Part of the WebMEM® Protocol
Fragment Class: MetaFragment
Location: /protocol/fragments/metafragment/
Status: Current Draft
Last Updated: 2026-08-24
Overview
A MetaFragment is a WebMEM fragment class for publishing high-level metadata about a dataset, data source, collection, or other information resource.
It provides contextual and structural information such as title, description, creator, canonical source location, publication date, retrieval date, license, format, version, and other metadata required to identify and understand the resource.
Unlike a DataFragment, which represents source-observed factual knowledge about an entity or context, or a DerivedStatsFragment, which represents calculated or aggregated knowledge, a MetaFragment describes the information resource itself.
A MetaFragment can therefore provide a stable metadata and provenance anchor that other WebMEM fragments, Machine Fact Objects, or knowledge structures reference when they depend on the same source.
Semantic Purpose
The semantic purpose of a MetaFragment is to state:
This is the information resource, dataset, or source from which associated knowledge originates, and this metadata describes that resource.
A MetaFragment may establish:
- the identity of the dataset or information resource;
- its title and description;
- the person or organization responsible for publishing it;
- its canonical source location;
- an archive or downloadable source location;
- publication, effective, or retrieval dates;
- license or usage information;
- format and version information;
- integrity information such as a checksum;
- semantic or domain scope;
- and other metadata necessary to identify or interpret the resource.
The MetaFragment allows this shared information to be represented once and referenced by other knowledge objects rather than duplicated throughout the SDT.
MetaFragment Within an SDT
A MetaFragment exists as one modular knowledge object within a Semantic Data Template (SDT).
The relationship may be represented as:
Web Resource → SDT → MetaFragment → Source or Dataset Metadata
An SDT may contain one MetaFragment or multiple MetaFragments depending on the number of datasets, sources, or information resources represented by the page-level machine knowledge.
For example, an SDT may use:
- one MetaFragment describing an authoritative product dataset;
- another MetaFragment describing an enrollment dataset;
- and another MetaFragment describing a geographic source.
Other fragments can then reference the appropriate metadata and provenance object according to the source supporting their knowledge.
MetaFragment Requirements
A conforming MetaFragment should establish enough information to identify and understand the information resource it describes.
| Element | Purpose |
|---|---|
| Fragment Class | Identifies the knowledge object as a MetaFragment. |
| Fragment ID | Provides a stable identifier for the MetaFragment within the published knowledge structure. |
| Resource ID | Provides a stable identity for the dataset, source, or information resource being described. |
| Title | Provides the human-readable name of the resource. |
| Description | Describes the contents, purpose, or scope of the resource. |
| Creator | Identifies the person or organization responsible for creating or publishing the resource. |
| Canonical Source | Identifies the canonical location or documentation resource associated with the source. |
| Publication Information | Identifies publication, release, effective, or other applicable temporal information. |
| Provenance | Provides or participates in the source and lineage information used by other WebMEM knowledge objects. |
Additional metadata may be included when required by the source, domain, provenance model, or applicable conformance profile.
Resource Identity
A MetaFragment describes an identifiable information resource.
The identity of that resource is distinct from the identity of the MetaFragment itself.
For example:
- Fragment ID: identifies the specific MetaFragment published within the SDT.
- Resource ID: identifies the dataset or source described by the fragment.
A resource identifier might represent:
dataset:cms-ma-landscape-2026The same source identity may be referenced by DataFragments, DerivedStatsFragments, IndexFragments, FAQFragments, or other WebMEM knowledge objects that depend on that resource.
This allows source identity to remain stable independently of the individual fragments that use it.
Metadata Fields
A MetaFragment may contain metadata appropriate to the resource it describes.
| Metadata | Purpose |
|---|---|
| Title | Official or canonical name of the dataset or source. |
| Description | Explanation of the resource’s contents, structure, purpose, or coverage. |
| Creator | Person or organization responsible for creating or publishing the resource. |
| Home | Canonical homepage, documentation page, or source landing page. |
| Archive | Direct location of an archived, downloadable, or version-specific source when available. |
| License | License, public-domain status, or other applicable usage rights. |
| Published | Publication or release date of the resource. |
| Retrieved | Date the publisher retrieved, imported, or otherwise acquired the source data. |
| Format | Format of the source, such as CSV, JSON, XML, ZIP, or XLSX. |
| Version | Version, release identifier, edition, or other source version information. |
| Year | Calendar, policy, plan, or other applicable year associated with the resource. |
| Checksum | Cryptographic hash or other integrity value associated with a specific source artifact. |
| Guidelines | Canonical documentation, regulatory guidance, methodology, or interpretive material associated with the source. |
| Tags | Optional classification terms associated with the resource. |
| Related Terms | Defined terms associated with the resource when explicit semantic alignment is useful. |
Not every MetaFragment requires every metadata field. The appropriate fields depend on the type of information resource being described and the provenance requirements of the knowledge that references it.
Example MetaFragment
The following example represents metadata for a CMS Medicare Advantage source dataset using the current WebMEM HTML serialization.
<template
data-webmem-fragment
data-fragment-class="MetaFragment"
data-fragment-id="meta-cms-ma-landscape-2026">
<section
data-entity-type="Dataset"
data-entity-id="dataset:cms-ma-landscape-2026">
<h3 data-role="title">
CMS Medicare Advantage Landscape, 2026
</h3>
<p data-role="description">
CMS-published dataset listing approved Medicare Advantage
plans and their service areas for the 2026 plan year.
