shelf (Maven module plus r-iota configuration extension) checked in the supervision configuration property of the project; Weighted Graph Distance Strategy and KPIL indicators recalculation are checked by default. It replaces the former F-score and D3 detection methods. See Supervision engine and Supervision strategies.
- Failure category
- One of the five ways a supervised solution can fail: Uncovered (new problems not treated), Failing (resources of remaining tasks gone, the process cannot complete), Obsolete (treated problems vanished), Wrong (indicators drift from the deduced values), Improvable (better resources available). Strategies feed them with values normalized by their thresholds, 1.0 meaning action needed; a category takes the highest value. An alert category newly crossing 1 opens a single alert box in r-iota; Improvable only shows a blue badge and a light banner.
- Deduction · deduction strategy
- Generating collaborative process candidates from objectives, functions, actors and resources. Strategies are
AbstractProcessGeneratorimplementations: Drools profiles and the native Basic Bloc Satisfaction. See Deduction & planning. - Planner
- Assigns actors and resources, sequences tasks and stamps expected dates on a deduced plan; only the Default planner is built.
- Campaign · campaign solution · signature
- A campaign groups the solutions of one deduction or experiment; each solution carries indicators and a signature used to drop duplicates. See Campaigns.
- FIRST_GOOD_SOLUTION · ALL_SOLUTIONS
- Deduction modes: stop at the first unique solution and publish it, or explore every combination.
- Process instance · exchange
- A running deployed process and each monitored SOAP exchange (task call, response, fault) recorded by the monitoring service.
- Human task · to-do list
- Task created by the PRIO engine for a person or group, stored by the human-task service and shown in the to-do lists (web applications, r-io4U).
- Mock endpoint
- Simulated SOAP or REST partner standing in for a Good’s real service; answers process tasks with recorded exchanges. See Simulation.
- Game scenario · dataset · sensor controller
- A replayable configuration of sensor controllers and mock endpoints; a dataset is the recorded sequence a controller plays into a sensor.
- Topic · TopicSet · WS-Notification
- Named event streams declared by producers in TopicSet files and routed by the event broker; browsers receive them through scoped WebSocket topics. See Event broker.
- Time series
- MongoDB time-series collection per sensor storing every measurement with second precision.
- Profile · wave
- A container profile lists which web applications and backends run; backends start in four waves. See Configuration & ports.
- Standalone
- The single-JVM distribution that runs all services and web applications with embedded databases.
- Driver
- Language-native façade (Java, Python, C#, Dart) over the REST resources and WebSocket topics. See API reference.
- WebJar
- Compiled Angular application packaged as a Maven artifact and served by its Dropwizard web module.
- GJaxb classes
- Java types generated from the XSD schemas (prefix
GJaxb), mirrored in TypeScript, Dart, Python and C#.
