stage 1 initial exploration

---
layout: raw-data.njk
title: "stage 1 initial exploration"
---

# Stage 1: Initial Orientation and Semantic Exploration

## Date and Agent
- Date: 2025-11-19
- Agent: Claude (general-purpose agent)

## Objective

Understand the identification standards landscape through systematic exploration of the MCP server and available data assets. This stage aims to familiarize with the MCP server capabilities, document collection structure, asset organization, and perform initial semantic searches to understand content distribution and relationships across the 30-document collection.

## Methodology

### Tools and Approaches Used

1. **MCP Server Exploration**:
   - Used `get_document_stats` to understand collection size and structure
   - Used `get_schema` to identify available node properties and relationships
   - Used `run_cypher_query` to identify all 30 documents and their node counts
   - Performed 6 semantic searches on key topics using `semantic_search` tool with min_score=0.75

2. **File System Exploration**:
   - Read LLM_MCP_SERVER_GUIDE.md (507 lines) to understand MCP server capabilities and best practices
   - Read ProjectDirectoryTree.md to understand project structure
   - Examined markdown document samples to understand custom markdown style
   - Reviewed TomNotesManualReview.md to understand manual review findings
   - Examined annotation set structure (JSON format)

3. **Asset Organization Review**:
   - Explored MarkdownVersionsOfDocRefDocuments/ directory structure (26 document folders)
   - Examined ChecklistsAndTablesFromConformingPageIdentificationStandards/ (6 Word documents)
   - Reviewed annotation data organization (23 JSON annotation files)

## Key Findings

### MCP Server Capabilities and Structure

The `identification-management-standards` MCP server provides comprehensive access to New Zealand's identification management standards through a Neo4j graph database with the following characteristics:

**Database Statistics**:
- **9,374 total DocumentNode entities** across 30 documents
- **7,454 nodes with embeddings** (79.5% coverage using 768-dimensional vectors)
- **Two relationship types**:
  - CHILD_OF: 9,090 hierarchical parent-child relationships
  - SEMANTIC_SIMILARITY: 74,485 precomputed semantic similarity links (K=10 neighbors per node)
- **229 virtual nodes** created to fill gaps in document hierarchies

**Content Categories** (by node count):
1. text (3,384 nodes) - Primary content and paragraphs
2. structural (1,065) - Sections, parts, and structural elements
3. metadata (669) - Document metadata and descriptions
4. list (611) - List items and enumerated content
5. virtual (229) - Virtual nodes for missing hierarchy parents
6. example (178) - Example scenarios and use cases
7. definition (107) - Term definitions and glossary entries
8. table (106) - Tabular content and data
9. figure (48) - Diagrams, charts, and figures
10. heading (37) - Section headings and titles
11. root (12) - Document root nodes

**Key Node Properties Available**:
- uri, url - Canonical identifiers and full URLs for citations
- content - Document text
- type, category - Node classification
- level - Hierarchy depth
- document - Document URI prefix
- docrefId - Original hierarchical ID (e.g., "part1-section2-para1")
- isVirtual - Boolean flag for virtual nodes
- parentUri, ancestorUris - Hierarchical relationships
- embeddings - 768-dimensional float arrays
- tags, annotations - Metadata as JSON strings

**MCP Server Tools Assessment**:
The server provides 7 query tools with different performance characteristics:
- `semantic_search` - Natural language queries (slowest, requires API call)
- `find_semantic_neighbors` - Fast graph traversal of precomputed similarities
- `search_by_document` - Retrieve all nodes from specific document
- `get_hierarchical_context` - Navigate parent/child/sibling relationships
- `run_cypher_query` - Custom Neo4j queries for advanced analysis
- `get_schema` and `get_document_stats` - Metadata and statistics

### Document Collection Overview

The database contains **30 documents** organized into several distinct categories:

**Four Core Identification Standards** (text CANNOT be modified, only structure):
1. **Federation Assurance Standard** (431 nodes) - Requirements for credential presentation by facilitation providers ([DocRef](https://docref.digital.govt.nz/nz/identification-management/federation-assurance-standard/2025/en/#h1))
2. **Information Assurance Standard** (169 nodes) - Robustness of processes to establish quality and accuracy of information ([DocRef](https://docref.digital.govt.nz/nz/identification-management/information-assurance-standard/2024/en/#h1))
3. **Authentication Assurance Standard** (328 nodes) - Controls to ensure authenticators remain in control of authorized holder ([DocRef](https://docref.digital.govt.nz/nz/identification-management/authentication-assurance-standard/2024/en/#h1))
4. **Binding Assurance Standard** (161 nodes) - Robustness of processes to bind entity to information ([DocRef](https://docref.digital.govt.nz/nz/identification-management/binding-assurance-standard/2024/en/#h1))

