diff --git a/README.md b/README.md
index c991c2c..ec26c01 100644
--- a/README.md
+++ b/README.md
@@ -1,257 +1,98 @@
-# Printer Profiling UI Project: Comprehensive Requirements & Architecture Roadmap
+# ICCery 🎨
 
-This document serves as the detailed feature set roadmap and architectural blueprint for integrating ArgyllCMS into a modern, native GUI application specifically dedicated to **Printer Profiling**. The goal is to abstract the complexities of the command-line utilities into a seamless, visual, and user-friendly experience.
+> Modern, cross-platform native desktop application for printer profiling, powered by ArgyllCMS.
+
+[![Release](https://img.shields.io/badge/version-v0.2.0-blue.svg)](https://git.i3omb.com/gronod/ICCery)
+[![Platform](https://img.shields.io/badge/platform-Windows%20%7C%20Linux-lightgrey.svg)](https://git.i3omb.com/gronod/ICCery)
+[![Framework](https://img.shields.io/badge/framework-Tauri%20v2%20%2B%20Rust-orange.svg)](https://tauri.app)
+[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
+
+**ICCery** is a native GUI frontend designed to make creating custom ICC/ICM printer profiles seamless, visual, and reliable. It wraps the powerful color management capabilities of [ArgyllCMS](https://www.argyllcms.com/) within an intuitive, artefact-gated 5-stage wizard.
 