</p>
<span data-role="creator">
Centers for Medicare & Medicaid Services (CMS)
</span>
<a
data-role="home"
href="https://www.cms.gov/">
CMS
</a>
<meta
data-role="license"
content="Public Domain">
<time
data-role="published"
datetime="2026-06-01">
2026-06-01
</time>
<time
data-role="retrieved"
datetime="2026-06-28">
2026-06-28
</time>
<meta
data-role="format"
content="ZIP (XLSX)">
<meta
data-role="version"
content="2026">
</section>
</template>
The fragment establishes that:
- the knowledge object is a
MetaFragment; - the fragment has its own stable identity;
- the information resource being described is an identifiable dataset;
- the dataset’s title, description, creator, source location, license, dates, format, and version are explicitly represented;
- and other WebMEM knowledge objects can reference this resource and its metadata when establishing provenance.
MetaFragment and Provenance
MetaFragment and provenance are closely related but serve different purposes.
A MetaFragment describes the information resource itself.
Provenance describes the lineage relationship between published knowledge and the sources from which that knowledge originated.
A MetaFragment can therefore function as a reusable source-description object within the provenance architecture.
For example:
Machine Fact Object → Provenance Reference → MetaFragment → Source Dataset
or:
DerivedStatsFragment → Provenance Reference → MetaFragment → Source Dataset
This allows source metadata to be declared once while individual facts and fragments preserve explicit relationships to that source.
Read the Provenance specification →
MetaFragment and DataFragment
| DataFragment | MetaFragment |
|---|---|
| Represents source-observed facts about an entity or context. | Represents metadata about an information resource or dataset. |
| Answers: What facts are known about this object? | Answers: What is this source or dataset? |
| Example: a plan’s monthly premium is $35. | Example: identifies and describes the dataset from which the premium originated. |
| Publishes factual values from a source. | Describes the source itself. |
A DataFragment may therefore reference a MetaFragment when its facts originate from the information resource described by that MetaFragment.
MetaFragment and DerivedStatsFragment
A DerivedStatsFragment publishes calculated knowledge produced from underlying source data.
A MetaFragment may describe the dataset or datasets used to produce those calculations.
For example:
Derived Statistic → DerivedStatsFragment → Provenance → MetaFragment → Source Dataset
This allows a derived value to remain connected not merely to a generic source label, but to an explicitly described information resource.
MetaFragment and Identity
The information resource described by a MetaFragment has its own identity.
This identity should remain distinct from the identity of the MetaFragment that describes it.
A dataset may therefore be referenced consistently across multiple SDTs, fragments, facts, and provenance relationships using the same stable resource identifier.
This distinction becomes particularly important when:
- the same dataset supports multiple pages;
- multiple versions of a dataset exist;
- different source files belong to the same publishing series;
- or a machine must distinguish the abstract dataset from a particular publication or retrieved artifact.
Read the Identity specification →
Multiple Sources
An SDT may contain multiple MetaFragments when its knowledge depends on multiple information resources.
For example, a resource may combine:
- plan availability data from one source;
- enrollment data from another source;
- quality ratings from another source;
- and geographic information from another source.
Each source can be represented independently by its own MetaFragment.
Other fragments and Machine Fact Objects can then reference the appropriate source metadata according to their actual evidentiary lineage.
This prevents source relationships from being flattened into a single generic page-level citation when the underlying knowledge comes from different datasets.
Versioning and Source Artifacts
A MetaFragment may distinguish between a continuing dataset or information resource and a specific version or artifact of that resource.
For example:
- the dataset series may have a stable identity;
- a plan-year release may have its own version identity;
- and a particular ZIP or CSV file may have its own archive location and checksum.
Publishers should preserve enough version information to identify the actual source supporting published knowledge when source versions can materially change the meaning or validity of that knowledge.
MetaFragment and Resolution
A MetaFragment may participate in machine resolution when a source identifier, provenance reference, dataset identity, or related knowledge object can be resolved to the metadata describing that source.
For example:
Fact → Provenance Reference → MetaFragment → Canonical Source
This allows a machine to move from a factual assertion toward information describing its evidentiary origin.
The MetaFragment does not itself define the WebMEM Resolution Model. It participates in resolution by providing an identifiable source object and explicit relationships to canonical source information.
Read the WebMEM Resolution Model →
Conformance
A conforming MetaFragment must:
- identify itself as a
MetaFragment; - have an identifiable fragment scope;
- identify the dataset, source, or information resource it describes;
- provide a human-readable title or equivalent identifying label;
- provide sufficient metadata to distinguish and understand the resource;
- identify the creator or publisher when known and applicable;
- identify a canonical source location when one exists;
- preserve applicable publication, version, or temporal information required to identify the source;
- and conform to the applicable WebMEM serialization requirements.
Additional metadata fields are optional unless required by an applicable domain, provenance, or conformance profile.
Relationship to WebMEM 1.x
The MetaFragment class originated in the WebMEM 1.x Semantic Data Template specification and remains a defined fragment class in the current WebMEM Protocol.
Earlier specifications described MetaFragment primarily as a trust-layer anchor for datasets and associated it with ProvenanceMeta, W3C PROV alignment, semantic digests, glossary alignment, export indexing, and memory-layer behavior.
The current protocol preserves the semantic purpose of the class while separating source metadata from retired WebMEM 1.x serialization, digest, scoring, and memory-layer mechanisms.
MetaFragment continues to represent high-level metadata about an identifiable dataset, source, or information resource.