**Four Implementation Guides** (paired with each core standard):
5. **Implementing the Federation Assurance Standard** (438 nodes) - Largest implementation guide ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-federation-assurance-standard/2025/en/#h1))
6. **Implementing the Authentication Assurance Standard** (280 nodes) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-authentication-assurance-standard/2024/en/#h1))
7. **Implementing the Information Assurance Standard** (214 nodes) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-information-assurance-standard/2024/en/#h1))
8. **Implementing the Binding Assurance Standard** (158 nodes) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-binding-assurance-standard/2024/en/#h1))

**Foundational and Reference Materials**:
9. **Identification Terminology** (383 nodes) - Agreed and evolving terms with dictionary and international sources ([DocRef](https://docref.digital.govt.nz/nz/identification-management/identification-terminology/2025/en/#h1))
10. **Levels of Assurance** (90 nodes) - Overview of assurance levels across standards
11. **Overview of the Identification Standards** (51 nodes) - High-level introduction
12. **Identification Standards** (26 nodes) - Landing/navigation page

**Guidance and Supporting Materials**:
13. **Assessing Identification Risk** (225 nodes) - Risk assessment methodology
14. **Conforming with the Identification Standards** (205 nodes) - Conformance process and assessment procedures
15. **Counter Fraud Techniques** (212 nodes) - Fraud detection and prevention
16. **Authenticator Types** (204 nodes) - Types of authentication factors
17. **Derived Information** (177 nodes) - Using derived values from information
18. **Authority to Act for Another Entity** (154 nodes) - Delegation and representation
19. **Using Documents as Evidence** (125 nodes) - Document verification
20. **About Identification Management** (60 nodes) - Introduction and context
21. **Guidance** (32 nodes) - Navigation page for guidance materials
22. **Training and Clinics** (45 nodes) - Training resources
23. **Resource Material Evidence of Identity Standard** (32 nodes) - Historical reference
24. **Superseded Standards** (20 nodes) - Historical versions
25. **Contact the Identification Management Team** (12 nodes) - Contact information
26. **Digital Identity Programme** (6 nodes) - Related programme information

**Related Legal/Regulatory Documents** (not identification management core):
27. **Privacy Act 2020** (3,611 nodes) - nz/pri/193/en/ - Largest document in collection
28. **Digital Identity Services Trust Framework Act 2023** (935 nodes) - nz/distf/14/en/
29. **Digital Identity Services Trust Framework Regulations** (242 nodes) - nz/distfr/13/en/
30. **Digital Identity Services Trust Framework Rules** (348 nodes) - nz/dia-distfr/2/en/

### Semantic Search Results by Topic

#### 1. Authentication Methods and Requirements

Search query: "authentication methods and requirements for verifying credentials"

**Key findings**: Authentication requirements are concentrated in the Authentication Assurance Standard with strong semantic connections to:
- Requirements for Authentication section within the standard (score: 0.933) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/authentication-assurance-standard/2024/en/#part4-title))
- Scope explicitly related to authenticators and authentication processes (score: 0.890) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/authentication-assurance-standard/2024/en/#part1-para3))
- Conformance processes for establishing authenticators (score: 0.887) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/conforming-with-the-identification-standards/2025/en/#part3-subpart1-section2-tb1-tr2-td1-line5))
- Trust Framework Rules requiring credential verification (scores: 0.886) ([DocRef](https://docref.digital.govt.nz/nz/dia-distfr/2/en/#part2-rule8-para4))
- Definition of authentication process from Authenticator Types guidance (score: 0.886) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/authenticator-types/2021/en/#part2-para1))

**Pattern observed**: Authentication content spans both normative standards and guidance documents, with Trust Framework legal documents also containing authentication requirements that may need alignment.