 ---
 
-## 1. Architectural Overview & Licence Boundary
+## Key Features
 
-To maintain compliance with the ArgyllCMS AGPLv3 licence whilst allowing the host UI application to be closed-source, a strict inter-process communication (IPC) boundary must be maintained.
+- 🪄 **Linear 5-Stage Wizard Workflow**:
+  1. **Stage 1 — Patch Generation (`targen`)**: Configure RGB (driver-managed) or CMYK (RIP-managed) patch sets with custom counts, quality presets, and neutral/grey axis boosting.
+  2. **Stage 2 — Target Creation & Raw Printing (`printtarg`)**: Format patch targets for spectrophotometers (i1Pro, i1Pro2, ColorMunki, SpyderPrint). View high-resolution downscaled TIFF previews and print directly using native OS raw unmanaged pathways (Windows GDI uncorrected / Linux CUPS `raw`).
+  3. **Stage 3 — Interactive Measurement (`chartread`)**: Real-time instrument calibration prompts, interactive strip reading state machine, and a live swatch grid featuring instant per-patch CIEDE2000 ($\Delta E_{00}$) quality indicators.
+  4. **Stage 4 — Profile Calculation (`colprof`)**: Generate high-precision cLUT mathematical ICC/ICM profiles with configurable algorithm quality, descriptions, and copyright tagging.
+  5. **Stage 5 — Verification & 3D Gamut (`profcheck` + `iccgamut`)**: Comprehensive mathematical validation report (Peak, Average, RMS $\Delta E$) paired with an interactive 3D CIELAB color volume viewer and bundled sRGB reference wireframe comparison.
+- 🛡️ **Disk Artefact Gating**: Stepper navigation strictly verifies generated artefacts on disk (`.ti1`, `.ti2`, `.ti3`, `.icc`/`.icm`), preventing out-of-order execution while preserving backward navigation.
+- 🌐 **Platform-Aware**: Automatic handling of platform profile conventions (`.icm` on Windows, `.icc` on Linux/macOS) and native OS printer subsystems.
+- ⚖️ **Clean AGPL Boundary**: Complete isolation of AGPLv3 binaries via asynchronous tokio IPC process pipelines.
+
+---
+
+## Architectural Overview
+
+ICCery is built on **Tauri v2** and **Rust**, coupled with a reactive Vanilla JavaScript frontend and **Three.js** WebGL visualization:
 
 ```mermaid
 flowchart TD
-    subgraph host_app ["Host Application (Closed Source)"]
-        UI[Native User Interface]
-        State[State Machine & Orchestrator]
-        IO[Subprocess IPC Manager]
-        
+    subgraph Host ["ICCery Host (Tauri + Rust + Vanilla JS)"]
+        UI[Wizard UI & Swatch Grid]
+        ThreeJS[3D CIELAB Gamut Viewer]
+        State[Wizard State & Artefact Verifier]
+        PrintEngine[Raw Print Subsystem (GDI / CUPS)]
+        ProcMgr[Async Subprocess IPC Manager]
+
         UI <--> State
-        State <--> IO
+        State <--> ProcMgr
+        ProcMgr --> ThreeJS
+        UI --> PrintEngine
     end
-    
-    subgraph argyll_cms ["ArgyllCMS (AGPLv3)"]
+
+    subgraph Argyll ["ArgyllCMS Subprocesses (AGPLv3)"]
         BIN_TAR[targen]
         BIN_PRT[printtarg]
         BIN_CHR[chartread]
         BIN_COL[colprof]
+        BIN_CHK[profcheck]
+        BIN_GAM[iccgamut]
     end
-    
-    IO -- stdin/stdout/stderr pipes --> BIN_TAR
-    IO -- stdin/stdout/stderr pipes --> BIN_PRT
-    IO -- stdin/stdout/stderr pipes --> BIN_CHR
-    IO -- stdin/stdout/stderr pipes --> BIN_COL
-    
-    classDef hostStyle fill:#0f52ba,stroke:#000,stroke-width:2px,color:#fff;
-    classDef agplStyle fill:#800000,stroke:#000,stroke-width:2px,color:#fff;
-    
-    class UI,State,IO hostStyle;
-    class BIN_TAR,BIN_PRT,BIN_CHR,BIN_COL agplStyle;
-```
 
-### Core IPC Guidelines
-1. **No Library Linking**: The host application must never link against `libinst`, `libicc`, or any other Argyll C library.
-2. **Subprocess Management**: Binaries must be invoked via OS process spawning (e.g., `std::process::Command` in Rust, `System.Diagnostics.Process` in C#).
-3. **Data Passing**: Data is passed via file paths (e.g., `.ti1`, `.ti2`, `.ti3` files) and standard streams (`stdin` / `stdout`).
-
----
-
-## 2. Technology Stack Recommendations
-
-Given the requirements for a cross-platform (Windows, macOS, Linux), compiled (non-scripted), and closed-source compliant application, the following technology stacks are highly recommended:
-
-1.  **Rust with Tauri, Slint, or Iced**
-    *   **Pros**: Compiles to a highly optimised native binary, memory-safe, fearless concurrency for background tasks.
-    *   **Licence**: Permissive (MIT / Apache 2.0). Fully safe for closed-source.
-2.  **C# with .NET (Avalonia UI or MAUI)**
-    *   **Pros**: Strong object-oriented ecosystem, highly productive. Avalonia provides pixel-perfect identical rendering across all OSs.
-    *   **Licence**: Permissive (MIT). Fully safe for closed-source.
-3.  **C++ with wxWidgets or Qt**
-    *   **Pros**: Industry standard for cross-platform C++ UIs. Highly performant.
-    *   **Licence**: wxWidgets uses a modified LGPL that permits static linking in closed-source apps. Qt requires dynamic linking under LGPLv3 or purchasing a commercial licence.
-
-*(Note: Scripted frameworks like Python/PyQt and heavy web-wrappers like Electron/React are explicitly excluded per architectural constraints).*
-
----
-
-## 3. The Complete Profiling Workflow
-
-The application must guide the user through a linear wizard-style process.
-
-```mermaid
-sequenceDiagram
-    participant User
-    participant UI as Host UI
-    participant TAR as targen
-    participant PRT as printtarg
-    participant CHR as chartread
-    participant COL as colprof
-    
-    Note over User, COL: Stage 1: Patch Generation
-    User->>UI: Select RGB/CMYK & Patch Count
-    UI->>TAR: targen -v -d [colour_space] -f [count] target_name
-    TAR-->>UI: Generates .ti1 file
-    
-    Note over User, COL: Stage 2: Target Creation
-    User->>UI: Select Instrument & Page Size
-    UI->>PRT: printtarg -v -i [instr] -p [size] -t 300 target_name
-    PRT-->>UI: Generates .ti2 & .tif files
-    UI->>User: Display .tif for printing (Colour Management OFF)
-    
-    Note over User, COL: Stage 3: Measurement
-    User->>UI: User clicks "Start Reading"
-    UI->>CHR: chartread -v -u target_name
-    CHR-->>UI: Live JSON Stream (ROW_COLORS_JSON)
-    UI-->>User: Live Visual Updates (Swatches, Delta E)
-    CHR-->>UI: Generates .ti3 file upon completion
-    
-    Note over User, COL: Stage 4: Profile Calculation
-    User->>UI: Select Profile Quality & Name
-    UI->>COL: colprof -v -q [quality] target_name
-    COL-->>UI: Generates .icc / .icm profile
-    UI->>User: Display Gamut & Quality Report
+    ProcMgr -- stdin/stdout/stderr pipes --> BIN_TAR
+    ProcMgr -- stdin/stdout/stderr pipes --> BIN_PRT
+    ProcMgr -- stdin/stdout/stderr pipes --> BIN_CHR
+    ProcMgr -- stdin/stdout/stderr pipes --> BIN_COL
+    ProcMgr -- stdin/stdout/stderr pipes --> BIN_CHK
+    ProcMgr -- stdin/stdout/stderr pipes --> BIN_GAM
 ```
 
 ---
 
-## 4. Stage 1: Patch Set Creation (`targen`)
+## Building from Source
 
-The first step is configuring the numerical colour combinations to be printed.
+### Prerequisites
+- [Node.js](https://nodejs.org/) (v18 or newer)
+- [Rust](https://www.rust-lang.org/) (1.78+ stable)
+- Operating system dependencies:
+  - **Windows**: Microsoft Visual Studio C++ Build Tools & WebView2 runtime.
+  - **Linux (Debian/Ubuntu)**: `libwebkit2gtk-4.1-dev`, `build-essential`, `curl`, `wget`, `file`, `libxdo-dev`, `libssl-dev`, `libayatana-appindicator3-dev`, `librsvg2-dev`, `libcups2-dev`.
 
