interactive computing
Enable an agent to recognise interactive computing, assess whether its ongoing input-response loop is usable, and determine which interventions preserve user control and computational state.
Bundle → Layer → Finding → Questions Filled
6 bundles · 11 layers · 17 findings · 27 questions
Interaction loop The reciprocal relationship between user input and computational response.
Input interpretation
How a user action becomes a computational request.
Action-to-request mapping
Record accepted inputs, their interpretation in the current context, and how acceptance or rejection is communicated.
- Which user actions supply computational input, and how does the current mode or focus change their meaning? definition
- What observable response distinguishes accepted input from ignored or rejected input? measurement
Response-guided continuation
How computational output informs the next user action.
Continuing user influence
Establish which responses let the user revise, extend, or redirect the ongoing activity.
- After receiving a response, what can the user change about subsequent computation? action
- Which parts of the activity support continuing interaction, and which are submitted jobs whose execution cannot be influenced? boundary
Timing and concurrency Response delays and the treatment of input while computation is underway.
Feedback and completion
The separate timelines for acknowledging input, reporting progress, and producing results.
Observable response milestones
Record task-specific timing expectations against defined response events rather than assuming one universal interactivity threshold.
- What are the measured delays from input to acknowledgement, useful feedback, and completion under the intended workload? measurement
- What task requirement or user evidence establishes acceptable delay at each milestone? provenance
Overlapping actions
How pending work interacts with newer user input.
Pending input and result order
Record whether new actions queue, replace, or overlap earlier work and how outdated responses are handled.
- What happens when another input arrives before the previous request completes? definition
- How can an agent determine whether a displayed response belongs to the current request or an earlier one? action
Session and working state The context that makes successive interaction steps part of one activity.
Context dependence
Working state that changes the interpretation or result of subsequent input.
Inspectable interaction context
Identify consequential state such as selected objects, command environment, active modes, and partially entered requests.
- Which retained values, selections, modes, or prior results can change what the next input does? definition
- How can an agent inspect or establish that context before acting? action
Interruption and resumption
Continuity when an interaction pauses or loses its connection.
Resumption state reconciliation
Record what survives interruption and how the participant resolves uncertainty about work that may have continued.
- Which working state and pending computations survive suspension, disconnection, or session termination? boundary
- Before retrying or resuming, how can the agent establish whether the last request completed or produced partial effects? action
User control and recovery The user's ability to intervene in processing and recover from an unintended interaction.
Intervention during execution
Controls that alter work already in progress.
Interruption semantics
Distinguish pausing execution, requesting cancellation, and confirming that processing has stopped.
- At which stages can the user pause, cancel, or redirect computation, and what continues after each intervention? action
- What observable evidence confirms that an intervention took effect? measurement
Correction after response
Recovery from rejected input, mistaken commands, or unwanted completed effects.
Correction and reversal boundaries
Record how input can be repaired and which effects can be undone, repeated safely, or compensated.
- How can the user correct an invalid or unintended input while preserving useful working context? action
- Which completed effects support undo or compensation, and which remain after the visible interface is reset? boundary
Execution and effect boundaries The relationship between visible interaction, underlying computation, and consequential effects.
Interface to execution
Where requests execute and how that execution relates to the interaction channel.
Execution coupling
Identify the computation behind each interaction and whether it continues independently of the interface.
- Which process or service executes the request, and what links its status to the response shown to the user? relation
- Does closing or losing the interface stop computation, detach from it, or leave its condition uncertain? boundary
Commitment and authority
When an interaction creates effects and who may initiate them.
Preview-to-committed effect
Distinguish exploration and staged changes from actions that commit effects under the participant's authority.
- Which interaction step commits an effect, and what evidence distinguishes a preview, an accepted request, and a completed change? boundary
- What authority or user instruction permits the agent to perform that committing action? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Check these first
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these check these first hold for the sense of interactive computing this model covers, and on what evidence? provenance
Kinds and varieties
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these kinds and varieties hold for the sense of interactive computing this model covers, and on what evidence? provenance
Standards and regulation
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these standards and regulation hold for the sense of interactive computing this model covers, and on what evidence? provenance
Real-world use
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these real-world use hold for the sense of interactive computing this model covers, and on what evidence? provenance
Typical measurements
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these typical measurements hold for the sense of interactive computing this model covers, and on what evidence? provenance
Failure modes and hazards
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these failure modes and hazards hold for the sense of interactive computing this model covers, and on what evidence? provenance
Neighbouring kinds and how to tell them apart
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these neighbouring kinds and how to tell them apart hold for the sense of interactive computing this model covers, and on what evidence? provenance
Classifiers Filled
- Family
- Thing Registry
- Category
- Physical world and living systems
- Entry kind
- thing
- Plane
- PHY
- Domain
- PHY.OBJ
- Other names and narrower kinds
- time-travel debugging
What it is Filled
Interactive computing is a mode of computer use in which a user supplies input and receives responses during an ongoing session, allowing subsequent actions to depend on intermediate results.
