Glossary

Artifact

A file or binary-like object related to a conversation or message. An artifact may be present in a provider export, referenced without payload, or added later through manual capture.

Artifact Hydration

The act of attaching a payload to an already known artifact reference by appending an ArtifactPayloadCaptured event. Hydration does not rewrite the earlier ArtifactReferenced history.

Bounded Context

A semantic boundary that defines vocabulary, rules, and responsibility for a part of the system.

Canonical

In NEXUS, canonical does not mean universally true or final.

It means:

• the preferred normalized representation inside a specific bounded context
• the form the system chooses so different source inputs can be treated consistently
• a stable working shape for storage, dedupe, projection, and later interpretation

In the current Ingestion and CanonicalHistory bounded contexts, canonical means:

• append-only
• provenance-preserving
• provider-neutral where practical
• reparsable from preserved raw source artifacts

It does not mean:

• raw provider truth
• the final NEXUS graph ontology
• the only valid interpretation for all later domains and lenses

Canonical History

The append-only normalized event history maintained by NEXUS.

In the current bounded context, it is the canonical store for observed acquisition history. It is derived from acquisition inputs and intended to be durable, provenance-preserving, and reparsable.

CommandSlice

In the Event Modeling lens, a CommandSlice is a slice centered on user or system intent that asks for a change.

When working in Event Modeling contexts in NEXUS, CommandSlice is the preferred term for the intent-side slice that should lead to one or more durable events.

Domain

A broad area of reality or work that a part of the graph concerns, such as Ingestion or SoftwareDevelopment.

Flow

An ordered sequence of Event Modeling slices that accomplishes something useful.

Event

An append-only record in the canonical history that states what was observed, referenced, captured, or completed.

Fact

A graph-level assertion derived from canonical history and tied back to provenance.

FnHCI

The broader NEXUS interaction concern and top interaction namespace line, covering visual UI, CLI, API, accessibility, and other human-computer interaction surfaces without collapsing them into one renderer or one interface type.

FnUI

The narrower visual/UI-specific system and likely public/package line within FnHCI.

FnUI is the line most likely to represent the Bolero-replacement system on top of Blazor while also providing the NEXUS GUI-facing application shell, navigation, views, and operator workflows without becoming the source of truth for NEXUS state.

Graph Substrate

The shared structural layer of nodes, edges, and assertions derived from observed history.

Imprint

A domain-neutral meaning applied to structure when that structure is understood as the persistent result of causality and a source for later interpretation.

In the current ontology direction, Imprint is a role, not a structural primitive.

Import

A single acquisition run that processes one or more provider artifacts and appends canonical history.

Intake Channel

A stable classification for the path through which a LOGOS signal entered NEXUS, such as AI conversation, forum thread, Discord thread, email thread, bug report, or app feedback.

Access Context

A stable classification for the visibility or authority level under which material was observed, such as public-anonymous, registered-user, owner, admin, bot, or api-client.

Acquisition Kind

A stable classification for how material entered NEXUS, such as manual-note, web-scrape, api-pull, manual-export, or live-capture.

Attribution Reference

The explicit license, source, or attribution text/reference that should be carried forward when a rights policy requires prominent attribution in later public-facing use.

Lens

A perspective over the underlying graph that localizes meaning, naming, grouping, and emphasis for a purpose or audience.

LOGOS

The NEXUS concept area for intake, refinement, and retention of knowledge-bearing signals.

LOGOS is broader than any one storage or retrieval technology.

LOGOS Intake Note

A durable Markdown seed note under docs/logos-intake/<pool>/ that records explicit source-system, source-instance, access, acquisition, rights, locator, and handling metadata for an intake signal before full ingestion exists for that source type.

LOGOS Sanitized Note

A derived Markdown note under docs/logos-intake-derived/<pool>/ that preserves source classification, access, rights, and handling-policy provenance while intentionally excluding raw locators and raw copied source text.