-### UI Configuration Details
-*   **Printer Colour Space**: Toggle between RGB (Driver-managed) or CMYK (RIP-managed).
-    *   *Command flag*: `-d 2` (RGB) or `-d 4` (CMYK).
-*   **Total Patch Count**: Controls profile accuracy vs. reading time. Provide presets (e.g., "Draft (400)", "Standard (800)", "High Quality (1500)", "Ultra (2500+)") and a custom input.
-    *   *Command flag*: `-f <number>`.
-*   **Neutral Axis Boosting**: Inject extra grey/black patches for smoother B&W printing.
-    *   *Command flag*: `-e <white_patches> -B <black_patches>`.
-
-### Execution
-The UI launches the subprocess and monitors `stderr` for completion.
+### Development Mode
 ```bash
-targen -v -d 2 -f 800 -e 4 -B 8 my_profile
+# Clone the repository
+git clone https://git.i3omb.com/gronod/ICCery.git
+cd ICCery
+
+# Install frontend dependencies
+npm install
+
+# Run the development app
+npm run tauri dev
 ```
 
----
-
-## 5. Stage 2: Target Image Generation (`printtarg`)
-
-Once the `.ti1` patch file exists, it is converted into a physical page layout formatted specifically for the user's measurement hardware.
-
-### UI Configuration Details
-*   **Hardware Instrument**: Dropdown list of supported devices.
-    *   *i1Pro/i1Pro2*: `-i i1`
-    *   *ColorMunki*: `-i CM`
-    *   *SpyderPrint/SpectroScan*: `-i SS`
-*   **Page Size**: Dropdown for media dimensions.
-    *   *Command flag*: `-p A4` or `-p Letter`.
-*   **Output Format**: Default to high-resolution TIFF for maximum compatibility.
-    *   *Command flag*: `-t 300` (8-bit) or `-T 300` (16-bit).
-
-### Execution & Printing Instructions
+### Production Build
 ```bash
-printtarg -v -i i1 -t 300 -p A4 my_profile
-```
-> [!IMPORTANT]
-> The UI must explicitly warn the user: **The generated TIFF files MUST be printed with colour management completely disabled in the printer driver.** (e.g., "No Colour Adjustment" on Epson, "Off (No Colour Adjustment)" on Canon). Failure to do so ruins the profile.
-
----
-
-## 6. Stage 3: Interactive Measurement (`chartread`)
-
-This is the most complex stage for the UI, requiring real-time bi-directional communication with `chartread` utilising the `-u` stream.
-
-### Interaction State Machine
-```mermaid
-stateDiagram-v2
-    [*] --> Initialize
-    Initialize --> AwaitCalibration: Spawn chartread -u
-    AwaitCalibration --> Calibrating: Send Spacebar
-    Calibrating --> AwaitStrip: Calibration OK
-    
-    AwaitStrip --> ReadingStrip: Wait for hardware trigger / button
-    ReadingStrip --> ProcessRow: Read success
-    ReadingStrip --> ErrorState: Read failed (Too fast/slow)
-    
-    ErrorState --> AwaitStrip: Retry
-    ProcessRow --> AwaitStrip: Row complete, parse JSON
-    ProcessRow --> Finished: Last row completed
-    
-    Finished --> [*]: .ti3 generated
-```
-
-### IPC Mechanics
-1. **Stdout Parsing**: The UI reads `stdout` line-by-line.
-    *   Lines matching `/^ROW_COLORS_JSON: (.*)/` are parsed as JSON to update the live patch grid and Delta E traffic lights.
-    *   Unprefixed lines (e.g., "Hit [Space] to read strip A") trigger state changes in the UI prompt text.
-2. **Stdin Injection**: The UI provides buttons (e.g., "Calibrate", "Skip") that inject keystrokes (`" \n"`, `"s\n"`) into the child process's `stdin`.
-
-### Visualising Delta E
-For live feedback, the JSON stream contains `expected.Lab` and `measured.Lab`. The UI should:
-1.  Implement the **CIEDE2000 ($\Delta E_{00}$)** formula.
-2.  Calculate the difference for every patch as it arrives.
-3.  Display a traffic light indicator (e.g., Green: $\Delta E < 2$, Amber: $\Delta E < 5$, Red: $\Delta E > 5$) on the UI swatch grid to instantly flag printing or reading errors.
-
----
-
-## 7. Stage 4: Profile Calculation (`colprof`)
-
-Once reading is finished and the `.ti3` file is generated, the mathematical ICC profile generation begins.
-
-### UI Configuration Details
-*   **Profile Quality**: Controls LUT resolution and optimisation passes.
-    *   *Command flag*: `-q l` (Low), `-q m` (Medium), `-q h` (High), `-q u` (Ultra). *Note: Ultra can take a very long time.*
-*   **Profile Description Name**: The internal name embedded in the ICC file that appears in Photoshop/Lightroom dropdowns.
-    *   *Command flag*: `-D "My Profile Name"`.
-*   **Algorithm**: Typically `cLUT` for printers.
-    *   *Command flag*: `-a l`.
-
-### Execution
-This process is highly CPU intensive. The UI must launch this in an asynchronous background thread and display an indeterminate progress spinner.
-```bash
-colprof -v -q h -a l -D "Canon Pro-100 Luster Paper" my_profile
+# Build desktop packages (MSI/NSIS on Windows, DEB/AppImage on Linux)
+npm run tauri build
 ```
 
