Section 8 – The JanOS File System

JanOS introduces a storage model designed around narrative, semantics, and intent. In JanOS, storage is not treated as a passive hierarchy of byte sequences, but as a long-lived record of human work: what was attempted, what changed, what was decided, and what concluded.

The JanOS file system is called NAFS: Narrative File System. It is intentionally designed to support comprehension and continuity over time, rather than only persistence and location.

Semantic Organization

In JanOS, files and artifacts are organized primarily by meaning rather than by path.

Artifacts can be grouped automatically around:

• tasks and intents
• projects and workstreams
• timelines and phases
• entities (people, customers, systems, cases)
• user-defined narratives and organizational categories.

This reduces the cognitive effort required to locate and interpret information. It also enables retrieval that matches how humans remember work: by purpose, context, and progression, not by the last folder it was saved into.

Versioning and History

NAFS maintains lightweight, continuous versioning as a default property of storage.

Unlike snapshot-based systems that capture state at arbitrary intervals, JanOS aligns version boundaries with meaningful events in the lifecycle of work, such as:

• task and intent transitions
• completion, abandonment, or transformation of intent
• explicitly marked milestones or decisions
• policy-governed contextual markers (optional).

Where organizational policy allows, the system can preserve sufficient historical context to support accountable digital archaeology: understanding how a result came to be, not merely what the latest state is. This addresses a common failure mode of modern computing, where crucial work context exists briefly and then disappears into transient UI states, ad-hoc messaging, or unstructured revisions.

Replay and Temporal Navigation

Because history is recorded semantically, JanOS can provide temporal navigation through work.

Users can navigate backward through the history of an intent, task, or project, including:

• file and artifact states
• relevant context and relationships
• interface states derived from declarative UI
• intent markers and lifecycle events (including closure).

Replay is intended to support understanding, support, auditing and learning; not surveillance. It enables users and organizations to revisit work as a coherent sequence, rather than as disconnected files and logs.

Safety and Isolation

NAFS is designed to be safe by default. Applications do not receive all-encompassing access to user storage. Instead, they operate within:

• an isolated storage namespace by default
• explicit, declarative access granted via application manifest
• sandboxed interactions unless policy grants broader privileges.

This enforces deny-by-default at the storage layer and drastically reduces the impact of common classes of attack, including mass deletion, ransomware-style encryption, and unauthorized exfiltration of unrelated user artifacts.

By combining isolation with verifiable application identity and lineage, JanOS can make access decisions that are explainable and auditable: who accessed what, under which policy, and as part of which intent.

Contextual Relationships and Semantic Queries

In long-lived digital work, understanding an artifact often requires more than access to its contents. People need to know why something exists, how it came to be, who was involved, and what other work it relates to. Conventional file systems provide little structural support for these questions, leaving users and organizations to reconstruct context manually from fragmented tools, messages, and institutional memory.

To address this, NAFS maintains an explicit set of semantic relationships between artifacts, intents, identities, and other system-recognized entities. These relationships are not incidental metadata or application-specific annotations, but first-class system entities with identity, lineage, and policy context.

Relationships in NAFS may express, for example:

• that an artifact was produced as part of a specific intent or task
• that a document was referenced in a communication or review
• that a decision or approval applies to a particular artifact
• that multiple artifacts belong to the same narrative phase or workstream
• that an artifact derives from or supersedes another
• that specific individuals or roles participated in its creation or modification.

By maintaining such relationships explicitly, JanOS enables contextual understanding without relying on behavioral inference, heuristic reconstruction, or application-level indexing. The system records what is known, how it became known, and under which authority or policy, rather than attempting to infer meaning retrospectively.

NAFS exposes this structure through a constrained set of system-defined contextual queries. These queries allow users and applications to explore the neighborhood of an artifact safely and purposefully, answering questions such as:

• what other artifacts are directly related to this one
• which intents or tasks this artifact contributed to
• which people or roles were involved, subject to policy and authorization
• how the artifact evolved over time within a broader narrative
• which communications, reviews, or decisions reference it.

These queries operate within the same identity, lineage, and policy framework as all other JanOS subsystems. Access to relationship information is governed by role, intent context, and organizational policy. Where content cannot be disclosed, the system may return redacted or structural results that preserve contextual understanding without exposing protected information.

Importantly, NAFS does not attempt to model all possible relationships, nor does it automatically link artifacts based on speculative inference. Relationships are created through explicit system actions, application declarations, or user-acknowledged operations, and each relationship retains its provenance and scope of validity. This restraint limits semantic drift and preserves long-term legibility.

When an intent reaches closure, the set of relationships associated with that intent may be compacted, summarized, or sealed according to policy. In this way, completed work becomes a stable and interpretable part of organizational memory rather than an ever-expanding web of unresolved associations.

Through explicit semantic relationships and bounded contextual queries, NAFS allows digital artifacts to be understood as part of coherent human work, rather than as isolated files. This capability supports comprehension, collaboration, replay, and long-term continuity while remaining aligned with JanOS principles of clarity, accountability, and respect for human attention.