Conformance

Part of the WebMEM® Protocol
Location: /protocol/conformance/
Status: Current Draft
Last Updated: 2026-08-25

Overview

Conformance defines the requirements an implementation must satisfy before its representation can legitimately be described as conforming to the WebMEM Protocol.

WebMEM conformance is not established merely by using:

WebMEM terminology;
WebMEM Fragment Class names;
WebMEM data-* attributes;
HTML <template> containers;
Semantic Data Templates;
or structures that resemble WebMEM examples.

A representation may use those mechanisms while failing to preserve the semantic distinctions required by the protocol.

WebMEM conformance means preserving the semantic distinctions required by the protocol using a conforming representation and, where applicable, a conforming serialization or protocol profile.

The governing principle is:

WebMEM Conformance
        ↓
Semantic Integrity
        +
Structural Integrity
        +
Serialization Integrity
        +
Applicable Component Requirements

A representation that satisfies the syntax of a WebMEM serialization while violating the WebMEM knowledge model is not conforming.

Purpose of Conformance

The purpose of conformance is to establish a predictable boundary between:

Uses WebMEM concepts

and:

Conforms to the WebMEM Protocol

That distinction is necessary for:

interoperability;
validation;
implementation testing;
consistent use of Fragment Classes;
semantic preservation across serializations;
Resolver interoperability;
tooling;
documentation;
and meaningful conformance claims.

Without a defined conformance boundary, two implementations could use the same WebMEM vocabulary while representing materially different knowledge semantics.

Normative Language

The WebMEM Protocol uses the following terms to express normative requirements:

Term	Meaning
MUST	The requirement is mandatory for conformance where it applies.
MUST NOT	The prohibited behavior is incompatible with conformance where the requirement applies.
SHOULD	The behavior is recommended. An implementation may depart from it when a valid implementation reason exists and required semantics remain preserved.
SHOULD NOT	The behavior is generally discouraged. Departure may be appropriate when justified and when required semantics remain preserved.
MAY	The behavior or capability is optional.

These terms apply only where the corresponding protocol requirement is relevant to the knowledge or component being represented.

Conformance Is Requirement-Based

WebMEM does not define conformance as a checklist requiring every implementation to use every capability of the protocol.

A conforming SDT does not need to contain all 15 Fragment Classes.

A representation containing no derived knowledge does not need to invent derivation metadata.

A representation containing no material cross-object relationship does not need to invent relationships.

A representation whose applicability is already unambiguous does not need to manufacture an unresolved state.

A resource that is not a Resolver does not need to satisfy Resolver-specific requirements.

Conformance requires preservation of applicable semantics, not artificial population of every WebMEM capability.

Three Layers of Conformance

WebMEM conformance operates across three principal layers:

Protocol Conformance
│
├── Semantic Conformance
│
├── Structural Conformance
│
└── Serialization Conformance

These layers address different aspects of the representation.

Semantic Conformance

Semantic conformance requires the representation to preserve the meaning established by the WebMEM knowledge model.

Where applicable, this includes preserving:

the semantic role of each Fragment Class;
the subject identity of assertions and structures;
the distinction between entities, fragments, datasets, SDTs, and resources;
source-observed versus derived knowledge;
provenance and evidentiary lineage;
relationship predicates and direction;
collection membership;
geographic applicability;
temporal applicability;
jurisdictional, population, eligibility, or other material scope;
semantic ordering where order carries meaning;
and unresolved ambiguity where the represented knowledge does not support unique resolution.

For example:

segment:H5521-290-001
monthly_premium = 18.50 USD

is not semantically equivalent to:

plan:H5521-290
monthly_premium = 18.50 USD

if the premium applies only to the segment.

The numerical value is identical.

The subject is not.

A representation that collapses that distinction is not semantically conforming.

Structural Conformance

Structural conformance requires the components of the WebMEM representation to be organized in a manner consistent with the architecture defined by the protocol.

