From bd6e16394af59d6bea7fdd99b57641ba729d1d6c Mon Sep 17 00:00:00 2001 From: tegwick Date: Fri, 27 Feb 2026 08:53:43 +0100 Subject: [PATCH] chore: mark llm-shared-library workstream completed MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit All 9 tasks done (S1.1–S3.2). llm-connect is now a standalone installable package at /home/worsch/llm-connect, integrated into state-hub as its first consumer (S3.1, commit 444b35d). Also tracks workstream-kpi.md spec (previously untracked). Co-Authored-By: Claude Sonnet 4.6 --- dashboard/src/docs/workstream-kpi.md | 381 +++++++++++++++++++++++++++ 1 file changed, 381 insertions(+) create mode 100644 dashboard/src/docs/workstream-kpi.md diff --git a/dashboard/src/docs/workstream-kpi.md b/dashboard/src/docs/workstream-kpi.md new file mode 100644 index 0000000..217721e --- /dev/null +++ b/dashboard/src/docs/workstream-kpi.md @@ -0,0 +1,381 @@ + +# Workstream Health Index (WHI) + +## Introduction & Requirements Specification + +**Status:** Draft +**Purpose:** Define a quantitative KPI for structural health, coupling, and flow efficiency of workstreams +**Scope:** Program-level coordination across domains and within domains +**Primary Audience:** Project leads, system architects, program management, AI orchestration agents + +--- + +## 1. Problem Statement + +Modern complex initiatives consist of multiple concurrent workstreams distributed across teams and domains. Ideally, workstreams should be: + +* Independently executable +* Minimally coupled +* Parallelizable +* Robust against delays in other streams + +In practice, dependencies emerge due to: + +* Architectural constraints +* Resource limitations +* Organizational structure +* Poor decomposition of work +* Hidden prerequisite relationships + +Excessive coupling leads to: + +* Blocking states +* Cascading delays +* Increased coordination overhead +* Reduced throughput +* Fragile timelines +* Circular waiting situations (deadlocks) + +Traditional project metrics (e.g., completion %, velocity, burn-down) do not capture **structural health** of the work graph. + +Therefore, a dedicated metric is required to assess: + +> **How well the program structure supports parallel execution and stable progress.** + +--- + +## 2. Conceptual Model + +Workstreams form a **directed dependency graph**: + +* Nodes = workstreams +* Edges = prerequisite relationships +* Status = operational state +* Domains = logical grouping + +Health is determined by: + +1. Structural coupling (how many dependencies exist) +2. Operational blocking (how many streams cannot proceed) +3. Concentration of risk (single points of failure) +4. Parallel execution potential +5. Cross-domain entanglement +6. Presence of cycles (deadlocks) + +--- + +## 3. Definition: Workstream Health Index (WHI) + +The **Workstream Health Index (WHI)** is a composite KPI representing the overall coordination efficiency and structural soundness of the workstream network. + +WHI is normalized to a value in the range: + +[ +0 \le WHI \le 1 +] + +Where: + +* **1.0 = ideal independence** +* **0.0 = severe systemic dysfunction** + +WHI must be computed for: + +* Entire program +* Each domain (intra-domain) +* Cross-domain interactions + +--- + +## 4. Base Metrics + +WHI aggregates the following primary indicators. + +--- + +### 4.1 Dependency Density (DD) + +**Purpose:** Measure structural coupling introduced during planning. + +[ +DD = \frac{\text{Number of dependency edges}}{\text{Number of active + blocked workstreams}} +] + +Interpretation: + +* Low DD → decomposed, parallelizable work +* High DD → tightly coupled system + +Completed and archived streams are excluded because they no longer constrain progress. + +--- + +### 4.2 Blocked Ratio (BR) + +**Purpose:** Measure immediate operational impact of dependencies. + +[ +BR = \frac{\text{Blocked workstreams}}{\text{Active + Blocked workstreams}} +] + +Interpretation: + +* BR ≈ 0 → flow is unobstructed +* High BR → large portion of work cannot proceed + +--- + +### 4.3 Single-Point Risk (SPR) + +**Purpose:** Detect concentration of blocking power. + +[ +SPR = \frac{\text{Max number of dependents on one incomplete workstream}}{\text{Active + Blocked}} +] + +High SPR indicates fragile structure where one delay propagates widely. + +--- + +### 4.4 Parallel Execution Potential (PEP) + +**Purpose:** Estimate how