 ---
 
-## 8. Display of Results & Verification
+## Licence
 
-### Post-Processing Integration
-1.  **Mathematical Quality Score (`profcheck`)**:
-    Silently execute `profcheck` to compare the generated profile against the measurement data.
-    ```bash
-    profcheck -k my_profile.ti3 my_profile.icc
-    ```
-    *Parse the output*: Extract "Peak error" and "Average error". A healthy profile typically has an average $\Delta E$ below 1.5. Display this clearly in the UI.
-2.  **3D Gamut Visualisation**: 
-    To provide a premium feel, integrate a native 3D WebGL (if using Tauri) or OpenGL (if using C++/C#) viewer. Parse the gamut boundary data to render a rotatable 3D plot of the printer's colour volume in CIELAB space.
-
----
-
-## 9. Proposed ArgyllCMS Upstream Modifications
-
-To ensure maximum value, stability, and responsiveness in the UI, several low-level features should be developed within the ArgyllCMS codebase. These modifications primarily revolve around structured data output (JSON) to prevent the UI from relying on fragile regex string parsing of human-readable console text.
-
-### 9.1 Implemented Enhancements
-*   **`chartread -u` (JSON Streaming)**: *Already developed and merged.* We added the `-u` flag to `chartread`, which emits a real-time, flush-forced `ROW_COLORS_JSON: { ... }` stream. This solved the critical UI requirement of displaying live swatches and computing Delta E ($\Delta E$) mid-measurement without waiting for the `.ti3` file to close.
-
-### 9.2 Recommended Future Enhancements
-
-#### `targen`: Structured Progress
-*   **Feature**: Add a `-u` flag to emit JSON progress events during generation (e.g., `{"event": "progress", "stage": "optimising", "percent": 45}`).
-*   **Rationale**: Advanced patch generation (especially iterative OFPS with adaptation) can take several minutes. A structured progress stream prevents the UI from appearing frozen.
-
-#### `printtarg`: Structured Target Manifest
-*   **Feature**: Add a `-u` flag to output a JSON manifest upon completion, detailing the exact filenames generated (e.g., `target_01.tif`, `target_02.tif`), their physical dimensions, and patch counts per page.
-*   **Rationale**: Currently, `printtarg` generates multiple pages implicitly. A UI requires this structured manifest to reliably load and display the correct number of pages for printing without guessing the filenames based on the basename.
-
-#### `colprof`: Structured Profile Calculation Progress
-*   **Feature**: Add a `-u` flag to emit JSON progress during profile calculation (e.g., `{"event": "gamut_mapping", "percent": 75}`).
-*   **Rationale**: High and Ultra quality profile generation is highly CPU-intensive and time-consuming. Parsing dots (`.`) or numbers from `stdout` for progress bars is unreliable; structured progress ensures smooth UI feedback.
-
-#### `profcheck`: Structured Delta E Report
-*   **Feature**: Add a `-u` flag to output the final $\Delta E$ report in JSON format (e.g., `{"peak_de": 2.34, "avg_de": 0.51, "rms": 0.61}`).
-*   **Rationale**: The UI currently has to use regex to parse human-readable text (e.g., `errors(CIEDE2000): max. = ...`). A JSON payload guarantees robust extraction of quality metrics.
-
-#### Instrument Enumeration (Global)
-*   **Feature**: A flag (e.g., `chartread -u -c list`) to query connected USB instruments and return a JSON array of device ports and names.
-*   **Rationale**: Allows the UI to automatically populate the "Select Instrument" dropdown without relying on OS-level USB polling or brittle string parsing of `chartread -?`.
-
----
-
-## 10. Development Milestones
-
-*   **Milestone 1: Project Scaffolding & State Management**: Build the core host application (Rust/C#/C++) and the non-blocking subprocess invocation abstractions to safely manage AGPL binaries.
-*   **Milestone 2: Target Generation Pipeline**: Implement UI forms and validation for Stages 1 & 2 (`targen` and `printtarg`), yielding printable TIFF files.
-*   **Milestone 3: Interactive Reading Engine**: Build the state machine, live swatch grid, $\Delta E$ math, and `chartread -u` IPC pipe handler.
-*   **Milestone 4: Profiling & Verification**: Implement the `colprof` execution thread and `profcheck` parsing for the final report.
-*   **Milestone 5: Advanced Visualisations**: Integrate 3D gamut plotting and UX polish.
+ICCery is licensed under the [MIT Licence](LICENSE). ArgyllCMS binaries and source code are licensed under the GNU Affero General Public License (AGPLv3).
diff --git a/ROADMAP.md b/ROADMAP.md
index c991c2c..a9b2c5d 100644
--- a/ROADMAP.md
+++ b/ROADMAP.md
@@ -1,257 +1,72 @@
-# Printer Profiling UI Project: Comprehensive Requirements & Architecture Roadmap
+# ICCery Development Roadmap & Architecture Strategy
 