Conceptually:

Web Resource
        ↓
Page-Level Dataset / SDT
        ↓
Fragments
        ↓
Assertions and Structures

Cross-Cutting Semantics
├── Identity
├── Provenance
├── Relationships
├── Context
└── Resolution

Structural conformance includes preserving distinctions such as:

Web Resource
≠ SDT
≠ Fragment
≠ Represented Entity

and:

Fragment Identity
≠ Subject Identity

and:

Collection Membership
≠ Identity Equivalence

A representation that contains the correct values but organizes them in a way that destroys these distinctions may fail structural conformance.

Serialization Conformance

Serialization conformance requires the concrete representation to satisfy the requirements of the WebMEM serialization being used.

The current normative serialization is HTML-in-HTML.

For HTML-in-HTML, serialization conformance includes requirements such as:

a valid SDT container;
an identifiable SDT;
valid fragment containers;
fragment identity;
declared Fragment Classes;
parseable assertions and structures;
valid datatypes and units where required;
resolvable references;
correct context inheritance;
relationship structure;
and valid HTML parsing behavior.

Serialization conformance is necessary but not sufficient for full WebMEM conformance.

Valid WebMEM syntax does not rescue invalid WebMEM semantics.

Read the Serialization specification →

Read the HTML-in-HTML Serialization specification →

Conformance Subjects

Conformance claims should identify what is claimed to conform.

The current WebMEM Protocol recognizes several possible conformance subjects.

WebMEM Representation

A machine-facing representation of WebMEM knowledge.

Semantic Data Template

A page-level SDT and its contained knowledge objects.

WebMEM Fragment

An individual fragment conforming to the core fragment requirements and the requirements of its declared Fragment Class.

WebMEM Resolver

A Resolver resource satisfying the applicable Resolution Model and Resolver requirements.

WebMEM Serialization

A concrete representation satisfying the requirements of a defined WebMEM serialization.

WebMEM Parser or Validator

A tool that claims to process or validate a defined portion of the WebMEM representation according to the requirements applicable to that capability.

Conformance of a parser or validator does not imply conformance of every representation it processes.

Publisher-Side Conformance

The WebMEM Protocol is principally a publisher-side knowledge representation protocol.

Its normative requirements therefore apply primarily to how publishers represent and serialize knowledge.

WebMEM does not define conformance requirements for how search engines, answer engines, language models, agents, retrieval systems, or other consumers must reason over WebMEM knowledge.

A publisher can produce a conforming WebMEM representation regardless of whether any particular consuming system chooses to use it.

Conformance describes the representation published by the publisher. It does not prescribe consumer behavior.

Core Protocol Conformance

A WebMEM representation claiming core protocol conformance MUST preserve every core semantic distinction applicable to the knowledge it represents.

At minimum, where applicable, the representation MUST preserve:

the identity of the page-level SDT;
the identity and Fragment Class of each fragment;
the subject identity required to interpret assertions and structures;
the distinction between source-observed and derived knowledge;
material provenance;
material relationships;
applicability context;
semantic ordering;
collection membership;
resolution-relevant distinctions;
and material unresolved ambiguity.

A representation MUST NOT introduce semantic claims unsupported by the represented knowledge merely to satisfy a WebMEM structure.

Fragment Conformance

A WebMEM fragment MUST satisfy:

Core Fragment Requirements
        +
Declared Fragment Class Requirements

For example:

DataFragment
→ Core Fragment Requirements
→ DataFragment Requirements

while:

ProcedureFragment
→ Core Fragment Requirements
→ ProcedureFragment Requirements

and:

IdentityFragment
→ Core Fragment Requirements
→ IdentityFragment Requirements

A fragment MUST NOT claim a Fragment Class whose semantic role does not match the knowledge represented by the fragment.

For example, publisher-calculated statistics MUST NOT be represented as source-observed DataFragment assertions when the derivation is material to their meaning and lineage.

Read the WebMEM Fragments specification →

Fragment Class Requirements Are Normative

Each defined Fragment Class may specify additional conformance requirements appropriate to that class.

For example:

a DataFragment may require identifiable factual assertions and applicable provenance;
a DerivedStatsFragment may require derivation semantics and source lineage;
an IndexFragment may require identifiable collection membership;
a ProcedureFragment may require ordered procedural steps;
an EligibilityFragment may require a sufficiently defined eligibility structure;
an IdentityFragment may require an identifiable subject and sufficient identity structure;
and other Fragment Classes impose requirements appropriate to their semantic roles.