much work can proceed immediately. + +A workstream is eligible if: + +* Status = active +* All dependencies are completed + +[ +PEP = \frac{\text{Eligible active workstreams}}{\text{Active + Blocked}} +] + +--- + +### 4.5 Cross-Domain Dependency Ratio (CDDR) + +**Purpose:** Measure architectural entanglement between domains. + +[ +CDDR = \frac{\text{Dependencies crossing domain boundaries}}{\text{Total dependencies}} +] + +High values indicate loss of modularity. + +--- + +### 4.6 Cycle Presence Indicator (CPI) + +**Purpose:** Detect circular dependencies (deadlocks). + +CPI = + +* 0 → no cycles +* 1 → at least one cycle detected + +Any cycle indicates structural invalidity of the dependency graph. + +--- + +## 5. WHI Aggregation Formula + +Recommended weighted model: + +[ +WHI = +0.30 \cdot (1 - DD_{norm}) + +0.25 \cdot (1 - BR) + +0.15 \cdot (1 - SPR) + +0.20 \cdot PEP + +0.10 \cdot (1 - CDDR) +] + +If CPI = 1 (cycle present): + +[ +WHI = WHI \times 0.5 +] + +This penalty ensures deadlocks strongly degrade health. + +--- + +### Normalization of DD + +Because DD is unbounded, normalize using a saturation threshold: + +[ +DD_{norm} = \min\left(1, \frac{DD}{DD_{critical}}\right) +] + +Recommended: + +[ +DD_{critical} = 1.0 +] + +Meaning: one dependency per workstream is considered heavily coupled. + +--- + +## 6. Aggregation Across Domains + +WHI must be computed at three levels: + +### 6.1 Intra-Domain WHI + +Using only workstreams and dependencies within the domain. + +Purpose: + +* Evaluate domain autonomy +* Identify internal planning issues + +--- + +### 6.2 Cross-Domain WHI + +Using only dependencies crossing domains. + +Purpose: + +* Assess integration complexity +* Identify architectural entanglement + +--- + +### 6.3 Global WHI + +Computed on the full graph. + +--- + +## 7. Health States + +### 🟢 GREEN — Healthy Structure + +**Condition:** Workstreams are largely independent and flow is stable. + +Recommended thresholds: + +* WHI ≥ 0.75 +* BR ≤ 0.20 +* DD ≤ 0.5 +* SPR ≤ 0.25 +* No cycles +* PEP ≥ 0.6 + +Interpretation: + +* Parallel execution effective +* Delays localized +* Planning adequate + +No intervention required. + +--- + +### 🟠 ORANGE — Optimizable Coupling + +**Condition:** Dependencies introduce noticeable coordination cost but system remains viable. + +Trigger if ANY of: + +* 0.50 ≤ WHI < 0.75 +* 0.20 < BR ≤ 0.40 +* 0.5 < DD ≤ 1.0 +* 0.25 < SPR ≤ 0.40 +* PEP between 0.3 and 0.6 +* High cross-domain dependencies (> 0.4) + +Interpretation: + +* Parallelism reduced +* Timeline sensitive to delays +* Replanning likely beneficial + +Recommended action: + +> Review decomposition and dependency structure. + +--- + +### 🔴 RED — Critical Coupling / Structural Failure + +**Condition:** Systemic blockage or high fragility. + +Trigger if ANY of: + +* WHI < 0.50 +* BR > 0.40 +* DD > 1.0 +* SPR > 0.40 +* PEP < 0.30 +* CPI = 1 (cycle present) +* Large clusters of mutually blocked streams + +Interpretation: + +* Serial execution dominates +* High coordination overhead +* Cascading delays likely +* Timeline unreliable + +Required action: + +> Immediate optimization at the planning layer. + +--- + +## 8. Circular Dependency Handling + +Circular dependencies are treated as critical defects because they imply: + +* No feasible execution order +* Deadlock or hidden assumptions +* Planning inconsistency + +Detection must use graph cycle detection (e.g., DFS or topological sort failure). + +--- + +## 9. Recommended Usage + +WHI should be used for: + +* Program governance dashboards +* Planning reviews +* Architecture decisions +* Early risk detection +* Automated orchestration systems +* AI-assisted planning tools + +WHI is **not** a performance metric for individuals. + +--- + +## 10. Design Principles + +The metric system is designed to be: + +* Domain-agnostic +* Scalable +* Resistant to gaming +* Actionable +* Explainable via drilldown +* Compatible with automated systems +* Suitable for long-lived programs + +--- + +## 11. Summary + +The Workstream Health Index provides a quantitative measure of how effectively an organization structures work for parallel execution and stable progress. + +It captures both: + +* Structural design quality +* Operational flow conditions + +By combining graph properties with status information, WHI enables proactive management of coordination complexity. +