-This document serves as the detailed feature set roadmap and architectural blueprint for integrating ArgyllCMS into a modern, native GUI application specifically dedicated to **Printer Profiling**. The goal is to abstract the complexities of the command-line utilities into a seamless, visual, and user-friendly experience.
+This document outlines the architectural roadmap, completed milestones, and upcoming development goals for **ICCery**.
 
 ---
 
-## 1. Architectural Overview & Licence Boundary
+## 1. Architecture Summary
 
-To maintain compliance with the ArgyllCMS AGPLv3 licence whilst allowing the host UI application to be closed-source, a strict inter-process communication (IPC) boundary must be maintained.
-
-```mermaid
-flowchart TD
-    subgraph host_app ["Host Application (Closed Source)"]
-        UI[Native User Interface]
-        State[State Machine & Orchestrator]
-        IO[Subprocess IPC Manager]
-        
-        UI <--> State
-        State <--> IO
-    end
-    
-    subgraph argyll_cms ["ArgyllCMS (AGPLv3)"]
-        BIN_TAR[targen]
-        BIN_PRT[printtarg]
-        BIN_CHR[chartread]
-        BIN_COL[colprof]
-    end
-    
-    IO -- stdin/stdout/stderr pipes --> BIN_TAR
-    IO -- stdin/stdout/stderr pipes --> BIN_PRT
-    IO -- stdin/stdout/stderr pipes --> BIN_CHR
-    IO -- stdin/stdout/stderr pipes --> BIN_COL
-    
-    classDef hostStyle fill:#0f52ba,stroke:#000,stroke-width:2px,color:#fff;
-    classDef agplStyle fill:#800000,stroke:#000,stroke-width:2px,color:#fff;
-    
-    class UI,State,IO hostStyle;
-    class BIN_TAR,BIN_PRT,BIN_CHR,BIN_COL agplStyle;
-```
-
-### Core IPC Guidelines
-1. **No Library Linking**: The host application must never link against `libinst`, `libicc`, or any other Argyll C library.
-2. **Subprocess Management**: Binaries must be invoked via OS process spawning (e.g., `std::process::Command` in Rust, `System.Diagnostics.Process` in C#).
-3. **Data Passing**: Data is passed via file paths (e.g., `.ti1`, `.ti2`, `.ti3` files) and standard streams (`stdin` / `stdout`).
+ICCery is a native, cross-platform desktop application built with:
+- **Backend**: Rust + Tauri v2, managing asynchronous process pipes, native printer devmode configurations (Windows GDI & Linux CUPS), and filesystem operations.
+- **Frontend**: Vanilla JS (ES Modules) + HTML5/CSS3 with a modern dark theme and responsive layout.
+- **Visualisation**: Three.js WebGL engine for 3D CIELAB color gamut volumes and sRGB reference comparisons.
+- **Engine**: ArgyllCMS command-line utilities orchestrated over isolated standard stream IPC (`stdin`, `stdout`, `stderr`).
 
 ---
 
-## 2. Technology Stack Recommendations
+## 2. Completed Milestones
 
-Given the requirements for a cross-platform (Windows, macOS, Linux), compiled (non-scripted), and closed-source compliant application, the following technology stacks are highly recommended:
+### Milestone 1 — Core Scaffolding & Subprocess Engine (`v0.1.0`)
+- [x] Initialized Tauri v2 + Rust multi-threaded asynchronous process manager.
+- [x] Implemented non-blocking event-driven stream pipes (`process:stdout`, `process:stderr`, `process:exit`).
+- [x] Established strict AGPLv3 isolation boundary.
 
-1.  **Rust with Tauri, Slint, or Iced**
-    *   **Pros**: Compiles to a highly optimised native binary, memory-safe, fearless concurrency for background tasks.
-    *   **Licence**: Permissive (MIT / Apache 2.0). Fully safe for closed-source.
-2.  **C# with .NET (Avalonia UI or MAUI)**
-    *   **Pros**: Strong object-oriented ecosystem, highly productive. Avalonia provides pixel-perfect identical rendering across all OSs.
-    *   **Licence**: Permissive (MIT). Fully safe for closed-source.
-3.  **C++ with wxWidgets or Qt**
-    *   **Pros**: Industry standard for cross-platform C++ UIs. Highly performant.
-    *   **Licence**: wxWidgets uses a modified LGPL that permits static linking in closed-source apps. Qt requires dynamic linking under LGPLv3 or purchasing a commercial licence.
+### Milestone 2 — Patch Set & Target Generation (`v0.1.3`)
+- [x] Integrated Stage 1 `targen` for RGB/CMYK patch generation with presets and neutral axis boosting.
+- [x] Integrated Stage 2 `printtarg` multi-page layout generation for major spectrophotometers (i1Pro, i1Pro2, ColorMunki, SpyderPrint).
+- [x] Built working directory resolution and cross-platform binary detection.
 