Those class-specific requirements form part of WebMEM conformance when the corresponding Fragment Class is used.

SDT Conformance

A conforming Semantic Data Template MUST:

have a stable, identifiable SDT identity;
contain one or more conforming WebMEM fragments;
preserve page-level context where that context is required to interpret contained knowledge;
preserve distinctions among fragment identity, entity identity, dataset identity, and resource identity;
preserve material relationships among contained or referenced knowledge objects;
preserve applicable provenance;
avoid applying inherited context to knowledge for which that context is not valid;
and satisfy the requirements of the serialization used to represent the SDT.

An SDT MAY contain any subset of Fragment Classes appropriate to the resource.

An SDT MUST NOT include artificial fragments merely to satisfy an assumed completeness requirement.

Resolver Conformance

A WebMEM Resolver claiming conformance MUST satisfy:

Core WebMEM Requirements
        +
Resolution Model Requirements
        +
Resolver Requirements
        +
Serialization Requirements

A conforming Resolver MUST, where applicable:

identify or make resolvable its entry key;
preserve the type, namespace, or scope required to interpret that entry key;
establish a coherent resolution purpose;
preserve material identities and relationships;
preserve applicability conditions;
preserve legitimate option spaces;
preserve unresolved states where available context does not support unique resolution;
preserve applicable provenance;
and expose canonical continuation resources where those resources form part of the Resolver’s defined purpose.

A page does not become a conforming Resolver merely because it contains WebMEM markup and several links.

Read the Resolvers specification →

Compositional Conformance

WebMEM conformance is compositional.

Each component must satisfy the requirements applicable to its role.

For example:

WebMEM Resolver
│
├── Core Protocol Conformance
│
├── SDT Conformance
│
├── Fragment Conformance
│   ├── IdentityFragment
│   ├── IndexFragment
│   ├── DataFragment
│   └── MetaFragment
│
├── Resolution / Resolver Conformance
│
└── HTML-in-HTML Serialization Conformance

The whole representation depends on the conformance of the material components from which it is composed.

This does not mean that one optional component absent from the representation creates a failure.

It means that components actually used MUST satisfy the requirements applicable to them.

Applicable Semantics

Conformance requirements apply according to the knowledge represented.

For example, if a fragment contains:

source-observed facts

then source provenance may be required.

If a fragment contains:

publisher-derived statistics

then derivation lineage may be required.

If a resource contains:

Plan
→ Segment
→ Geography

and those relationships determine applicability, the relationships and applicable scope must be preserved.

If an information space contains:

Candidate A
Candidate B
→ insufficient context to distinguish

then the ambiguity must not be collapsed merely to create a single value.

Conformance follows semantic necessity.

Identity Conformance

A conforming representation MUST preserve identity at the semantic level necessary to interpret the represented knowledge correctly.

It MUST NOT:

collapse distinct entities merely because they are related;
treat names as equivalent to canonical identity when ambiguity exists;
treat parent and child identities as interchangeable;
treat fragment identity as subject identity;
treat a web resource as identical to the entity it represents unless genuine equivalence is explicitly intended;
or attach an assertion to a broader or different subject than the knowledge supports.

Identity conformance preserves the correct subject.

Read the Identity specification →

Provenance Conformance

A conforming representation MUST preserve provenance sufficient to understand material evidentiary lineage.

It MUST NOT:

represent publisher-derived knowledge as though it were directly asserted by the source;
collapse materially different source datasets into an undifferentiated source label when those distinctions affect lineage;
apply fragment-level provenance to an assertion whose actual source differs;
erase transformations or derivations necessary to understand how an assertion was produced;
conflate the creator of a source dataset with the publisher of the WebMEM representation;
conflate publication, retrieval, effective, applicability, generation, or modification dates when those temporal distinctions are material;
or represent a citation as though the citation alone necessarily establishes the complete lineage of the assertion.

Provenance conformance preserves recoverable lineage.

Read the Provenance specification →

Relationship Conformance

A conforming representation MUST preserve material relationships among identifiable knowledge objects.

Where a relationship is represented, the source, predicate, and target MUST be identifiable or unambiguously recoverable from context.

