metadata

Author:	Ralph B. Holland
Affiliation:	Arising Technology Systems Pty Ltd
Contact:	ralph.b.holland [at] gmail.com
version:	1.0.1
Updates:	2026-01-02T18:05Z - minor edits and replaced markdown ** with mark-up ''' (for bold).
Publication Date:	2025-12-30T07:02Z
Provenance:	This is an authored paper maintained as a MediaWiki document; clarified MWDUMP as the authoritative, permission-granting artefact governing allowable reasoning across sessionstory reflects editorial changes, not collaborative authorship.
Status:	First release

Metadata (Normative)

The metadata table immediately preceding this section is CM-defined and constitutes the authoritative provenance record for this MWDUMP artefact.

All fields in that table (including artefact, author, version, date, local timezone, and reason) MUST be treated as normative metadata.

The assisting system MUST NOT infer, normalise, reinterpret, duplicate, or rewrite these fields. If any field is missing, unclear, or later superseded, the change MUST be made explicitly by the human and recorded via version update, not inferred.

Curator Provenance and Licensing Notice

This document predates its open licensing.
As curator and author, I apply the Apache License, Version 2.0, at publication to permit reuse and implementation while preventing enclosure or patent capture. This licensing action does not revise, reinterpret, or supersede any normative content herein.
Authority remains explicitly human; no implementation, system, or platform may assert epistemic authority by virtue of this license.

Open, Unclassified Requirements for CM-Compatible LLM Selection in RT-KE Systems

This document is open, unclassified, and vendor-neutral. It defines capability requirements only and contains no operational, organisational, or classified information.

1. Scope and Intent

This document defines open, unclassified requirements for selecting large language models (LLMs) suitable for Cognitive Memoisation (CM) and Round-Trip Knowledge Engineering (RT-KE) workflows.

The intent is to enable transparent evaluation of LLM platforms for governed, durable knowledge engineering, rather than transient conversational use.

These requirements apply equally to:

research systems
enterprise deployments
on-premise installations
API-based services
self-hosted models

They are independent of licensing, vendor, hosting model, or user interface.

2. Non-Negotiable Functional Requirements

2.1 Semantic Artefact Ingestion (Ingress)

An LLM platform MUST support:

Ingestion of structured knowledge artefacts (e.g. XML, TOML, JSON)
Semantic binding of artefacts at session start
Treatment of artefacts as authoritative premises, not attachments
Artefact sizes exceeding conversational message buffer limits
Explicit confirmation of which artefacts are in scope

If structured artefacts cannot be ingested as semantic inputs, CM and RT-KE cannot operate.

2.2 Durable Artefact Emission (Egress)

An LLM platform MUST support:

Emission of generated knowledge artefacts to durable storage
Stable identifiers for emitted artefacts (paths, hashes, or URLs)
Persistence of artefacts beyond a single session
Re-ingestion of emitted artefacts without manual transcription

Inline-only output is architecturally insufficient for RT-KE.

2.3 Explicit Context Control

An LLM platform MUST provide:

Deterministic control over session context
Explicit inclusion and exclusion of artefacts
No silent truncation or implicit context loss
Clear separation between dialogue content and knowledge artefacts

Opaque or implicit context management is incompatible with governed knowledge engineering.

2.4 Transport Independence

CM workflows MUST NOT depend on:

Browser UI behaviour
Client-side buffering limits
Chat history as the sole state carrier

The same artefact must retain identical semantics across different transport surfaces.

2.5 Governance Compatibility

An LLM platform MUST ALLOW:

Curator-governed artefacts
Explicit provenance and authorship
Versioning of knowledge artefacts
Auditability of artefact usage

Stateless, ephemeral interaction models are insufficient for governed KE systems.

3. Explicit Non-Requirements

The following capabilities are NOT required by CM:

Conversational polish
Personality or stylistic continuity=Cognitive Memoisation: LLM Systems Requirements for Knowledge Round Trip Engineering=
Hidden memory or implicit recall
Vendor-specific tooling

CM prioritises knowledge durability, governance, and reproducibility over conversational fluency.

4. Evaluation Consequence

Any LLM platform that lacks either semantic artefact ingestion or durable artefact emission CANNOT support round-trip knowledge engineering, regardless of model reasoning capability or benchmark performance.

This is an architectural limitation, not a performance deficiency.

5. Classification and Disclosure

This document is:

Open: Cognitive Memoisation: LLM Systems Requirements for Knowledge Round Trip Engineering=
Unclassified
Vendor-neutral
Free of operational or organisational detail

It is published to enable open discussion, comparative evaluation, and independent verification of LLM suitability for CM and RT-KE workflows.