-*(Note: Scripted frameworks like Python/PyQt and heavy web-wrappers like Electron/React are explicitly excluded per architectural constraints).*
+### Milestone 3 — Interactive Measurement & Live Delta E (`v0.1.6`)
+- [x] Developed bidirectional `chartread -u` state machine with prompt-driven calibration and strip reading.
+- [x] Real-time JSON stream parsing (`ROW_COLORS_JSON`) for dynamic swatch grid updates.
+- [x] Implemented live CIEDE2000 ($\Delta E_{00}$) calculation with traffic light quality indicators.
+
+### Milestone 4 — Profile Calculation & Quality Verification (`v0.1.9`)
+- [x] Integrated Stage 4 `colprof` cLUT profile mathematical generation.
+- [x] Integrated Stage 5 `profcheck` numerical verification report (Peak, Average, RMS $\Delta E$).
+- [x] Added automated dark-theme UI with responsive status indicators.
+
+### Milestone 5 — Native OS Printing & Previews (`v0.1.13`)
+- [x] Added raw unmanaged print pipelines (Windows GDI unmanaged printing + Linux CUPS `raw` queue printing).
+- [x] Added high-resolution base64 TIFF thumbnail rendering pipeline.
+- [x] Optimized window constraints to 1280×800 minimum bounds.
+
+### Milestone 6 — 3D Gamut Visualization & Platform Extension Handling (`v0.1.14`)
+- [x] Extracted `.gam` gamut datasets via `iccgamut` and integrated 3D CIELAB WebGL viewer.
+- [x] Bundled reference `sRGB.gam` wireframe overlay for comparative volume analysis.
+- [x] Resolved platform profile extensions (`.icm` on Windows vs. `.icc` on Unix).
+
+### Milestone 7 — Wizard Disk-Artefact Gating & v0.2 Ship (`v0.2.0`)
+- [x] Implemented disk-artefact gating for wizard stepper navigation (`.ti1` → `.ti2` → `.ti3` → `.icc`/`.icm`).
+- [x] Eliminated all hardcoded placeholder and fallback crutches across JavaScript modules.
+- [x] Comprehensive documentation, release testing, and packaging automation.
 
 ---
 
-## 3. The Complete Profiling Workflow
+## 3. Future Roadmap
 
-The application must guide the user through a linear wizard-style process.
+### Milestone 8 — Advanced Measurement & Workflow Enhancements (`v0.3.0`)
+- [ ] **Averaging & Multi-Pass Reading**: Integrate Argyll's `average` utility to merge multiple measurement sheets for enhanced accuracy.
+- [ ] **Ambient & Display Profiling**: Extend wizard beyond reflective targets to emissive displays (`dispwin`, `dispread`).
+- [ ] **Preset Management**: Save and load reusable profiling recipes (paper types, patch counts, quality presets).
+- [ ] **Full 3D Convex Hull in CIELAB**: Replace 2D projected Delaunay triangulation with full 3D Delaunay/Convex Hull in Lab space for complex non-convex gamuts.
 
-```mermaid
-sequenceDiagram
-    participant User
-    participant UI as Host UI
-    participant TAR as targen
-    participant PRT as printtarg
-    participant CHR as chartread
-    participant COL as colprof
-    
-    Note over User, COL: Stage 1: Patch Generation
-    User->>UI: Select RGB/CMYK & Patch Count
-    UI->>TAR: targen -v -d [colour_space] -f [count] target_name
-    TAR-->>UI: Generates .ti1 file
-    
-    Note over User, COL: Stage 2: Target Creation
-    User->>UI: Select Instrument & Page Size
-    UI->>PRT: printtarg -v -i [instr] -p [size] -t 300 target_name
-    PRT-->>UI: Generates .ti2 & .tif files
-    UI->>User: Display .tif for printing (Colour Management OFF)
-    
-    Note over User, COL: Stage 3: Measurement
-    User->>UI: User clicks "Start Reading"
-    UI->>CHR: chartread -v -u target_name
-    CHR-->>UI: Live JSON Stream (ROW_COLORS_JSON)
-    UI-->>User: Live Visual Updates (Swatches, Delta E)
-    CHR-->>UI: Generates .ti3 file upon completion
-    
-    Note over User, COL: Stage 4: Profile Calculation
-    User->>UI: Select Profile Quality & Name
-    UI->>COL: colprof -v -q [quality] target_name
-    COL-->>UI: Generates .icc / .icm profile
-    UI->>User: Display Gamut & Quality Report
-```
+### Milestone 9 — macOS Native Support & Enhanced Print Spooling (`v0.4.0`)
+- [ ] **macOS Platform Bundle**: Build and sign universal macOS `.dmg` bundles.
+- [ ] **macOS Raw Spooling**: Native CoreGraphics/CUPS raw print dialog bypass.
+- [ ] **SpectroScan & Automated Table Support**: Support XY automated scanning tables in `chartread`.
 