#### 2. Binding Assurance Processes

Search query: "binding assurance processes for linking credentials to identity"

**Key findings**: Binding assurance content is distributed across multiple documents:
- Increasing assurance levels in binding processes (score: 0.894) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/binding-assurance-standard/2024/en/#part1-subpart2-para3-3))
- Trust Framework Rules definition of binding assurance (score: 0.893) ([DocRef](https://docref.digital.govt.nz/nz/dia-distfr/2/en/#part1-rule4-para9))
- Implementation guidance explaining binding reduces identity theft (score: 0.886) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-binding-assurance-standard/2024/en/#part1-para2))
- Federation Standard's objective on maintaining binding assurance (score: 0.884) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/federation-assurance-standard/2025/en/#part5-subpart2))
- Conformance guidance references to Binding Assurance Standard (multiple links)

**Pattern observed**: Binding assurance is a crosscutting concern referenced in Federation Standard, suggesting interdependencies between standards that could inform restructuring.

#### 3. Federation Standards

Search query: "federation standards for identity providers and service providers"

**Key findings**: Federation content has strong connections to Trust Framework legal documents:
- Trust Framework provider requirements for digital identity services (score: 0.892) ([DocRef](https://docref.digital.govt.nz/nz/distf/14/en/#P2-s12-title))
- Trust Framework Rules requiring compliance with Federation Assurance Standard (score: 0.890) ([DocRef](https://docref.digital.govt.nz/nz/dia-distfr/2/en/#part2-rule9-para4))
- Implementing Federation Assurance Standard applies to credential and facilitation providers (score: 0.885) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-federation-assurance-standard/2025/en/#part1-para2))
- Conformance guidance for facilitation providers (score: 0.885) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/conforming-with-the-identification-standards/2025/en/#part3-subpart1-section2-tb1-tr3))

**Pattern observed**: Federation Standard is the primary linkage point between identification management standards and the Digital Identity Services Trust Framework (DISTF) legal regime. This relationship should be prominent in restructured materials.

#### 4. Conformance Processes

Search query: "conformance processes and compliance assessment procedures"

**Key findings**: Conformance guidance is consolidated in a single document but has lower semantic scores (0.85-0.88 range):
- Guidance describes types of conformance and assessment process (score: 0.882) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/conforming-with-the-identification-standards/2025/en/#h1-subtitle))
- Three key stages in formal conformance process (score: 0.881) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/conforming-with-the-identification-standards/2025/en/#part3-para1))
- Two assessment types: qualified (lighter) and audited (robust with statement) (scores: 0.878-0.861) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/conforming-with-the-identification-standards/2025/en/#part3-subpart3-section6-para1-2))
- Checklists and templates for documenting evidence (score: 0.854) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/conforming-with-the-identification-standards/2025/en/#part3-subpart2-section5))

**Pattern observed**: Conformance process is somewhat isolated from other standards content semantically, suggesting it may be "tucked away" as Tom's notes indicate. However, conformance is likely the primary reason users engage with standards.

#### 5. Privacy and Security Requirements

Search query: "privacy and security requirements for identity information"

**Key findings**: Privacy requirements are distributed across multiple documents with strong emphasis on Privacy Act compliance:
- Trust Framework Rules define privacy and confidentiality requirements (score: 0.893) ([DocRef](https://docref.digital.govt.nz/nz/dia-distfr/2/en/#part1-rule4-para35))
- Privacy Act identity information access requirements (score: 0.885) ([DocRef](https://docref.digital.govt.nz/nz/pri/193/en/#part7-subpart2-s168-subs2-b))
- Trust Framework Act requirements for dealing with personal/organizational information (score: 0.884) ([DocRef](https://docref.digital.govt.nz/nz/distf/14/en/#P2-s12))
- Information Assurance Standard Objective 2 on protecting information (score: 0.880) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/information-assurance-standard/2024/en/#part3-subpart1))
- Standards work together to prevent identity theft, fraud and loss of privacy (score: 0.874) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/identification-standards/2025/en/#h1-subtitle))

**Pattern observed**: Privacy is a fundamental concern but requirements are fragmented across identification standards, Privacy Act, and Trust Framework documents. Tom's notes suggest linking to external privacy/security standards "without specificity is not helpful."