A conforming representation MUST NOT:

replace a known specific relationship with a materially weaker generic relationship when doing so changes meaning;
reverse a directional relationship without preserving the correct inverse semantics;
treat related entities as equivalent entities without evidence of genuine equivalence;
assume that relationships are transitive when their semantics do not establish transitivity;
propagate parent, child, membership, or applicability relationships to related entities without support;
treat hyperlinks as substitutes for semantic relationships;
treat document proximity or shared labels as proof of a relationship;
or remove relationship scope when geography, time, jurisdiction, population, or another condition materially affects the relationship.

Relationship conformance preserves the correct connections.

Read the Relationships specification →

Resolution Conformance

Where a WebMEM representation exposes resolution structure, it MUST preserve the distinctions necessary to determine applicability correctly.

A conforming representation MUST NOT:

represent one candidate as uniquely resolved while material ambiguity remains;
select a child-specific assertion at the parent level without resolving the applicable child;
assume that parent-level knowledge necessarily applies to every child;
discard geographic, temporal, jurisdictional, population, eligibility, or other context necessary for resolution;
substitute retrieval relevance for semantic resolution;
substitute document proximity, URL structure, or lexical similarity for represented resolution structure;
collapse a legitimate option space into one option without a semantic basis for selection;
or represent incomplete resolution as complete resolution.

When the represented knowledge does not support unique resolution, a conforming representation MUST preserve the material ambiguity rather than manufacture certainty.

Resolution conformance preserves correct applicability.

Read the WebMEM Resolution Model →

Serialization Conformance

A representation claiming conformance to a particular WebMEM serialization MUST satisfy both:

Core WebMEM Semantic Requirements
        +
Serialization-Specific Requirements

For the current HTML-in-HTML serialization, this means that valid HTML syntax alone is insufficient.

For example:

<template
  data-webmem-sdt
  data-sdt-id="https://example.org/resource/#webmem">

  ...

</template>

may be syntactically valid HTML and still fail WebMEM conformance if its contents:

misidentify subjects;
misclassify fragments;
erase provenance;
collapse relationships;
lose applicability;
or otherwise violate the WebMEM semantic model.

Serialization syntax carries the knowledge model. It does not override it.

Validation and Conformance

Validation and conformance are related but different concepts.

Validation asks whether a representation satisfies requirements that can be tested mechanically.

Conformance asks whether the representation satisfies the applicable requirements of the WebMEM Protocol.

A validator may be able to determine that:

an SDT has an identifier;
a fragment declares a recognized Fragment Class;
fragment IDs are unique within an SDT;
a relationship contains a source, predicate, and target;
a provenance reference resolves;
a currency assertion declares an applicable unit;
a value is compatible with its declared datatype;
or an HTML structure is parseable.

A generic validator may not be able to determine that:

the publisher selected the correct source record;
the asserted relationship accurately reflects the real-world relationship;
an assertion has been attached to the correct entity;
a derivation method is substantively appropriate;
a source itself is accurate;
or the represented knowledge is factually true.

Therefore:

Successful Validation
≠
Proof of Semantic Truth

and:

Valid Syntax
≠
Full Protocol Conformance

Conformance Does Not Certify Truth

WebMEM conformance does not certify that represented knowledge is true.

A conforming representation may accurately preserve the identity, provenance, relationships, applicability, and derivation of an assertion whose underlying source contains an error.

Likewise, a publisher may make a substantively incorrect assertion while representing that assertion in structurally conforming WebMEM.

Conformance evaluates whether the protocol’s semantic and structural requirements have been honored.

It does not function as an external fact-checking or certification authority.

Conformance preserves the integrity of the representation. It does not guarantee the truth of everything represented.

Conformance Does Not Certify Authority

A conforming representation does not establish that the publisher is authoritative for every entity, source, policy, definition, or assertion it represents.

Authority may be evaluated from:

provenance;
source identity;
publisher identity;
jurisdiction;
methodology;
corroboration;
or other evidence available to the consuming system.

WebMEM exposes the structure necessary to make those distinctions.

It does not assign authority merely through conformance.