----
-
-## 4. Stage 1: Patch Set Creation (`targen`)
-
-The first step is configuring the numerical colour combinations to be printed.
-
-### UI Configuration Details
-*   **Printer Colour Space**: Toggle between RGB (Driver-managed) or CMYK (RIP-managed).
-    *   *Command flag*: `-d 2` (RGB) or `-d 4` (CMYK).
-*   **Total Patch Count**: Controls profile accuracy vs. reading time. Provide presets (e.g., "Draft (400)", "Standard (800)", "High Quality (1500)", "Ultra (2500+)") and a custom input.
-    *   *Command flag*: `-f <number>`.
-*   **Neutral Axis Boosting**: Inject extra grey/black patches for smoother B&W printing.
-    *   *Command flag*: `-e <white_patches> -B <black_patches>`.
-
-### Execution
-The UI launches the subprocess and monitors `stderr` for completion.
-```bash
-targen -v -d 2 -f 800 -e 4 -B 8 my_profile
-```
-
----
-
-## 5. Stage 2: Target Image Generation (`printtarg`)
-
-Once the `.ti1` patch file exists, it is converted into a physical page layout formatted specifically for the user's measurement hardware.
-
-### UI Configuration Details
-*   **Hardware Instrument**: Dropdown list of supported devices.
-    *   *i1Pro/i1Pro2*: `-i i1`
-    *   *ColorMunki*: `-i CM`
-    *   *SpyderPrint/SpectroScan*: `-i SS`
-*   **Page Size**: Dropdown for media dimensions.
-    *   *Command flag*: `-p A4` or `-p Letter`.
-*   **Output Format**: Default to high-resolution TIFF for maximum compatibility.
-    *   *Command flag*: `-t 300` (8-bit) or `-T 300` (16-bit).
-
-### Execution & Printing Instructions
-```bash
-printtarg -v -i i1 -t 300 -p A4 my_profile
-```
-> [!IMPORTANT]
-> The UI must explicitly warn the user: **The generated TIFF files MUST be printed with colour management completely disabled in the printer driver.** (e.g., "No Colour Adjustment" on Epson, "Off (No Colour Adjustment)" on Canon). Failure to do so ruins the profile.
-
----
-
-## 6. Stage 3: Interactive Measurement (`chartread`)
-
-This is the most complex stage for the UI, requiring real-time bi-directional communication with `chartread` utilising the `-u` stream.
-
-### Interaction State Machine
-```mermaid
-stateDiagram-v2
-    [*] --> Initialize
-    Initialize --> AwaitCalibration: Spawn chartread -u
-    AwaitCalibration --> Calibrating: Send Spacebar
-    Calibrating --> AwaitStrip: Calibration OK
-    
-    AwaitStrip --> ReadingStrip: Wait for hardware trigger / button
-    ReadingStrip --> ProcessRow: Read success
-    ReadingStrip --> ErrorState: Read failed (Too fast/slow)
-    
-    ErrorState --> AwaitStrip: Retry
-    ProcessRow --> AwaitStrip: Row complete, parse JSON
-    ProcessRow --> Finished: Last row completed
-    
-    Finished --> [*]: .ti3 generated
-```
-
-### IPC Mechanics
-1. **Stdout Parsing**: The UI reads `stdout` line-by-line.
-    *   Lines matching `/^ROW_COLORS_JSON: (.*)/` are parsed as JSON to update the live patch grid and Delta E traffic lights.
-    *   Unprefixed lines (e.g., "Hit [Space] to read strip A") trigger state changes in the UI prompt text.
-2. **Stdin Injection**: The UI provides buttons (e.g., "Calibrate", "Skip") that inject keystrokes (`" \n"`, `"s\n"`) into the child process's `stdin`.
-
-### Visualising Delta E
-For live feedback, the JSON stream contains `expected.Lab` and `measured.Lab`. The UI should:
-1.  Implement the **CIEDE2000 ($\Delta E_{00}$)** formula.
-2.  Calculate the difference for every patch as it arrives.
-3.  Display a traffic light indicator (e.g., Green: $\Delta E < 2$, Amber: $\Delta E < 5$, Red: $\Delta E > 5$) on the UI swatch grid to instantly flag printing or reading errors.
-
----
-
-## 7. Stage 4: Profile Calculation (`colprof`)
-
-Once reading is finished and the `.ti3` file is generated, the mathematical ICC profile generation begins.
-
-### UI Configuration Details
-*   **Profile Quality**: Controls LUT resolution and optimisation passes.
-    *   *Command flag*: `-q l` (Low), `-q m` (Medium), `-q h` (High), `-q u` (Ultra). *Note: Ultra can take a very long time.*
-*   **Profile Description Name**: The internal name embedded in the ICC file that appears in Photoshop/Lightroom dropdowns.
-    *   *Command flag*: `-D "My Profile Name"`.
-*   **Algorithm**: Typically `cLUT` for printers.
-    *   *Command flag*: `-a l`.
-
-### Execution
-This process is highly CPU intensive. The UI must launch this in an asynchronous background thread and display an indeterminate progress spinner.
-```bash
-colprof -v -q h -a l -D "Canon Pro-100 Luster Paper" my_profile
-```
-
----
-
-## 8. Display of Results & Verification
-
-### Post-Processing Integration
-1.  **Mathematical Quality Score (`profcheck`)**:
-    Silently execute `profcheck` to compare the generated profile against the measurement data.
-    ```bash
-    profcheck -k my_profile.ti3 my_profile.icc
-    ```
-    *Parse the output*: Extract "Peak error" and "Average error". A healthy profile typically has an average $\Delta E$ below 1.5. Display this clearly in the UI.
-2.  **3D Gamut Visualisation**: 
-    To provide a premium feel, integrate a native 3D WebGL (if using Tauri) or OpenGL (if using C++/C#) viewer. Parse the gamut boundary data to render a rotatable 3D plot of the printer's colour volume in CIELAB space.
-
----
-
-## 9. Proposed ArgyllCMS Upstream Modifications
-
-To ensure maximum value, stability, and responsiveness in the UI, several low-level features should be developed within the ArgyllCMS codebase. These modifications primarily revolve around structured data output (JSON) to prevent the UI from relying on fragile regex string parsing of human-readable console text.
-
-### 9.1 Implemented Enhancements
-*   **`chartread -u` (JSON Streaming)**: *Already developed and merged.* We added the `-u` flag to `chartread`, which emits a real-time, flush-forced `ROW_COLORS_JSON: { ... }` stream. This solved the critical UI requirement of displaying live swatches and computing Delta E ($\Delta E$) mid-measurement without waiting for the `.ti3` file to close.
-
-### 9.2 Recommended Future Enhancements
-
-#### `targen`: Structured Progress
-*   **Feature**: Add a `-u` flag to emit JSON progress events during generation (e.g., `{"event": "progress", "stage": "optimising", "percent": 45}`).
-*   **Rationale**: Advanced patch generation (especially iterative OFPS with adaptation) can take several minutes. A structured progress stream prevents the UI from appearing frozen.
-
-#### `printtarg`: Structured Target Manifest
-*   **Feature**: Add a `-u` flag to output a JSON manifest upon completion, detailing the exact filenames generated (e.g., `target_01.tif`, `target_02.tif`), their physical dimensions, and patch counts per page.
-*   **Rationale**: Currently, `printtarg` generates multiple pages implicitly. A UI requires this structured manifest to reliably load and display the correct number of pages for printing without guessing the filenames based on the basename.
-
-#### `colprof`: Structured Profile Calculation Progress
-*   **Feature**: Add a `-u` flag to emit JSON progress during profile calculation (e.g., `{"event": "gamut_mapping", "percent": 75}`).
-*   **Rationale**: High and Ultra quality profile generation is highly CPU-intensive and time-consuming. Parsing dots (`.`) or numbers from `stdout` for progress bars is unreliable; structured progress ensures smooth UI feedback.
-
-#### `profcheck`: Structured Delta E Report
-*   **Feature**: Add a `-u` flag to output the final $\Delta E$ report in JSON format (e.g., `{"peak_de": 2.34, "avg_de": 0.51, "rms": 0.61}`).
-*   **Rationale**: The UI currently has to use regex to parse human-readable text (e.g., `errors(CIEDE2000): max. = ...`). A JSON payload guarantees robust extraction of quality metrics.
-
-#### Instrument Enumeration (Global)
-*   **Feature**: A flag (e.g., `chartread -u -c list`) to query connected USB instruments and return a JSON array of device ports and names.
-*   **Rationale**: Allows the UI to automatically populate the "Select Instrument" dropdown without relying on OS-level USB polling or brittle string parsing of `chartread -?`.
-
----
-
-## 10. Development Milestones
-
-*   **Milestone 1: Project Scaffolding & State Management**: Build the core host application (Rust/C#/C++) and the non-blocking subprocess invocation abstractions to safely manage AGPL binaries.
-*   **Milestone 2: Target Generation Pipeline**: Implement UI forms and validation for Stages 1 & 2 (`targen` and `printtarg`), yielding printable TIFF files.
-*   **Milestone 3: Interactive Reading Engine**: Build the state machine, live swatch grid, $\Delta E$ math, and `chartread -u` IPC pipe handler.
-*   **Milestone 4: Profiling & Verification**: Implement the `colprof` execution thread and `profcheck` parsing for the final report.
-*   **Milestone 5: Advanced Visualisations**: Integrate 3D gamut plotting and UX polish.
+### Milestone 10 — Enterprise Colour Workflow (`v1.0.0`)
+- [ ] **Custom CGATS Export**: Export and import industry-standard `.ti3` / `.txt` CGATS datasets.
+- [ ] **Batch Verification & Drift Tracking**: Track printer drift over time by comparing periodic verification measurements against a baseline profile.
+- [ ] **Multi-Language Localization**: Full UI internationalization (English, German, French, Japanese).