#### 6. Information Assurance Frameworks

Search query: "information assurance frameworks and controls"

**Key findings**: Information assurance implementation guidance scores highest:
- Implementing Information Assurance Standard subtitle (score: 0.911) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-information-assurance-standard/2024/en/#h1-subtitle))
- Guidance materials navigation reference (score: 0.911) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/guidance/2025/en/#subpart2-para1))
- Controls for levels of information assurance (score: 0.907) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/derived-information/2024/en/#part4-det8-para2))
- Trust Framework Rules definition of information assurance (score: 0.877) ([DocRef](https://docref.digital.govt.nz/nz/dia-distfr/2/en/#part1-rule4-para24))
- Information Assurance Standard itself (score: 0.871) ([DocRef](https://docref.digital.govt.nz/nz/identification-management/information-assurance-standard/2024/en/#h1))

**Pattern observed**: Implementation guidance is semantically closer to user queries than the standard itself, suggesting users may find guidance more accessible than normative standards. This supports consolidation approach.

### Document Categories and Taxonomy

Based on MCP server exploration and file system review, the 30 documents can be categorized into a clear taxonomy:

**1. Core Normative Standards** (4 documents - text cannot be modified)
- Federation Assurance Standard
- Information Assurance Standard
- Authentication Assurance Standard
- Binding Assurance Standard

**2. Implementation Guides** (4 documents - paired 1:1 with standards)
- Implementing the [Standard Name]
- These provide guidance on complying with controls

**3. Foundational Reference** (4 documents)
- Identification Terminology (definitions and evolving terms)
- Levels of Assurance (crosscutting framework)
- Overview of the Identification Standards
- Identification Standards (navigation/landing)

**4. Process Guidance** (9 documents)
- Assessing Identification Risk
- Conforming with the Identification Standards
- Counter Fraud Techniques
- Authenticator Types
- Derived Information
- Authority to Act for Another Entity
- Using Documents as Evidence
- About Identification Management
- Guidance (navigation page)

**5. Supporting Resources** (3 documents)
- Training and Clinics
- Resource Material Evidence of Identity Standard (historical)
- Superseded Standards (historical)
- Contact information

**6. Related Legal Framework** (4 documents - external)
- Privacy Act 2020 (3,611 nodes - largest document)
- Digital Identity Services Trust Framework Act, Regulations, and Rules

**7. Administrative** (2 documents)
- Contact and Digital Identity Programme pages (minimal content)

### Asset Organization and Relationships

**Markdown Documents** (MarkdownVersionsOfDocRefDocuments/):
- 26 folders corresponding to identification management documents (27-30 are external legal documents)
- Each folder contains dated markdown file: `YYYY--YYYY-MM-DD--en.md`
- Some folders contain `/files/` subdirectory for images (e.g., levels-of-assurance.png)
- These demonstrate the **custom markdown style** required for Phase 2 output

**Custom Markdown Style Observations**:
From examination of identification-terminology and authentication-assurance-standard samples:
- Clean heading hierarchy using `#`, `##`, `###`
- Tables using standard markdown table syntax with alignment
- Bold text using `**text**`
- Italic using `*text*`
- Internal links: `[Link text](/relative/path/)` or `[Link text](full-url/)`
- Image syntax with figure wrapper: `::: fig` ... `:::`
- Detail expanders using `+++` delimiter (Tom wants these removed)
- Unordered lists with `*` or `-`
- Ordered lists standard
- Emphasis on scannable structure with clear headings
- Metadata subtitle line using `###` after main heading

**JSON Annotation Sets** (ManualReviewIdentificationStandardsAnnotationSets/):
- 23 JSON files for identification management documents (no annotations for legal docs)
- Structure: `{"name": "...", "jurisdiction": "nz", "slug": "...", "language": "en", "version": "YYYY", "au": "Tom Barraclough", "annotations": [...]}`
- Each annotation has: `id`, `bid` (block ID matching docrefId), `a` (annotation text), `au` (author), plus empty arrays for tags/users/notes/replies
- Annotations provide granular feedback on specific document sections
- Example annotation types: questions about authority/process, suggestions for clarity, critique of passive voice, concerns about terminology

**Checklists and Tables** (ChecklistsAndTablesFromConformingPageIdentificationStandards/):
- 6 Word documents (.docx format) for conformance assessment
- Authentication Factor Level Table
- Conformance Checklist - Authentication Assurance
- Conformance Checklist - Credential Establishment
- Conformance Checklist - Facilitation Mechanisms
- Conformance Checklist - Information & Binding Assurance
- Levels of Assurance Table
- These are referenced in conformance guidance but stored separately
- Should be integrated or converted to markdown in restructuring