Conformance Does Not Guarantee Machine Use

A conforming WebMEM representation does not guarantee:

search-engine indexing;
answer-engine ingestion;
retrieval;
ranking;
citation;
AI visibility;
memory retention;
knowledge-graph inclusion;
or use by any particular machine system.

Those outcomes depend on the architecture, policies, capabilities, and decisions of consuming systems.

The WebMEM Protocol defines the publisher-side representation.

Conformance means that the publisher has represented the knowledge according to the protocol. It is not a promise about what any external system will do with that representation.

Conformance Does Not Prescribe Consumer Reasoning

A consuming system is not required to:

trust a conforming WebMEM representation;
prefer it over another source;
follow every represented relationship;
accept a publisher’s derivation;
use a Resolver;
retain the knowledge;
or follow a particular reasoning algorithm.

A consumer may evaluate WebMEM knowledge according to its own source-selection, retrieval, ranking, safety, reasoning, or trust policies.

WebMEM conformance establishes what the publisher represented and how that representation preserves meaning.

Extensions

WebMEM implementations MAY extend the protocol with domain-specific:

fields;
attributes;
relationship predicates;
identifier systems;
enumerations;
structures;
Fragment Class profiles;
Resolver Profiles;
or other additional semantics.

For example, a healthcare implementation may use:

data-plan-year
data-contract-id
data-plan-id
data-segment-id

while an automotive implementation may use:

data-vin
data-model-year
data-engine-code

Extensions are permitted because WebMEM cannot define every domain vocabulary required by every implementation.

Extension Conformance

An extension MUST NOT:

redefine the meaning of a reserved WebMEM concept;
contradict core WebMEM semantics;
change the meaning of a defined Fragment Class while continuing to claim that Fragment Class;
collapse identity distinctions required by the core protocol;
erase required provenance;
change relationship semantics without declaring a distinct relationship predicate;
or make the core representation dependent on an undisclosed interpretation that reverses or contradicts its apparent WebMEM meaning.

Extensions MAY add meaning.

They MUST NOT silently reverse core meaning.

A generic WebMEM processor that does not understand a domain-specific extension SHOULD still be able to recover the core WebMEM structure to the extent that the representation otherwise conforms.

Extensions may enrich WebMEM. They must not redefine WebMEM.

Unknown Extensions

A processor may encounter attributes, predicates, fields, or structures it does not recognize.

For example:

<template
  data-webmem-fragment
  data-fragment-class="DataFragment"
  data-entity-id="plan:H5521-290-0"
  data-plan-year="2026"
  data-cms-contract-type="H">

  ...

</template>

A generic processor may not understand data-cms-contract-type.

It should nevertheless remain able to recognize:

Fragment Class:
DataFragment

Entity:
plan:H5521-290-0

Unknown extensions SHOULD NOT cause a processor to reinterpret recognized core WebMEM semantics.

Reserved Vocabulary

Names and attribute families explicitly reserved by a WebMEM specification SHOULD be used only with the meanings defined by that specification.

For the current HTML-in-HTML serialization, reserved attribute families include:

data-webmem-*
data-sdt-*
data-fragment-*
data-entity-*
data-provenance-*
data-defined-term-*
data-relationship-*
data-collection-*
data-member-*
data-resolution-*

Implementations SHOULD use clearly domain-scoped names for extensions where naming collisions are reasonably possible.

No Conformance Levels

The current WebMEM Protocol does not define:

Level 1 WebMEM
Level 2 WebMEM
Level 3 WebMEM

or equivalent tiers that allow required semantic protections to be omitted while retaining a lower conformance designation.

For example, an implementation cannot claim a reduced level of WebMEM conformance merely because it intentionally omits required provenance or collapses required identity distinctions.

Requirements apply according to the knowledge and components represented.

WebMEM conformance is requirement-based, not tier-based.

Partial Implementations

An implementation MAY support only part of the WebMEM Protocol.

For example, a software library may support:

HTML-in-HTML parsing;
only selected Fragment Classes;
validation but not generation;
generation but not Resolver validation;
or another explicitly limited capability.

Such an implementation SHOULD describe its supported scope precisely.