In scope
- The coupling between user actions, computational processing, and observable responses
- Interaction sessions, their working state, and continuity across interruptions
- Response timing, progress visibility, and handling of concurrent or repeated input
- Available user interventions, including correction, cancellation, and commitment
- The relationship between interface actions and their computational effects
Out of scope
- Hardware architecture and maintenance of terminals, computers, and input devices
- Application-specific domain rules and the correctness of domain results
- Batch computing that provides no opportunity to influence processing through an ongoing input-response loop
- Autonomous machine-to-machine protocols without a user interaction boundary
- Human-computer interaction as a research discipline
- General networking, operating-system, and cybersecurity models
Why it exists Filled
Enable an agent to recognise interactive computing, assess whether its ongoing input-response loop is usable, and determine which interventions preserve user control and computational state.
Distinguishing features Filled
- A user can provide an input, observe a response, and choose a subsequent input within the same continuing activity.
- User input can influence what computation happens next; merely watching an unchangeable output stream does not establish interactive computing.
- Interactivity can be demonstrated through a command line, graphical interface, or another input-response channel; a graphical display is not required.
- An automatic feedback controller does not qualify solely because it reacts to sensor data; the model requires an identifiable user interaction boundary.
- A submission-and-result workflow qualifies only to the extent that it supports an ongoing interaction loop; a launch button alone does not make the underlying job interactive.
What robots and AI may and may not do Filled
Must not
- Keep interactive sessions open on shared systems in ways that expose them to others.
- Act in a user's interactive session without the user knowing.
- Record a user's inputs and responses beyond what the user agreed to.
Only with a human decision
- Taking over or terminating another user's interactive session.
May
- Explain interactive computing and how it differs from batch processing.
- Run interactive sessions on systems the agent is authorized to use.
Moral aspects Filled
- Interactive systems shape how people think and work; their design affects accessibility.
- Session data can reveal a user's work and private activity.
Who is affected
- Users
- System administrators
Owners Filled
Steward
Nobody owns the concept; the operator of a system answers for its sessions.
Links to other meta-models Filled
parent
- Q7397 - registry parent class
What else AI and robots need to interact with it Filled
Identity and identifiers required Filled
- Vercy registry: vr.tr.interactive-computing
- Wikidata: Q956238 (https://www.wikidata.org/wiki/Q956238)
Direct properties required Filled
- Interaction participants: User or delegated agent ↔ interface ↔ computational process
- Input modality: Text commands, pointer actions, touch, speech, gesture, assistive input, or combinations
- Input-to-feedback latency: Milliseconds or seconds, with the triggering input, feedback event, workload, and percentile specified
- Input-to-result latency: Milliseconds or seconds, with completion criteria and workload specified
- Session condition: Ready, processing, awaiting user input, suspended, disconnected, failed, or ended; concurrent conditions recorded where applicable
- Working-state lifetime: Per action, per session, persistent across sessions, or mixed
- Pending-input policy: Queue, reject, replace, merge, or execute concurrently, specified by action type
- Intervention support: Pause, cancel, edit pending input, undo, redo, or compensate, with conditions specified
- Effect commitment: Preview, staged, executing, committed, partially committed, rolled back, or unknown
Recognition required Filled
- A user types or clicks and sees responses immediately on a screen, as in a terminal or notebook.
Capabilities and actions required Filled
- Identify available inputs and establish how their effects will be observed.
- Perform an authorised interaction step and select the next step from its response.
- Inspect progress and distinguish delayed computation from an unresponsive interaction channel.
- Correct input, cancel pending work, or undo completed work when the relevant capability exists.
- Resume an interrupted session after checking retained state and unresolved effects.
- Assess whether latency, feedback, and available controls support the intended user task.
Hazards and failure modes required Filled
- Unattended sessions can be misused by others.
- Interactive changes can be made without the review a batch process would have.
Standards and interfaces required Filled
- ISO 9241-110, issued by the International Organization for Standardization, defines interaction principles relevant to interactive systems.
- ISO 9241-210, issued by the International Organization for Standardization, specifies human-centred design principles and activities for interactive systems.
Context of use required Filled
- Interactive computing is encountered in command sessions, graphical applications and other software that responds while a person is using it.
- It supports exploration, editing and decision-making by allowing users to adjust subsequent input in response to intermediate results.
Sources Filled
- Wikidata item Q956238: interactive computing - identity and sense of the item
- Wikipedia: Interactive computing - general description of the item
Open questions
- Does the registry intend interactive computing to denote a mode of computation, a class of systems, or a physical installation, and how should that interpretation relate to its current PHY / PHY.OBJ placement?
- Which existing Vercy world model, if any, already owns interactive computing and should serve as its single source of truth?
- Should agent-mediated interaction qualify when no human participates in the immediate loop, and what boundary separates it from autonomous machine-to-machine processing?
- What task-specific evidence should establish acceptable feedback and completion delays across command-line, graphical, conversational, and other interactive forms?
- Which primary sources and applicable standards should substantiate the terminology, intervention semantics, and session-continuity expectations proposed in this draft?
Machine files
Provenance
thing registry research (pass 2) · unreviewed
Built from: models/things/publications/thing-q956238/spec.json