**DocRef JSON Files** (DocRefJSONFiles/):
- Raw JSON datasets organized by document collections
- Three subdirectories: DISTFActRegsRules/, IdentificationStandardsAndGuidance/, PrivacyAct/
- These are the source data loaded into the MCP server Neo4j database
- Contain full text, metadata, URLs, IDs, and hierarchical structure
- Can be used to verify MCP server data or examine raw structure if needed

**Relationship Between Assets**:
- Markdown files in MarkdownVersionsOfDocRefDocuments/ correspond to documents in MCP server
- Document URIs from MCP server map to folder names in markdown directory
- Annotation JSON files use `bid` field to reference `docrefId` property in MCP server nodes
- TomNotesManualReview.md provides high-level synthesis of annotation findings
- Checklists reference controls from standards but are separate resources

## Supporting Evidence

### Evidence of Document Fragmentation

Tom's manual review notes identify fragmentation as a key issue:
> "Standards are standards. Guidance is guidance. Does having separate pages really help? It's not for reading, it's for working through." ([Source](file:///Users/tombarraclough/projects/official/IdentificationStandards19November2025/ManualReviewIdentificationStandardsAnnotationSets/TomNotesManualReview/))

The MCP server data confirms this: each of the 4 core standards (avg 272 nodes) has a paired implementation guide (avg 272 nodes), creating 8 separate documents for 4 topics. Additional process guidance adds 9 more documents.

### Evidence of Passive Voice Usage

Tom's notes repeatedly emphasize this issue:
> "A lot of passive voice rather than active voice - ie 'Credential enrolment' rather than 'X enrols the credential'."
> "The whole document is passive but should really be addressed to the person reading it - ie, 'When you or your organisation are doing X, this is what you must do:'"

Sample from Authentication Assurance Standard confirms passive construction:
- "This standard applies to any Relying Party (RP)."
- "Application of the controls in this standard will contribute to..."
- "The scope of the requirements in this standard is explicitly related to..."

### Evidence of Conformance Process Visibility Issue

Tom's notes state:
> "It feels like the process of conforming with the standards is the whole point of even reading the standards in the first place. People are coming to the standards to try and conform with them, or to understand others' conformance with them. It's not clear why this is tucked away."

The semantic search on conformance returned scores in the 0.85-0.88 range (lower than other topics at 0.89-0.93), suggesting conformance content may be less semantically connected to other materials.

### Evidence of Privacy Act Integration Issue

Tom's notes observe:
> "In the guidance materials on other pages, a lot is made of the fact that identification and privacy overlap but aren't the same, but then a lot of the standards are explicitly stated to be an application of an information privacy principle. That seems inconsistent and makes things harder than they need to be in terms of comprehension and familiarity."

Information Assurance implementation guidance confirms this: "This is the application of Information privacy principle 1 of the Privacy Act 2020" ([DocRef](https://docref.digital.govt.nz/nz/identification-management/implementing-the-information-assurance-standard/2024/en/#part2-subpart1-section3))

### Evidence of Terminology Authority Concerns

Tom's notes question terminology approach:
> "Not clear why dictionary definitions are useful in such a specialised area. Also, if DIA is the identification authority in some way, then it can fairly be empowered to declare its own definitions as an authority, and all it needs to do is disambiguate terms or enable standardisation with other external standards."

Annotation on terminology page:
> "The function of this page needs some thought. If DIA publishes the standards, DIA can declare the meaning of terms. The meanings should be consistent with other relevant standards and instruments, including the law, but relying on a dictionary definition for such a specialised area is not a useful source of authority."

### Evidence of Detail Expander Usage

Tom's notes clearly state:
> "Get rid of all detail expanders"

Authentication Assurance Standard markdown shows detail expander syntax:
```
+++ Detailed description of diagram
[content]
+++
```

## Initial Observations

### 1. Clear Document Hierarchy Exists

The collection has a well-defined structure:
- 4 core standards that cannot have text modified
- 4 paired implementation guides that can be restructured
- Foundational reference materials (terminology, levels of assurance)
- Process guidance that can be consolidated
- External legal framework that provides context

This hierarchy should inform the Phase 2 restructuring approach.

### 2. Standards-Guidance Separation Creates Friction

Each core standard (avg 272 nodes) is paired with an implementation guide (avg 272 nodes) of similar size. Semantic searches show implementation guidance often scores higher than standards for user queries, suggesting users find guidance more accessible.