For example:

Supports parsing conforming WebMEM HTML-in-HTML SDTs
and DataFragment, MetaFragment, and IndexFragment structures.

is preferable to:

Fully WebMEM compliant.

when the implementation does not support the full set of capabilities implied by the broader claim.

Non-Conforming Representations

A representation that intentionally omits or contradicts a requirement applicable to the knowledge it represents MUST NOT be described as a conforming WebMEM representation.

For example:

WebMEM conforming except provenance

is not an appropriate conformance claim when provenance is required to preserve the lineage of the represented knowledge.

Likewise:

WebMEM conforming except identity scope

is not appropriate where identity scope is necessary to distinguish the subjects of assertions.

An implementation may accurately describe itself as:

inspired by WebMEM;
using selected WebMEM concepts;
supporting selected WebMEM Fragment Classes;
or exporting a subset of WebMEM-compatible structures;

provided that such language does not imply protocol conformance that has not been achieved.

Conformance Claims

A conformance claim SHOULD identify the subject and applicable specification.

Examples:

This resource contains a WebMEM-conforming Semantic Data Template
serialized using HTML-in-HTML.

or:

This resource implements a WebMEM-conforming Resolver
using the HTML-in-HTML serialization.

or:

This fragment conforms to the WebMEM DataFragment specification.

Claims SHOULD NOT imply certification, endorsement, licensing status, or third-party validation unless such status has actually been granted through a separately defined process.

Self-Declared Conformance

Unless a separate certification program states otherwise, WebMEM conformance may be self-declared by an implementer.

Self-declaration does not reduce the applicable protocol requirements.

An implementer making a conformance claim is asserting that the representation satisfies the requirements applicable to that claim.

A self-declared conformance claim is not equivalent to independent certification.

Conformance Metadata

A serialization or implementation profile MAY define machine-readable metadata indicating the specification or profile against which a representation claims conformance.

Such metadata does not itself establish conformance.

For example:

declared_conformance:
WebMEM HTML-in-HTML

is a claim about the representation.

The representation must still satisfy the corresponding requirements.

Conformance metadata can declare a claim. It cannot make a non-conforming representation conforming.

Conformance and Licensing

Technical conformance and licensing are separate concerns.

A representation may technically satisfy the requirements of the WebMEM Protocol while its implementation, distribution, deployment, or commercial use remains subject to applicable WebMEM license terms.

Likewise, possession of a license does not make a technically non-conforming representation conforming.

Conceptually:

Technical Conformance
→ Does the representation satisfy the protocol?

Licensing
→ Is the implementation or use authorized under applicable terms?

Neither condition should be represented as a substitute for the other.

Current licensing terms are published separately at the WebMEM license resource.

Conformance and Dataset Declaration

Use of Schema.org Dataset JSON-LD does not by itself establish WebMEM conformance.

The Dataset declaration identifies and describes the page-level machine dataset.

The associated SDT must independently satisfy applicable WebMEM semantic, structural, and serialization requirements.

Likewise, an otherwise conforming SDT does not become non-conforming merely because a Dataset declaration is absent unless an applicable WebMEM specification, profile, or implementation requirement specifically requires that declaration.

Where Dataset JSON-LD is used to declare the WebMEM dataset, the Dataset identity and SDT identity SHOULD be mapped consistently according to the Dataset Declaration and SDT Mapping specification.

Read Dataset Declaration and SDT Mapping →

Conformance and Semantic Equivalence

Conformance does not require two representations of the same WebMEM knowledge to be textually identical.

Two conforming representations may differ in:

element ordering where order is not semantically meaningful;
human-readable labels;
optional descriptive metadata;
reference organization;
domain-specific extensions;
or serialization syntax where multiple conforming serializations are defined.

They may still be semantically equivalent if the material WebMEM knowledge remains the same.

Conceptually:

Representation A
        ↓
Recover WebMEM Model

Representation B
        ↓
Recover WebMEM Model

Material Semantics
        =
Equivalent

Semantic equivalence requires preservation of the distinctions material to the represented knowledge.

For example, equivalent representations must not disagree materially about:

subject identity;
Fragment Class;
asserted values;
provenance;
source-observed versus derived status;
relationship meaning;
applicability;
or resolution state.