LOGOS Handling Policy

An explicit handling classification for a LOGOS intake signal covering sensitivity, sharing scope, sanitization status, and retention class.

LOGOS Handling Report

A derived audit view over docs/logos-intake/ and docs/logos-intake-derived/ that surfaces how LOGOS notes are currently classified for access, acquisition, rights, sensitivity, sharing scope, sanitization status, retention, and likely attribution obligations.

Entry Pool

The explicit pool path where a LOGOS note enters or lands: raw, private, or public-safe.

Private Pool

A handling pool for owner or restricted internal use where sensitive detail may still be retained for legitimate work.

Public-Safe Pool

A handling pool for explicitly approved material that is safe for public-facing or broadly shared downstream use only when both handling and rights boundaries are satisfied.

Raw Pool

A handling pool for preserved intake with maximal fidelity and provenance, regardless of whether the material is safe for broader sharing.

Semantic Role

A meaning classification applied to a node without changing the node's structural identity.

Normalized Import Snapshot

A per-import derived snapshot written from one parsed provider payload before canonical dedupe.

It captures what that import payload contained at the normalized layer, so imports can be compared with snapshot semantics instead of only additive canonical changes.

Normalization Version

The version label of the parser/canonicalizer shape that produced a canonical observation.

It exists so NEXUS can distinguish:

• a provider object changing under the same normalization rules
• NEXUS re-observing the same provider object after its own parser or canonicalizer changed

Observed History

The provenance-first layer that records what NEXUS actually encountered from exports, manual additions, and later capture paths.

Projection

A rebuildable materialized view derived from canonical history.

Retention Class

A stable classification for how long or how durably an intake signal should be retained, such as ephemeral, case-bound, or durable.

Provider Artifact

An original source object received from a provider, such as a ChatGPT, Claude, or Grok export zip.

Raw Object

A preserved source or derived extraction stored in the object layer.

Sanitization Status

A stable classification for whether an intake signal or derived artifact is still raw, has been redacted, has been anonymized, or has been approved for wider sharing.

Scope

An explicit filtered boundary over data or graph material, such as one provider, one conversation, one import batch, or one node neighborhood.

Sensitivity

A stable classification for how sensitive an intake signal is, such as personal-private, customer-confidential, internal-restricted, or public.

Sharing Scope

A stable classification for who may access or use an intake signal or derivative, such as owner-only, case-team, project-team, or public.

Slice

In NEXUS terminology, Slice is reserved for the Event Modeling sense: a unit of change or read.

View

In the Event Modeling lens, a View is the dataset or structure shown to an actor in the business/UI lens.

When working in Event Modeling contexts in NEXUS, View is the preferred term over read model when discussing the derived structure behind what a ViewSlice shows.

ViewSlice

In the Event Modeling lens, a ViewSlice is a slice centered on what an actor can see and decide from the current business state.

When working in Event Modeling contexts in NEXUS, ViewSlice is the preferred term over ReadSlice for the read-side slice in Event Modeling work.

Overlap Reconciliation

The explicit, traceable, and reversible process of linking multiple acquisition sources that may describe the same underlying interaction.

In NEXUS, overlap is preserved separately first and reconciled later by explicit logic rather than silently collapsed during import.

Signal Kind

A stable classification for the kind of knowledge-bearing signal represented in LOGOS, such as conversation, message, feedback, bug report, or support question.

Source System

A stable classification for the originating system or surface from which a LOGOS signal came, such as ChatGPT, Claude, Grok, Codex, a forum, Talkyard, Discord, or an app feedback surface.

Source Instance

A stable slug for one concrete source authority or deployment within a broader source system, such as one specific Talkyard host, Discord server, wiki, or repository surface.

Rights Policy

A stable classification for the reuse boundary governing a LOGOS signal or derivative, such as owner-controlled, personal-training-only, site-terms-restricted, cc-by, cc-by-sa, api-contract-restricted, customer-confidential, or review-required.

Batch

An import-bounded or materialization-bounded contribution unit, especially in the graph working layer.