Tom's annotations and notes consistently question this separation. The guidance is authoritative (created by standards maker) and linked from standards, so there's no regulatory reason for separation.

**Implication**: Phase 2 should consider how to integrate standards and guidance while maintaining clear distinction between normative requirements and explanatory material.

### 3. Conformance Is Central But Not Prominent

Tom's notes correctly identify that conformance is likely the primary user goal. The conformance guidance is comprehensive (205 nodes) with practical checklists, but:
- It's a separate document rather than integrated throughout
- Semantic searches show lower connectivity (0.85-0.88 scores vs 0.89-0.93 for other topics)
- Checklists are Word documents rather than integrated markdown

**Implication**: Phase 2 restructuring should make conformance process more prominent, possibly as an organizing framework for the entire resource.

### 4. Trust Framework Connection Is Critical

The Digital Identity Services Trust Framework (DISTF) Act, Regulations, and Rules are the only cited mandatory use of the identification standards. Tom notes:
> "The standards are at pains to distinguish themselves from the DISTF framework, but when it comes down to it, the only mandatory use of the standards is the DISTF. That seems self-defeating."

Federation Standard is the primary linkage point (highest semantic similarity with DISTF legal documents). This relationship should be clarified and made prominent in restructured materials.

### 5. Privacy Act Integration Needs Clarification

Privacy Act 2020 is the largest document in the database (3,611 nodes, 38.5% of all nodes). Multiple identification standards explicitly reference Privacy Act principles (e.g., Information Assurance Standard controls reference IPP1).

However, materials attempt to distinguish identification management from privacy, creating potential confusion. The relationship needs clearer articulation.

### 6. Passive Voice Pervades Standards

Tom's notes and sample reading confirm extensive passive voice usage in standards. Examples:
- "This standard applies to..." vs "You must apply this standard when..."
- "Application of the controls will contribute..." vs "Apply these controls to..."
- "The scope of requirements is related to..." vs "These requirements cover..."

**Implication**: Guidance materials (not core standards) can be rewritten in active voice for Phase 2. Core standards text cannot change, but structure and presentation can be improved.

### 7. Terminology Page Function Unclear

The terminology page provides dictionary definitions and international standard references, but Tom's annotations question its value:
- Dictionary definitions may not be authoritative enough for specialized domain
- Terminology should help users understand complex terms, not just define them
- Examples and plain language explanations are lacking
- Some terms are marked "evolving" but unclear why

**Implication**: Phase 2 should reconsider how terminology supports user understanding vs serving as authoritative reference.

### 8. Detail Expanders Should Be Removed

Tom explicitly wants detail expanders removed. The markdown syntax `+++` ... `+++` creates collapsible sections that hide content. This conflicts with goal of making information findable and scannable.

**Implication**: Phase 2 restructuring should eliminate detail expanders, using clear heading hierarchy instead.

### 9. MCP Server Is Highly Capable

The MCP server provides excellent capabilities for Phase 2 content retrieval:
- 79.5% embedding coverage enables semantic search
- 74,485 precomputed similarity relationships enable fast neighbor finding
- Hierarchical relationships support context retrieval
- Full Cypher query capability enables advanced analysis
- All results include DocRef citations for traceability

**Implication**: Phase 2 content retrieval can be systematic and comprehensive using MCP server tools.

### 10. Custom Markdown Style Is Consistent

The markdown documents demonstrate consistent style conventions:
- Clean heading hierarchy
- Standard table and list syntax
- Clear image integration with figure wrappers
- Internal and external linking patterns
- Scannable structure

**Implication**: Style guide can be extracted from examples for Phase 2 writing (Stage 8).

## Questions and Uncertainties

### 1. Electronic Identity Verification Act Integration

Tom's notes ask:
> "Should we add the electronic identity verification act and regulations? They only seem to be referenced in the federation standard. Also, the identify verification act and regs don't seem to acknowledge the existence of the identification standards at all."

**Question for Stage 3**: Are EIVA requirements essential for identification management practice? Does conformance with identification standards satisfy EIVA obligations? Should EIVA be integrated or just referenced?

### 2. Extent of Standards-Guidance Integration

While Tom's feedback suggests integration, the extent is unclear:
- Should standards and guidance be in same document with clear visual distinction?
- Should guidance be inserted alongside relevant controls?
- Should there be a consolidated guide separate from normative standards?