Conformance and Versioning

A conformance claim SHOULD identify the applicable version of the WebMEM Protocol when different protocol versions could produce materially different requirements.

Implementations SHOULD NOT assume that conformance to an earlier WebMEM specification automatically establishes conformance to the current protocol.

Likewise, a representation conforming to the current protocol does not need to reproduce retired WebMEM 1.x mechanisms merely to maintain historical syntax compatibility.

Version transitions should preserve durable semantic concepts while allowing retired implementation mechanisms to remain retired.

Relationship to WebMEM 1.x

WebMEM 1.x did not define a sufficiently complete protocol-wide conformance model.

Earlier specifications described requirements within individual Fragment Classes, YAML structures, provenance blocks, fragment headers, data bindings, and other implementation components, but did not consistently distinguish among:

semantic conformance;
structural conformance;
serialization conformance;
Fragment Class conformance;
Resolver conformance;
validation;
and consumer behavior.

The current protocol makes those distinctions explicit.

It also separates durable WebMEM semantics from retired WebMEM 1.x mechanisms associated with:

semantic digests;
retrieval hints;
memory conditioning;
trust scoring;
confidence weighting;
agent-specific retrieval behavior;
and YAML-specific representation.

The absence of those retired mechanisms does not constitute a conformance failure under the current protocol unless a current specification independently defines the same semantic requirement.

Conversely, use of historical WebMEM 1.x syntax does not by itself establish conformance to the current protocol.

Current WebMEM conformance is determined by the current protocol requirements, not by superficial compatibility with historical WebMEM syntax.

Conformance Failure Modes

Several patterns may create the appearance of WebMEM implementation while failing protocol conformance.

Syntax-Only Conformance

The representation uses WebMEM HTML attributes and containers but does not preserve the required semantic model.

Fragment-Class Misclassification

A fragment declares one Fragment Class while representing knowledge whose semantic role belongs to another class.

Identity Collapse

Distinct entities or identity scopes are merged in a way that changes the subjects of assertions or relationships.

Provenance Collapse

Material source, derivation, transformation, or lineage distinctions are removed.

Relationship Collapse

Known semantic relationships are flattened, reversed, generalized, or replaced by navigation or proximity.

Applicability Collapse

Material geographic, temporal, jurisdictional, population, eligibility, segment-level, or other applicability distinctions are omitted.

False Resolution

Material ambiguity remains but the representation claims or implies unique resolution.

Extension Override

A domain extension silently changes the meaning of a core WebMEM concept.

Validation Substitution

Successful mechanical validation is presented as proof of complete semantic conformance or factual truth.

Consumer-Outcome Claim

Protocol conformance is represented as guaranteeing indexing, retrieval, citation, ranking, AI memory, or another external system behavior.

Historical-Syntax Substitution

Use of retired WebMEM 1.x structures is treated as sufficient evidence of current protocol conformance.

These failure modes demonstrate why conformance must be evaluated against the applicable semantic and structural requirements rather than visual resemblance to WebMEM markup.

Minimum Conformance Model

A WebMEM conformance claim should be evaluated against the requirements applicable to the claimed subject.

At minimum, the evaluation should determine:

Conformance Subject: what resource, SDT, fragment, Resolver, serialization, parser, validator, or other implementation component is claiming conformance;
Applicable Protocol Version: which WebMEM specification governs the claim;
Core Semantic Requirements: which identity, provenance, relationship, applicability, assertion, composition, or resolution semantics apply;
Component Requirements: which Fragment Class, Resolver, SDT, or other component-specific requirements apply;
Serialization Requirements: which concrete serialization requirements apply;
Extension Requirements: whether domain-specific extensions preserve core semantics;
Validation Results: which mechanically testable requirements have been satisfied;
and Known Limitations: whether the implementation intentionally supports only a defined subset of WebMEM capabilities.

The evaluation does not require every possible WebMEM capability to be present.

It requires every requirement applicable to the claimed representation to be satisfied.