**Question for Stage 4-5**: What level of integration best serves user needs while maintaining regulatory clarity?

### 3. Terminology Approach

Tom questions dictionary-based definitions but solution is unclear:
- Should DIA assert authority to define terms?
- Should terminology align with DISTF legal definitions?
- Should plain language explanations replace formal definitions?
- Should examples be added to each term?

**Question for Stage 4**: How should terminology support both precision and understanding?

### 4. Conformance as Organizing Framework

Tom suggests conformance should be more central, but approach is unclear:
- Should restructured resource be organized around conformance process?
- Should checklists be integrated throughout rather than separate?
- Should levels of assurance expression be the primary organizing principle?

**Question for Stage 5**: How can conformance process organize restructured materials without overwhelming users?

### 5. Audience Segmentation

The standards serve multiple audiences (implementers, assessors, auditors, policy makers) but documents don't explicitly segment content:
- Should restructured materials have audience-specific pathways?
- Should technical vs non-technical content be distinguished?
- Should role-based views be created (e.g., Credential Provider, Relying Party, Facilitation Provider)?

**Question for Stage 4-5**: How should restructured materials serve diverse audiences?

### 6. Levels of Assurance Prominence

Levels of Assurance (LoA) is a foundational concept but has only 90 nodes:
- Should LoA be more prominent in restructured materials?
- Should controls be organized by LoA level?
- Should LoA expression be taught earlier?

**Question for Stage 4**: What role should LoA play in information architecture?

### 7. Historical Material Handling

Two documents contain historical/superseded content:
- Resource Material Evidence of Identity Standard (2021)
- Superseded Standards (2021)

**Question for Stage 6**: Should historical materials be included in restructured resource or archived separately?

### 8. Checklist Format and Integration

Six Word document checklists exist separately from standards:
- Should these be converted to markdown?
- Should they be integrated into main resource?
- Should they remain separate downloadable resources?

**Question for Stage 6**: How should checklists be handled in restructuring?

### 9. Diagram and Figure Approach

Standards include diagrams (e.g., element relationship diagram) using figure syntax:
- Are current diagrams effective?
- Should more diagrams be added?
- How should figures be referenced and integrated?

**Question for Stage 6**: What role should visual elements play in restructured materials?

### 10. Navigation Strategy

Current structure has many separate pages/documents:
- Should restructured resource be single document with good hierarchy?
- Should it be multi-document with clear navigation?
- How should cross-references and linking work?

**Question for Stage 6**: What navigation approach best serves users?

## Next Steps for Stage 2

Based on Stage 1 findings, Stage 2 (Cross-Document Pattern Analysis) should focus on:

1. **Deep Dive into Tom's Annotations**: Systematically review all 23 annotation sets to categorize feedback by theme (voice, structure, terminology, clarity, findability). The annotations provide granular feedback on specific issues.

2. **Analyze Previously Identified Issues**: Extract patterns from Tom's notes:
   - Passive voice → active voice opportunities (in guidance, not standards)
   - Content fragmentation → consolidation opportunities
   - Terminology inconsistencies → standardization needs
   - Structural problems → hierarchy improvements
   - Navigation issues → findability enhancements
   - Conformance visibility → prominence opportunities

3. **Use Semantic Neighbors to Trace Concepts**: Use `find_semantic_neighbors` on key nodes from Stage 1 semantic searches to:
   - Map where similar content appears across documents
   - Identify consolidation opportunities
   - Find content that should be linked but isn't
   - Understand cross-document dependencies

4. **Query for Structural Patterns**: Use Cypher queries to analyze:
   - Documents with highest cross-reference connectivity
   - Virtual nodes indicating hierarchy gaps
   - Content type distribution (definitions, examples, controls)
   - Depth and complexity of hierarchies
   - Identify over-fragmented vs under-structured areas

5. **Compare Document Types**: Analyze differences between:
   - Standards vs implementation guides (voice, structure, content design)
   - Process guidance vs standards (user focus, practicality)
   - Foundational reference vs operational guidance
   - Identify what works well and what needs improvement

6. **Map Improvement Opportunities**: Based on patterns identified:
   - Quick wins (e.g., remove detail expanders, improve headings)
   - Complex restructuring (e.g., standards-guidance integration)
   - Content consolidation (e.g., merge related guidance)
   - Navigation improvements (e.g., prominence of conformance)

This will provide the foundation for Stage 3 (evaluation of other materials) and Stage 4 (thematic synthesis).