Protocol-Wide Conformance Requirements

A representation claiming conformance to the WebMEM Protocol MUST:

preserve the semantic roles of the WebMEM components it uses;
preserve subject identity at the level required to interpret the represented knowledge correctly;
preserve material provenance and derivation lineage;
preserve material relationships and their semantic direction and scope;
preserve material applicability context;
preserve source-observed versus publisher-derived distinctions;
preserve semantically meaningful ordering;
preserve collection membership without treating membership as identity equivalence;
preserve material unresolved ambiguity;
use Fragment Classes consistently with their defined semantic roles;
satisfy the requirements of the serialization used;
satisfy applicable component-specific requirements;
ensure that extensions do not redefine or contradict core WebMEM semantics;
avoid introducing unsupported semantic claims merely to complete a WebMEM structure;
and avoid representing external machine behavior as a consequence guaranteed by protocol conformance.

A representation MUST NOT claim full WebMEM conformance when it knowingly violates a requirement applicable to the knowledge or component represented.

Protocol Integrity

Conformance is the mechanism through which the architectural principles defined throughout the WebMEM Protocol become enforceable implementation requirements.

The preceding protocol components establish different forms of semantic integrity:

Semantic Data Template
→ organizes page-level machine knowledge

Fragment Classes
→ preserve the kind of knowledge represented

Assertions
→ preserve what the publisher states

Identity
→ preserves the correct subject

Provenance
→ preserves recoverable lineage

Relationships
→ preserve the correct connections

Resolution
→ preserves correct applicability

Serialization
→ preserves those semantics in concrete syntax

Conformance
→ determines whether the applicable protocol contract was honored

Each layer protects a different part of the representation.

A failure in one layer cannot necessarily be repaired by correctness in another.

For example:

Correct Value
+
Correct Provenance
+
Wrong Identity
=
Wrong Assertion

or:

Correct Identity
+
Correct Value
+
Correct Provenance
+
Wrong Applicability
=
Wrong Resolution

or:

Correct Conceptual Model
+
Serialization That Erases Relationships
=
Incomplete Published Representation

Conformance exists to prevent an implementation from satisfying only the convenient portions of the architecture while still claiming the semantic integrity of the whole.

WebMEM conformance is not a claim that markup exists. It is a claim that the applicable meaning survived publication.

Conformance Statement

The core WebMEM conformance principle can therefore be stated simply:

A WebMEM-conforming representation preserves the semantic distinctions required by the WebMEM Protocol, satisfies the requirements applicable to the components it uses, and satisfies the requirements of its declared serialization without introducing unsupported meaning or erasing material identity, provenance, relationships, applicability, or ambiguity.

This principle governs conformance across WebMEM resources, SDTs, fragments, Resolvers, and serializations.

Overview

Purpose of Conformance

Normative Language

Conformance Is Requirement-Based

Three Layers of Conformance

Semantic Conformance

Structural Conformance

Serialization Conformance

Conformance Subjects

WebMEM Representation

Semantic Data Template

WebMEM Fragment

WebMEM Resolver

WebMEM Serialization

WebMEM Parser or Validator

Publisher-Side Conformance

Core Protocol Conformance

Fragment Conformance

Fragment Class Requirements Are Normative

SDT Conformance

Resolver Conformance

Compositional Conformance

Applicable Semantics

Identity Conformance

Provenance Conformance

Relationship Conformance

Resolution Conformance

Serialization Conformance

Validation and Conformance

Conformance Does Not Certify Truth

Conformance Does Not Certify Authority

Conformance Does Not Guarantee Machine Use

Conformance Does Not Prescribe Consumer Reasoning

Extensions

Extension Conformance

Unknown Extensions

Reserved Vocabulary

No Conformance Levels

Partial Implementations

Non-Conforming Representations

Conformance Claims

Self-Declared Conformance

Conformance Metadata

Conformance and Licensing

Conformance and Dataset Declaration

Conformance and Semantic Equivalence

Conformance and Versioning

Relationship to WebMEM 1.x

Conformance Failure Modes

Syntax-Only Conformance

Fragment-Class Misclassification

Identity Collapse

Provenance Collapse

Relationship Collapse

Applicability Collapse

False Resolution

Extension Override

Validation Substitution

Consumer-Outcome Claim

Historical-Syntax Substitution

Minimum Conformance Model

Protocol-Wide Conformance Requirements

Protocol Integrity

Conformance Statement

Related Protocol Documents